Nanotherapeutic treatment of the invasive glioblastoma tumor microenvironment

Abstract

Glioblastoma (GBM) is the most common malignant adult brain cancer with no curative treatment strategy. A significant hurdle in GBM treatment is effective therapeutic delivery to the brain-invading tumor cells that remain following surgery within functioning brain regions. Developing therapies that can either directly target these brain-invading tumor cells or act on other cell types and molecular processes supporting tumor cell invasion and recurrence are essential steps in advancing new treatments in the clinic. This review highlights some of the drug delivery strategies and nanotherapeutic technologies that are designed to target brain-invading GBM cells or non-neoplastic, invasion-supporting cells residing within the GBM tumor microenvironment.

1. Introduction

1.1. Glioblastoma Overview:

Glioblastoma (GBM, IDH wildtype grade 4 astrocytoma, IDH mutant grade 4 astrocytoma) [1,2] is an aggressive type of primary brain cancer, with a median survival of 15 months from the time of diagnosis [3]. Transcriptomic studies of human GBM biopsies have showcased that these tumors are highly heterogeneous [4,5], and neural stem cell-like glioma cells with high propensity for migration and invasion drive GBM pathology. These invasive glioma cells can migrate extensively into the brain and are often resistant to therapy [6,7]. Histopathologically, high-grade gliomas have areas of necrosis surrounded by highly cellular regions of viable tumor, beyond which tumor cells exhibit invasion into the surrounding healthy brain tissue. Currently, the standard-of-care treatment for GBM patients entails surgery for maximal safe tumor resection and subsequent radiation and oral chemotherapy using Temodar (i.e., temozolomide (TMZ)) [8,9]. The invasive nature of GBM makes complete surgical resection quite challenging without compromising neurological function; consequently, GBM recurs in 95% of the cases, with tumor recurrence commonly occurring near (~2–3 cm) the resection cavity [10,11]. Recurrence is associated with a dismal survival prognosis as evidenced by a median progression free survival of 1.5–6 months and median overall survival of 2–9 months [12,13]. Notably, there have only been marginal improvements in the survival statistics for GBM patients over the last 30 years [14–17].

A long-standing problem in GBM treatment is the delivery of therapeutics to brain-invading tumor cells without injuring adjacent neural cells and structures. Treatment of these invasive cells is significantly limited by the fact that most drugs accumulate in the brain at therapeutically low concentrations due to a unique physio-anatomic interface in the central nervous system (CNS) - the neuro-vascular unit (NVU) - commonly referred to as the blood–brain barrier (BBB) [18,19]. While the endothelial cell tight junctions, efflux pumps, and restricted transcytosis of the BBB effectively protect the CNS from exogenous biochemical threats, they also bar a majority of small molecule chemotherapies and almost all of the administered large therapeutics from reaching the brain and exerting target-specific efficacy [20]. Although some regions of brain tumors exhibit a disrupted BBB, the BBB typically remains functionally preserved in areas where brain-invading tumor cells intersperse with normal non-neoplastic cells [21]. Interstitial drug delivery and other local drug delivery approaches circumvent the transport barriers of the BBB, delivering therapies in situ and more extensively into the brain tissue [22]. The clinical feasibility and safety of these approaches have been repeatedly demonstrated in clinical trials; however, limited drug levels in the region of invasive tumor cells remain a critical challenge to generating meaningful, durable treatment outcomes for GBM patients [23]. Therapies that inhibit specific molecular targets driving GBM pathobiology have not yet been clinically successful, while current standard-of-care cytotoxic therapies (chemo-radiotherapy) do not reduce GBM invasion. Indeed, in a few instances radiation has been shown to promote glioma cell invasiveness [24]. In addition, recent studies suggest that intratumoral molecular heterogeneity and clonal evolution over time and in response to current treatments play central roles in GBM pathobiology and the development of treatment resistance, making invasive GBM extremely challenging to treat via single agent chemotherapies [25].

1.2. Nanomedicine approaches for GBM therapy

There has been a longstanding interest in developing advanced drug delivery systems for a multitude of diseases, with some nanotherapies approved for cancer treatment 20 years ago [26]. Nanoparticle (NP)-based drug delivery systems can overcome the limitations of conventional chemotherapy by improving systemic biodistribution, increasing drug circulatory half-life, and enhancing intracellular trafficking [27]. In addition, the physiochemical properties and surface profiles of NPs can be engineered with molecular ligands that enable tumor-specific targeting [28]. Furthermore, therapeutic NPs offer control over a drug’s release kinetics - thus reducing the need for repeated administration - and NPs can be formulated with multiple drugs for combination therapies [29].

Several NP systems have been evaluated in preclinical settings for GBM drug delivery, including lipid-based NPs [30,31], polymeric NPs [32,33], dendrimers [34], micelles [35], and inorganic NPs: gold, silica, iron oxides, quantum dots [36–41]. These NP systems differ in size, chemical composition, shape, surface properties and several other physiochemical parameters. The physiochemical properties of NPs dictate their ability to transport across or circumvent the BBB, navigate through the brain tissue microenvironment, and specifically traffic towards and into diseased cells, which affect their brain delivery and efficacy [42–46]. Studies looking at the diffusion of NPs within rat and human brains estimate the pore size distribution of brain extracellular spaces to be within the range of 100–200 nm [47]; hence, NPs that have an appropriately small hydrodynamic size (~100 nm) and are physiochemically slightly anionic or neutral in surface charge have been reported to be able to navigate through the brain parenchyma [32,48,49]. Depending on the preclinical brain tumor model, delivery strategy, and the NP physiochemical surface profile, successful delivery across the BBB into the brain has been reported using NPs possessing hydrodynamic diameters of 2–250 nm [50,51]. In some cases, the transport of larger NPs (>200 nm) across the BBB has been reported due to the compromised BBB in the tumor core [47,52].

1.2.1. Clinical nanoparticles used for GBM drug delivery

Despite many preclinical studies demonstrating NP-mediated therapeutic delivery and efficacy to the brain in various GBM animal models, there are no FDA-approved nanotherapeutics for GBM patients [42]. Several NP platforms are currently under evaluation in clinical trials for either imaging or therapy of brain tumors (Table 1). These nanotherapeutics are either lipid-based or inorganic in their material composition.

Table 1.

Summary of clinical trials investigating the use of NP-based delivery systems for brain tumor therapy.

Trade Name	Clinical Phase/NCT#	Nano formulation	Administration Route/Targeting Mode	Clinical Goals
NanoTherm	Phase II / Europe (2011)	Amino silane coated SPIONs	Local intracavity delivery	Effect of magnetic iron oxide NPs mediated intra-tumoral thermotherapy and concomitant external beam radiotherapy on overall survival in patients with recurrent GBM
Faraheme (AMAG) Ferumoxytol	Phase (undefined) NCT00660543 NCT02359097	Ferumoxytol: Iron polyglucose sorbitol carboxymethyl ether colloid	Non targeted/passive uptake across compromised BBB	MRI imaging evaluation using Ferumoxytol in assessing early response in patients with GBM receiving TMZ and radiation therapy
	Phase (undefined) (NCT00103038)	Ferumoxytol non-stoichiometric magnetite NPs	Intravenous/passive uptake across compromised BBB	Improving MR imaging of blood flow around brain tumors in patients diagnosed with high-grade brain tumors or cerebral metastases
EGFR(V)-EDV-Dox	Phase I (NCT02766699)	Bacterially derived minicell (400 nm) functionalized with EGFR bispecific antibodies and carrying doxorubicin as a cytotoxic payload	Intravenous/passive uptake across compromised BBB -BsAb-targeted, payload containing EDV nanocells transit across the compromised BBBs associated with tumors and deliver into the tumor microenvironment	Dose escalation study, evaluating the tolerability, safety, and immunogenicity of EGFR targeted doxorubicin containing NPs in patients with recurrent GBM
AGuIX ®	Phase I (NCT02820454) Phase II (NCT03818386)	Polysiloxane matrix /gadolinium chelates-based sub 5 nm NPs	Intravenous/passive uptake across compromised BBB	MRI imaging and radio sensitization of brain tumors
NU-0129	Phase 0 (NCT03020017)	Spherical nucleic acid gold NPs	Non targeted/passive uptake across compromised BBB	Safety and tolerability study of systemically administered NU-0129 NPs in patients with recurrent GBM or gliosarcoma undergoing resection surgery
Onyvide® (MM-398)	Phase I (NCT02022644) Phase I (NCT03086616)	PEGylated liposomal Irinotecan	Convection enhanced delivery	Dose finding study, evaluating MTD dosing of liposomal irinotecan, administered via CED. Phase I early efficacy study of MRI imaged assisted-CED administered liposomal irinotecan in children with DIPG.
SGT-53	Phase-II (NCT02340156)	Tumor-targeted cationic liposome loaded with a plasmid encoding for wild-type p53 DNA	Intravenous/ TfR targeting via a single-chain antibody fragment (TfRscFv)	Open label, single arm study, evaluating the safety and efficacy of SGT-53 NP and oral TMZ combination therapy in patients with GBM recurrence.
C225-ILs-Dox	Phase I (NCT03603379)	Doxorubicin-loaded anti-EGFR-antibody immunoliposomes	Intravenous/EGFR targeting via cetuximab/passive uptake across compromised BBB	Study evaluating the pharmacokinetics (delivery/NP accumulation within resected tumors/CSF/peripheral blood) of systemically administered C225-ILs-Dox NPs in GBM patients
Caleyx Peg-Dox	Phase I-II (NCT00944801)	PEGylated liposomal Doxorubicin	Non targeted/passive uptake across compromised BBB	Evaluating the safety and tolerance and efficacy of co-therapy including Caleyx Peg-Dox NPs in extended TMZ administration and radiotherapy in patients with new diagnosis of GBM
2B3–101	Phase I/II (NCT01386580)	Glutathione PEGylated Liposomal Doxorubicin	Active transport across BBB via GSH transporters	Open label study, evaluating, safety, tolerability and pharmacokinetics of 2B3-101 NPs when administered as a monotherapy or concomitantly with trastuzumab in patients with recurrent GBM, or brain metastasis.
Myocet	Phase I (NCT02861222)	Non-PEGylated liposomal doxorubicin	Non targeted/passive uptake across compromised BBB	Study evaluating safety and tolerance of liposomal doxorubicin NPs in adolescents and children with malignant gliomas.

CED: Convection enhanced delivery; MTD: Maximum tolerable dose; DIPG: Diffuse intrinsic pontine glioma.

1.2.1.1. Lipid-based nanoparticles.

Lipid-based NPs are the major category of FDA-approved nanomedicines [26,53]. Liposomes are a class of lipid-based NPs, composed of phospholipids that can be constituted into unilamellar or multilamellar vesicular structures, where hydrophilic, hydrophobic, lipophilic drugs or their combination can be co-encapsulated within the same liposomal system, thereby providing feasibility for combination drug therapy [54]. Liposome physiochemical attributes, including lipid composition, hydrodynamic size, surface charge, and surface profile engineering - with targeting ligands or polymers - dictate their in vitro as well as in vivo stability profiles [55]. Currently, there are a few liposomal nanoplatforms that are being evaluated in the clinic for glioma treatment. For example, the liposomal formulation of doxorubicin (Myocet) was recently studied in a dose finding phase I clinical trial in adolescents with refractory/relapsed malignant glioma (NCT02861222). Intravenously administered Myocet had acceptable safety profiles and enabled dosing at a maximum tolerable dose (MTD) which ensured ~ 60% of the doxorubicin in circulation remained intact in a liposomal form, which may enhance drug delivery outcomes [56].

Liposomes without appropriate surface modifications designed to extend in vivo circulation tend to be swiftly cleared via the reticuloendothelial system; hence, liposomes are often PEGylated to enable clinical use [31]. A PEGylated nanoliposomal formulation of irinotecan (Onyvide^® (MM-398) was evaluated, concomitantly with TMZ treatment, in a phase I dose finding trial including 12 patients with recurrent GBM. The phase I study indicated that the concomitant treatment of Onyvide^® (MM-398) at 50 mg/m²/2 weeks and TMZ at 50 mg/m²/day is well tolerated; albeit with no significant efficacy [57]. A phase I clinical trial testing administration of nanoliposomal-irinotecan by convection-enhanced delivery in subjects with recurrent GBM is underway (NCT02022644).

Similarly, administration of a PEGylated nanoliposomal formulation of doxorubicin (Caleyx Peg-Dox), combined with radiation, and a prolonged regimen of TMZ have been studied in Phase I/II study involving 63 newly diagnosed GBM patients [58]. Caleyx Peg-Dox administered at 20 mg/m² before radiation (60 Gy) and every two weeks after 4 weeks after radiation was found to be safe and tolerable along with the prolonged TMZ regimen (150–200 mg/m², days 1–5 on a 28-day cycle for 12 weeks). However, it was noted that such a combination treatment of Caleyx-Dox and a prolonged TMZ regimen did not significantly advance overall survival in patients (NCT00944801).

In another clinical trial, the ability of epidermal growth factor receptor (EGFR)-targeted immunoliposomes (C225-ILs-Dox) to deliver doxorubicin to brain tumor tissue in patients with relapsed GBM harboring an EGFR gene amplification has been evaluated [59]. The study found that intravenously administered C225-ILs-Dox had a maximal tolerable dose up to 50 mg/m² and doxorubicin was detected within the resected tumors; however, as free doxorubicin administration was not part of the study design, it remains to be seen whether doxorubicin delivery to the brain tumors was due to EGFR targeting or due to release of free doxorubicin from the C225-ILs-Dox during circulation. In addition, doxorubicin was not detected in the cerebrospinal fluid, indicating that C225-ILs-Dox likely accessed the brain tumors via a compromised BBB (NCT03603379).

Systemic administration of NPs without specific targeting functionalities may show preclinical efficacy in brain tumors. However, these NPs rely on passive uptake across a compromised BBB around proliferating tumors [60,61] and cannot effectively target invasive tumor cells. A Glutathione (GSH) functionalized PEGylated liposomal doxorubicin nanoformulation (2B3–101) designed to target GSH transporters in the BBB has recently been evaluated in phase I/II clinical trials (NCT01386580) in GBM patients as well as patients with breast cancer brain metastases. Intravenously administered 2B3–101 NPs in 28 patients with brain tumors revealed that this nanoformulation had acceptable safety profiles and showed signs of antitumor activity [62].

Transferrin receptor (TfR) targeting, using cationic liposomes surface functionalized with anti-TfR single-chain antibody fragment (TfRscFv) and loaded with a plasmid encoding human wild-type p53 (SGT-53), is being evaluated for GBM-specific targeting along with BBB penetration. In Phase II clinical trials, TMZ treatment in combination with tumor-targeted SGT-53 liposomes was under evaluation to establish safety and efficacy profiles in GBM patients with tumor recurrence or progression. However, the study was terminated due to the small number of enrolled patients, which prevented any meaningful statistical analysis (NCT02340156).

1.2.1.2. Inorganic nanoparticles.

There is a growing interest in using inorganic NPs for cancer drug delivery and imaging. Inorganic NPs have inherent material properties that can impart distinct physical, electrical, magnetic and optical properties, which can be leveraged to fill niche functional applications, requiring unique properties unattainable by polymeric or organic biomaterials [63]. For example, inorganic NPs such as silica, gold, and iron oxide have been widely developed as nanotherapeutics for GBM drug delivery and imaging applications [64].

Gold NPs possess free electrons on their surface, oscillating at frequencies dictated by their size and shape, which make them useful as photothermal agents, computed tomography contrast agents, and radiosensitizers [65–67]. In addition, gold NPs can be readily surface-functionalized with various biomolecules making them versatile platforms for drug delivery [68]. The safety and bioavailability of a spherical nucleic acid gold NP platform (NU-0129) were evaluated in a single-arm, open-label, “window of opportunity” phase 0 clinical trial in 8 patients undergoing GBM resection [69]. Intravenous administration of NU-1029 (0.04 mg/kg), and subsequent GBM tumor resection 8–48 h later, resulted in no significant treatment-related toxicities indicating that NU-0129 was well tolerated in GBM patients. Furthermore, quantitative analysis of resected GBM tumors showed initial evidence of NU-0129 crossing the BBB (NCT03020017).

Iron oxide-based NPs have been applied as contrast generating agents to enable magnetic resonance imaging (MRI) of various tumors [70]. Currently, all of the FDA-authorized iron oxide-based NPs are approved as iron replacement therapies and not for tumor diagnosis or therapy [71]. Several iron oxide-based NPs are presently under evaluation in clinical trials as contrast agents for GBM tumor imaging. For example, an iron polyglucose sorbitol carboxymethyl ether colloid (Faraheme: AMAG/Ferumoxytol) is being evaluated as an MRI imaging contrast agent for early tumor growth response in GBM patients receiving TMZ and radiation therapy (NCT00660543). In another clinical trial (NCT02359097), AMAG/Ferumoxytol is being used to evaluate improvement in MR imaging of blood flow around brain tumors in patients with high-grade brain tumors or cerebral metastases [26].

Apart from MRI contrast enhancements, iron oxide-based NPs, especially superparamagnetic iron oxide nanoparticles (SPIONs), can provide added therapeutic modalities via hyperthermia-mediated tumor treatment using alternating magnetic fields. NanoTherm is an example of an amino silane coated SPION that has been evaluated and approved for intra-tumoral thermotherapy in combination with external beam radiotherapy in patients with recurrent GBM in Europe and is purported to be considered for clinical trials in the USA [72]. Another emerging class of inorganic NPs under study in numerous clinical trials for cancer are silica-based NPs [73]. Amongst these of notable mention is an ultrasmall 5 nm NP comprised of polysiloxane matrix/gadolinium chelates (AGuIX^®), which is being evaluated for MRI imaging and radio-sensitization of brain tumors [74]. AGuIX^® was investigated in a first-in-human phase I clinical trial involving 15 patients diagnosed with brain metastases from diverse primary tumors (breast, lung, melanoma, and colon) (NCT02820454). Intravenously administered AGuIX^® was safe, exhibited an MTD of 100 mg/kg and was able to enhance contrast for imaging of brain metastases [75]. Subsequently, a phase II clinical trial evaluating AGuIX^® radio sensitization effects is underway (NCT03818386). However, it should be noted that the deposition of AGuIX^® in brain tissue was attributed to an already compromised BBB in brain metastases [76].

Despite substantial preclinical evidence of improved brain delivery of therapeutics and associated efficacies shown in various GBM animal models, the evaluation of nanotherapeutics in clinical trials for GBM has not yet been successful. Recent reports have highlighted that the delivery efficiency of various NPs to the brain is quite low (median 0.8% injected dose) [77,78]. This is primarily due to the shortcomings of current NP systems to overcome drug delivery barriers toward targeting invasive GBM cells. NP drug delivery barriers such as transit through the BBB and subsequent NP dispersion within the brain parenchyma present critical challenges to target invasive GBM cells responsible for GBM progression and recurrence. The following sections will describe delivery strategies that can enable NPs or other drug delivery technologies to target invasive GBM cells or an invasion supportive component associated with the glioma tumor microenvironment (TME).

2. Drug delivery strategies to target brain-invading glioma cells

A long-standing challenge to efficient drug delivery to brain-invading glioma cells is the unique and complex neuro-vascular unit that forms the BBB. The BBB is comprised of endothelial cells (EC) linked via tight junction proteins, abutted by a unique basal lamina, distributed between pericytes and astrocytic end feets and further sporadically interlinked by microglia and neuronal endings [79]. This neurovascular unit regulates the transport of fluid, ions, nutrients, proteins, and immune cells to and from the CNS while imposing a barrier to exogenous biochemical threats. The BBB, in its regulatory role, presents a primary barrier to the delivery of small molecule drugs, therapeutic antibodies, and drug-loaded NPs.

2.1. Systemic delivery strategies to traverse the BBB

NP transport across the BBB can either be passively directed or actively promoted via functionalization of NP surfaces to exploit various active transport systems across the BBB endothelium. Passive transport refers to the paracellular diffusion of hydrophilic NPs in the constrained space among ECs of the brain capillaries, or the passive diffusion of lipophilic NPs across the endothelial plasma membrane. Active transport mechanisms utilize several carrier-, adsorptive-, receptor- and cell-mediated pathways for endocytosis and transcytosis [80].

2.1.1. Nanoparticle surface profile engineering for enhancing BBB transport

NPs can potentially be delivered non-invasively across the BBB by harnessing endogenous transport processes, including, receptor-mediated transcytosis (RMT), carrier-mediated transcytosis (CMT), and adsorptive-mediated transcytosis (AMT). Amongst these processes, RMT-mediated pathways have primarily been utilized for NP delivery through the BBB (Fig. 1A). RMT-mediated transcytosis across the BBB begins on the luminal side of the brain capillary and microvascular ECs. The binding between a ligand and its cognate receptor initiates a sequential trafficking cascade entailing, receptor-mediated endocytosis - regulated via clathrin-decorated or non-clathrin-decorated vesicles - that are transported intracellularly and then sorted into multivesicular bodies that finally fuse along the abluminal side of the BBB, enabling transport and release of the trafficked constituents into the brain parenchyma [81]. Several ligands and antibodies which bind RMT-specific receptors on the BBB have been developed and recently reviewed [81].

Fig. 1. Drug delivery strategies to increase nanoparticle (NP) transport across the blood–brain barrier (BBB).

(A) NPs with surface functionalities that target active transport receptors on the endothelial cell layer can traverse across the BBB and reach brain-invading tumor cells. (B) Microbubbel-enhanced focused ultrasound (MB-FUS)-based BBB opening can transiently disrupt endothelial cell tight junctions, thus increasing NP transport and therapeutic delivery to brain-invading glioma cells. Hemispherical phased array transducers allow ultrasound beam steering and refocusing, enabling exact and controlled spatial activation of circulating microbubbles, which oscillate and expand/contract in the acoustic field resulting in spatially precise BBB opening.

2.1.1.1. BBB transport via targeting the transferrin receptor.

The transferrin receptor (TfR) is an extensively investigated target for RMT-based brain delivery strategies [82,83]. The TfR is abundantly expressed on the brain endothelium. It is also frequently overexpressed in highly proliferative cells, like glioma cells, thus enabling NP-based tumor-specific targeting [84,85]. Several recent reviews have summarized the various preclinical NP technologies that have utilized TfR-mediated BBB transport and showed therapeutic efficacy in glioma models [44,48,49]. Numerous preclinical studies utilizing TfR-mediated RMT to enable transport of therapeutic agents through the BBB have been reported [86]. For example, NP surface conjugation of antibodies against the TfR (clone R17217, and OX26) is reported to enhance NP-mediated delivery of therapeutic agents incapable of BBB transport [87–89]. However, the actual efficiency of TfR-targeting delivery systems, including NPs, to transport therapeutic cargos across an intact BBB is unclear [81,90,91]. Prior studies looking at the endocytic pathway of transferrin in rat BBB ECs, using in situ brain perfusion and electron microscopic evaluations, have shown that TfR is highly polarized at the BBB and spatially concentrated on the apical membrane. In addition, only a minimal amount of transferrin (~10%) remains associated with the perfused brain, while most of the transferrin is retro-endocytosed by brain ECs returning to the luminal face of the BBB [92]. Such TfR recycling suggests that therapeutics targeted to TfR, including NPs, are likely also recycled back to the luminal face of the BBB, limiting their delivery into the brain parenchyma.

The avidity of transferrin (Tf) itself or TfR antibody-conjugated NPs to TfR is critical to their RMT-mediated transit through the BBB and subsequent delivery into the brain parenchyma. Systemically administered Tf-conjugated PEGylated gold NPs (AuNPs) engineered to have a range of Tf avidities show differing BBB transport profiles and accumulation within the brain parenchyma of mice [93]. AuNPs with high Tf content (100–200 Tf units) and hence high avidity were consistently associated with blood vessels and thus showed lower accumulation in the brain parenchyma than non-targeted NPs. In contrast, AuNPs with lower Tf content (20–30 Tf units) and hence low avidity, showed higher accumulation in the brain parenchyma. This finding has also been recapitulated in TfR antibody-conjugated AuNPs as well as liposomal NPs [78]. Such low levels of NP transport can necessitate administration of higher systemic doses, which can induce off-target toxicities, and thus can be a significant obstacle for clinical translation [93–95].

Emerging strategies using PEGylated AuNPs (~80 nm) have shown that addition of an acid cleavable linker connecting Tf with the NP core could enable NP RMT across the BBB while maintaining high-affinity binding to endothelium TfR [96]. Upon systemic administration in mice, the cleavable Tf-conjugated AuNPs bound with high avidity along the luminal side of the BBB and released into the brain parenchyma in higher amounts, compared to high avidity NP controls containing non-cleavable Tf. Interestingly, adding an acid cleavable linker to TfR-specific high-affinity antibody-conjugated NPs did not improve their in vivo brain uptake as it did for Tf-containing NPs. Furthermore, brain uptake was decreased for TfR-antibody-conjugated NPs compared to Tf-conjugated NPs, indicating that native Tf ligand rather than TfR antibodies on NPs surfaces may allow them to avoid significant endothelium retention and confer a more remarkable ability to cross the BBB via RMT.

Another essential aspect to consider for TfR-mediated NP targeting of invasive cells across the BBB is the vascular location of NP engagement and trafficking at the BBB. A recent study examining transcytosis of TfR-targeted NPs highlights that post-capillary venules are preferred sites for RMT-mediated brain transport of TfR-conjugated NPs [97]. TfR-targeted liposomal NPs, grafted with high affinity anti-TfR antibodies (clone RI7217), upon systemic administration were sequestered within venules and capillaries, with the highest association of NPs being at the capillaries. Furthermore, although the NPs moved unobstructed within the endothelium at both sites of NP accumulation, transcytosis-mediated brain delivery occurred principally at post-capillary venules, while there was insignificant NP accumulation within capillaries. This vascular locus of successful NP entry is anatomically associated with the perivascular spaces, which facilitate rapid NP transit post-transcytosis. Currently, the evidence for of NP transcytosis through the BBB is controversial, particularly for high-affinity TfR antibody-conjugated NPs [93,96,98–100]; however, this new insight indicating that lack of perivascular spaces impede NP transit on the abluminal side of capillaries, highlights that access to perivascular space as well as intracellular sorting dynamics may independently or even interactively impact NP transport and drug delivery into the brain and the targeting of invading GBM cells.

2.1.1.2. Targeting the LDL and LRP1 receptors:.

The low-density lipoprotein-(LDL) receptor family is a class of cell surface receptors crucial for transporting triglycerides and cholesterol via controlling the intracellular transport of apolipoprotein-containing lipoprotein particles [101]. These receptors are expressed in several different tissues and can also be found in the brain [102]. Specifically, the EC of the BBB expresses LDL and low-density lipoprotein receptor-related protein 1 (LRP1) receptors that govern lipoprotein transport into the brain [103].

NP-based drug delivery strategies to the brain have utilized apolipoprotein E (ApoE) and LDL-mediated transcytosis to traffic therapeutic NPs across the BBB. For example, ApoE-functionalized polymeric NPs utilizing LDL receptor-mediated transcytosis to deliver various encapsulated chemotherapeutic drugs (as well as growth factors) through the BBB have been discussed in recent reviews [101]. Several NP delivery strategies have also leveraged the propensity of NP surfaces to adsorb a ubiquitous protein corona to achieve NP transport across the BBB. NP surface modifications using the surfactant polysorbate 80 (P80) led to a NP surface protein corona enriched in apolipoprotein-A, which has been shown to facilitate RMT across the BBB via mimicking lipoprotein particles traversing the BBB through LDL-mediated endocytosis [104,105]. Such a transport effect across the BBB was even further augmented when ApoE coatings were directly localized onto NP surfaces without any surfactant anchors; using these approaches, transport across the BBB has been shown for NPs up to 250 nm in hydrodynamic size [106].

LRP1-mediated therapeutic delivery to the brain has been explored via utilizing the expression of LRP1 on cells of the BBB. NPs surface-functionalized with the peptide angiopep-2 have shown sucessful transport and delivery through the BBB via RMT in several glioma animal models [107,108]. Recently, pH-sensitive polymersome NPs were reported to enable delivery of the chemotherapeutic doxorubicin across the BBB via angiopep-LRP1 mediated RMT and enhance the therapeutic effects of chemo-radiotherapy, increasing survival of nude mice bearing human U-87-MG glioma cell xenografts [109]. However, the adequacy of LRP1 expression among the cells of the BBB is still under investigation [81]. Recent studies looking at expression levels of LRP1 in mouse brain vasculature report modest to low expression of LRP1 in cells of the mouse BBB, especially in brain capillary ECs [110,111]. Thus, future studies elucidating the expression of LRP1 among cell types of the mouse and human BBB in a healthy state relative to a diseased state may influence LRP1-mediated drug delivery for treating GBM.

Overall, RMT-based approaches have significant potential for NP-mediated BBB transport and therapeutic targeting of invasive glioma cells in the brain. However, current approaches have several shortcomings that need further development. For optimal brain-specific delivery of therapeutic NPs across the BBB via RMT or CMT pathways, the target receptor should have high expression in the ECs of the cerebral vasculature and minimal expression in the peripheral vasculature to limit off-target effects [81]. Currently, there have been no such ideal target receptors identified; thus, most approaches continue to rely on ubiquitously expressed receptors like TfR, insulin receptor, iron transporters such as melanotransferrin (MTf), glucose transporter such as Glut1 trans membrane solute carriers (e.g., CD98 heavy chain (CD98hc), and members of the LDLR family [81]. However, recent studies dissecting the quantitative expression profiles of various RMT receptors in mouse and human BBB are currently emerging [112], which may advance RMT-mediated drug delivery approaches for treating invasive GBM. In addition, novel approaches to enable prolonged NP circulation and transport across the BBB are emerging. In one study, a C6 glioma cell membrane preparation was used to coat drug-loaded NPs, which were able to avoid systemic immune clearance and achieve glioma homotypic targeting across the BBB in a C6 intracranial glioma model [113]. Another study reported the use of activated effector/memory CD4 + helper T cells (CD4 + T_EM cells) as NP transport vehicles, wherein systemically-administered, NP-conjugated CD4 + TEM cells were shown to be able to traverse the BBB and deliver NPs to the brain parenchyma in allogenic mice [114].

2.1.2. Focused ultrasound (FUS)-mediated BBB opening

The BBB impedes NP-mediated therapeutic delivery to the invasive cells within the normal brain parenchyma as outlined in previous sections. Indeed, invasive cells continue spreading under the protective shield of the intact BBB in the brain regions past the contrast-enhancing edge of the tumor [115].

Focused ultrasound (FUS) has re-emerged as a viable therapeutic modality after the pioneering work of the Fry brothers and other contributors in the 1950s and 1960s propelled by recent technological advances in ultrasound electronics [116–118]. FUS can prime tissues non-invasively in various anatomical locations, most notably brain tissues [119]. A prominent feature of FUS has unfolded since the discovery that intravascularly circulating microbubbles (MBs) can act synergistically with FUS for the reversible and safe opening of the BBB [120]. Of note, the technique of MB-enhanced FUS (MB-FUS) can occur in an incision-less fashion via the intact skull, an impactful technological evolution compared to the first FUS systems developed by the Fry brothers that required craniotomies [116]. Not surprisingly, MB-FUS has increasingly gained attention as a platform for delivering nanotherapeutics to the brain (Fig. 1B). Timbie et al. showed in a rat glioma model that MB-FUS leads to an increase of drug-loaded brain penetrating NPs (BPNs) across the BBB and blood-tumor barrier (BTB), thereby achieving reduction in tumor burden and improved survival times [115]. The authors utilized fluorescently tagged PEGylated polystyrene (PS-PEG) BPNs along with biodegradable cisplatin-containing BPNs (CDDP-BPNs). This FUS study was the first to show improved therapeutic efficacy in treating GBM when systemically administered nanotherapeutics were applied in conjunction with MB-FUS.

While MB-FUS treatments in the clinical realm are primarily performed by magnetic resonance imaging (MRI)-guidance, advances in the field of ultrasound electronics and acoustic feedback have allowed for the real-time control and monitoring of safety features during treatment (e.g., harmonic dose) [121–124]. The technological capabilities of controlling and monitoring FUS in real-time during treatment have led to this technology gaining traction, as witnessed by an increasing number of clinical trials for treating a broad spectrum of indications ranging from neurode-generative brain pathologies to infiltrating gliomas [125–128]. Furthermore, the development of clinical-grade MBs, engineered initially as ultrasound imaging contrast agents [129,130], has rapidly spurred several clinical evaluations of MB-FUS BBBO, and various clinical trials in the United States (NCT04417088 NCT03551249, NCT03322813, NCT04667715) and internationally.

Most notably, there is mounting evidence that MB-FUS exerts secondary effects beyond BBB and BTB opening by modulating the interstitial space in the TME, another major impediment restricting the dispersion of nanotherapeutics [131]. In a recent study, Curley et al. showcased in a preclinical model the use of the MB-FUS platform for transfecting brain tumor cells by enabling delivery of systemically administered BPN gene vectors through the BBB and BTB [131]. This study concluded that MB-FUS-mediated BTB and BBB opening leads to significantly enhanced interstitial tumor flow leading to increased BPN transport through the tumor tissue [131]. While a part of such increased interstitial flow is attributed to lowering of hydraulic resistance within micro-vessel walls separating elevated pressure in the micro-vessels from lower pressures within the tumors, the study authors hypothesized that FUS could also directly enhance interstitial pore size and thus cause further reduction in hydraulic tissue resistance, an effect recapitulated in other studies [131,132].

2.2. Local delivery strategies bypassing the BBB to target invasive glioma cells:

Local delivery strategies that bypass the BBB can be used to administer therapeutics that are not effective via systemic delivery [22]. The BBB restricts the ability of systemically administered drugs to be delivered at therapeutic concentrations to residual postsurgical invasive glioma cells. In addition, systemic toxicities of free drugs limit their dosing, and their sub-therapeutic doses can enable the development of secondary resistance by GBM cells [133,134]. Therapeutics designed for local delivery approaches require physiochemical properties that minimize their clearance from the brain interstitial spaces while maximizing their distribution within the brain parenchyma.

2.2.1. Interstitial implants/intracavity drug delivery

As maximal safe resection is often the first stage of GBM treatment, many potential strategies are being investigated to deliver drugs directly to the resection cavity to treat residual invasive GBM cells [135]. Intra-cavity depot drug delivery systems are amongst the strategies that can locally deliver high drug doses within the resection cavity to treat residual infiltrative GBM cells while minimizing systemic exposure (Fig. 2B). Currently, carmustine-loaded biodegradable Gliadel^® wafers are the only clinically-approved formulation for local implantation into a brain tumor resection cavity, where carmustine can be released sustainably for 3 weeks [136]. However, GBM recurrence is not appreciably reduced by Gliadel^® wafer implantation, and thus these wafers only extend the median survival of GBM patients by 2 months [137–139]. The rigidity of the polymeric blend constituting Gliadel^® wafers makes them unconformable to the irregularly shaped resection cavities, limiting drug delivery to the entirety of the cavity and restricting carmustine diffusion into the brain parenchyma, preventing the eradication of residual infiltrative cells. Furthermore, the short half-life of carmustine (~15 mins) limits pharmacological activity over large diffusion distances and the heterogeneous nature of GBM makes recurrence likely as drug resistant tumor sub-clones rapidly overcome cytotoxic interference.

Fig. 2. Local drug delivery into the resection cavity can potentially treat residual glioma cells.

(A) Convection enhanced delivery has been a clinically relevant local delivery approach that can bypass the blood–brain barrier (BBB) and allow for transport of nanotherapeutics which can potentially improve therapeutic reach to invasive GBM cells. (B) Drug delivery technologies such as polymeric implants or hydrogel systems carrying small molecule drugs or nanotherapeutics can also be either locally implanted, injected or even sprayed within the resection cavity towards treating invasive GBM cells.

Several biomaterial-based approaches have emerged for efficacious intracavity drug delivery in preclinical brain tumor models to circumvent some of these limitations. These include engineered biodegradable polymeric gels containing encapsulated therapeutics that can be deployed locally in the resection cavity. OncoGel is an example of a thermo-responsive biodegradable hydrogel system constituting Poly (Ethylene glycol)-Poly(lactic-co-glycolic acid)- Poly (Ethylene glycol) (PEG-PLGA-PEG)-based triblock copolymers loaded with paclitaxel. OncoGel (0.63 % w/v paclitaxel) can form a gel in situ upon injection into the resection cavity and degrade within 4 to 7 weeks to provide controlled release of drug for ~50 days and reach therapeutic concentrations up to 9 mm from the injection site. A prior study in a rat 9L cell orthotopic gliosarcoma resection model demonstrated that OncoGel treatment significantly improved survival time, and the efficacy was further enhanced in combination with adjuvant radiotherapy [140]. This promising preclinical result has led to its further investigation in a Phase I/II clinical study, looking at the safety profiles, tolerability levels, and efficacy responses of OncoGel in patients with recurrent glioma (NCT00479765).

Recently, injectable NP-embedded viscoelastic gels made from lipid nanocapsules (GemC12-LNC) containing Lauroyl-Gemcitabine were tested in an orthotopic U87 glioma cell resection model [141]. The GemC12-LNC gels can adapt their shape to the unique topography of a resection cavity and conform to the brain parenchyma for sustained local drug delivery. However, no significant difference in delayed tumor recurrence was found in this model, compared to free gemcitabine administered intratumorally. However, the clinical applicability of this hydrogel formulation should be further evaluated in an infiltrative GBM model.

Recent studies demonstrated the in vivo efficacy of intracavity delivery of a combination of TMZ/etoposide-loaded biodegradable thermo-responsive PLGA/PEG paste that can mold to the uneven topography of a tumor resection cavity and sinters in situ, preserving immediate contact to the resected cavity [142]. Implantation of this biodegradable paste in a 9L orthotopic gliosarcoma resection model, followed by adjuvant radiotherapy, significantly increased survival. The loading of TMZ and etoposide in PLGA/PEG paste also improved the stability of the drugs as well as tolerance, suggesting that TMZ’s dose-limiting systemic toxicities were minimized via such intracavity delivery. These results, along with PLGA’s biocompatibility and safety in the brain tissue [143,144] bode well for the clinical potential of the TMZ/etoposide-loaded biodegradable paste.

Developing novel small molecules that specifically target oncogenic pathways while also possessing transport characteristics that enable their diffusion over long distances in an appropriate intracranial depot system constitutes a newly emerging therapeutic approach for treating invasive GBM. A novel small molecule, nbutylidenephthalide (BP), has been demonstrated to inhibit glioma cell migration and invasion by inhibiting Axl receptor tyrosine kinase activity, thus preventing tumor epithelial-to-mesenchymal cell transitions [145–147]. A recent study described the development of polyanhydride polymer-based BP-loaded wafers, which, when implanted into an orthotopic 9L GBM tumor model, enabled increased local drug concentration and drug diffusion over long distances and enhanced animal survival [148]. Furthermore, the BP wafer treatment significantly diminished the invasive GBM cell population compared to Gliadel^® treatments. Subsequent studies have shown that the BP wafer also enabled drug diffusion in excess of 20 mm in the canine brain and sustained BP at high concentrations (~3 mM, substantially higher than the IC₅₀ of the free BP) for 120 h.

2.2.2. Spray-based drug delivery:

Spray-based drug delivery systems utilize air pressure to deliver free drug or drug-loaded NPs within the resection cavity and into the brain parenchyma. Preclinical studies investigating the effectiveness of spray-based drug delivery to the resection cavity of GBM models are currently emerging. A recent study described the development of a sprayable bioadhesive thermosensitive formulation containing poly(lactic-co-glycolic acid) (PLGA) microspheres within a degradable poly(N-isopropylacrylamide) (PNIPAM) hydrogel [149]. This sprayable hydrogel system adhered to rat brains ex vivo, formed a gel in situ, and released encapsulated IgG and a model fluorescent dye in a controlled release pattern.

Bioadhesive hydrogels can prevent the clearance of NPs delivered to the resection cavity by holding them in place, while these NPs eventually diffuse into the brain parenchyma. However, such NPs must also possess the ability to spread within the brain tissue. A recent study demonstrated the development of a bioadhesive pectin-based hydrogel containing polylactic acid-polyethylene glycol (PLA-PEG) coated drug nanocrystals for delivery of a cytotoxic DNA crosslinker (Etoposide) and an inhibitor of poly-ADP ribose polymerase (olaparib) [150]. The pectin-based hydrogel adhered to the resection cavity in rat brains, forming a gel in situ via calcium-mediated cross-linking, and enabled diffusion of drug nanocrystals up to 1.5 cm into the brain parenchyma.

2.2.3. Convection enhanced drug delivery

NPs delivered into the brain must be capable of rapid transit within the brain extra cellular space and transport over many centimeters from their initial point of delivery. In such instances, diffusion alone may not facilitate spread within the brain and the application of facilitated diffusion mechanisms such as convective driven flow is necessary to establish therapeutic distribution within the brain [42]. Convection-enhanced delivery (CED) is a method for spatially targeted, local delivery to the brain that circumvents the BBB and allows for the administration of high drug doses with limited systemic distribution (Fig. 2A). CED disperses therapeutics based on a pressure gradient produced via a pump-catheter system, enabling convection-mediated delivery to enhanced volumes within the area of interest and to adjacent tissues [151]. Consequently, unlike local delivery approaches that rely on diffusion-mediated therapeutic delivery within the brain, CED facilitates a homogeneous concentration of drug over an enhanced volume of distribution (V_d) in the brain, as pressure gradients distribute therapeutics evenly over the V_d, independent of a therapeutic agent’s molecular weight or diffusivity characteristics [151]. Thus, in principle, CED is well suited to treat invasive GBM. However, clinical trials such as the TransMID and PRECISE trials, evaluating the use of molecularly targeted agents delivered via CED to treat gliomas, failed to demonstrate statistically significant improvements in patient survival relative to the Gliadel wafer treatments [152].

NPs can shield therapeutic drugs from clearance in the brain and serve as sustained release vessels. In several preclinical studies, CED slowly introduced sizeable volumes of NPs within the brain interstitium, increasing the residence time of NPs within the brain parenchyma, thus allowing the sustained release of encapsulated agents [49,153]. Gene therapy clinical trials using liposomal NPs delivered via CED have had limited therapeutic success due to the inability to achieve a homogenous V_d and transfection within the brain [151,154]. The inadequate spatial coverage of NP-based gene vectors in the brain is a longstanding hurdle in achieving clinical efficacy [155]. The heterogenous nature of the brain ECM presents a NP transport barrier, as adhesive or steric interactions between NPs and the brain ECM limit their spatial movement and can significantly impact their CED facilitated tissue distribution. Several studies have shown that, despite using CED, the adhesive interactions amongst cationic NPs and anionic nature of the brain ECM can confine the NPs within the injection site and to the perivascular spaces, limiting their penetration and subsequent distribution within the brain parenchyma [156–158]. Furthermore, studies have indicated that positively-charged NPs, even when shielded with PEGylation, have greatly restricted distribution away from the point of CED administration [156,159].

Recent studies have shown that high PEG densities on polyethylene glycol- polyethylenimine (PEG-PEI) based DNA NPs makes them non-adhesive to the rodent brain ECM, thus improving their distribution via CED and enabling widespread transfection while also mitigating any PEI-associated toxicity [160]. A relatively high amount of PEGylation was necessary to improve CED-facilitated distribution of PEG-PEI NPs, as insufficiently PEGylated NPs were restricted within the site of injection and thus were ineffective in inducing spatially widespread and sustained transgene expression in the brain post CED.

A substantial challenge in the delivery of NPs via CED includes a lack of understanding of tissue-specific factors that control the transport of NPs via convection. Tumor microenvironmental factors such as regional interstitial flow can influence the transport of NPs within areas of the brain harboring tumors. A recent study looked at the brain distribution of PLGA-based brain penetrating nanoparticles (BPNs) administered intratumorally via CED infusions in animals harboring U87 or RG2 cell intracranial tumors. The authors reported that although the overall NP V_d was comparable to those administered in healthy brains, tumor presence caused asymmetric and heterogenous distribution [161]. The extent of this non-uniform NP distribution was dependent on the heterogeneity of the tumors - size, degrees of necrosis, and changes in local interstitial flow. Interestingly, significant leakage of infused NPs into the peritumoral space was reported, with high concentrations of NPs accumulating by the tumor periphery, which may be advantageous for treating residual tumor cells in the GBM invasive rim.

Furthermore, CED to invasive regions of the brain parenchyma after resection is challenging as the presence of a low-pressure space created by the cavity leads to accumulation of infusate into the cavity, thus hindering effective distribution into the surrounding brain. In an alternative approach highlighting peritumoral infusion, CED of liposomal doxorubicin to the peritumoral brain tissue before tumor resection was demonstrated to increase NP delivery to the invasive regions, wherein, peritumoral infusion with intact tumor resulted in the most effective Vd in the brain surrounding the tumor [161]. In addition, peri-tumoral CED infusion of liposomes prior to tumor resection achieved more widespread V_d compared to intra-tumoral CED administration of liposomes. [162].

3. Examples of nanotherapeutics targeting the invasive rim region of GBM tumors

3.1. Nanoparticle targeting of invasive GBM cells

3.1.1. Fibroblast growth factor inducible 14-targeted nanoparticles

Targeting the invasive cells near tumor margins has remained a treatment challenge for glioma therapy. These cells are often left behind following debulking surgery and are difficult to eliminate after tumor resection due to their high resistance to clinical standard-of-care chemotherapy along with radiation. Among the emerging molecular targets implicated in invasive glioma biology is the TNF receptor superfamily member named fibroblast growth factor-inducible 14 (Fn14). This highly inducible transmembrane receptor is overexpressed in various solid tumors, including high-grade gliomas, especially in the mesenchymal GBM molecular subtype, which has a highly aggressive and invasive phenotype [163]. Fn14 activation in glioma cells triggers various downstream signaling pathways, including the NF-κB and Jak/STAT pathways, which are implicated as being critical drivers of tumor cell invasion in high-grade gliomas [164,165]. Furthermore, Fn14 expression is reported to be highly elevated in GBM cells near the tumor margins, in regions of high microvascular proliferation, and in recurrent tumors that have undergone standard-of-care treatments [166]. A recent study using the replication-competent avian sarcoma-leukosis virus/tumor virus A system (RCAS/tv-a) glioma modeling system has shown that Fn14 overexpression in rat gliomas promotes a more aggressive pathobiology due to increased brain invasion and an altered TME, ultimately resulting in significantly reduced animal survival times relative to animals harboring tumors with low Fn14 expression levels [167].

Our team has developed decreased adhesivity receptor targeting (DART) NPs that can selectively bind Fn14-positive glioma cells while maintaining minimal non-specific binding to surrounding brain extracellular matrix and blood components [168–170]. Such NPs optimized for maximum brain penetration and tumor cell targeting could be used to deliver a variety of payloads from traditional chemotherapeutics or radio-sensitizing agents to novel small molecule inhibitors of key glioma-driving pathways.

In recent years, as the importance of the TME composition related to invasive glioma biology and therapeutic success has become more clear, the focus of glioma therapy has also broadened to explore strategies that can concomitantly treat tumor cells as well as the key supporting non-tumor or immune cell types (e.g., tumor-supporting macrophages, activated microglia) [171]. Interestingly, analyses of single-cell RNA-seq data from human glioma samples reveal that Fn14 is also expressed by tumor-infiltrating macrophages in the glioma microenvironment [167]. This finding can potentially be leveraged to deliver novel immunomodulatory agents to the TME via the Fn14-directed DART NP platform. Emerging information about the impact of Fn14 in invasive glioma biology can thus be coupled with advances in NP engineering to foster the development of novel nano-enabled delivery schemes to overcome long-standing barriers in the treatment of invasive gliomas.

Furthermore, recent studies have leveraged 5-aminolevulinic acid (5-ALA) based fluorescence-assisted surgery to resect invasive margins of GBM tumors and further use tumor cell specific 5-ALA based metabolic fluorescence activated cell sorting to uniquely isolate invasive GBM cells [172]. Interrogation of these invasive GBM cells via single cell RNAseq have identified the human plasminogen activator inhibitor-1 (PAI-1) gene (SERPINE1) as a gene that is highly expressed in invasive GBM cells. Also, siRNA knockdown of SERPINE1 reduced glioma cell invasive capacity in vitro. Interestingly, SERPINE1 and Fn14 have been found to be frequently co-expressed in bulk tumor tissue as well as invasive tumor cells and higher expression of either is related to poorer prognosis in human GBM patients [167]. Taken together, SERPINE1 and Fn14 are important emerging molecular targets towards halting GBM invasion and development of NP-based therapies targeting these proteins may advance clinical therapy options for invasive GBM.

3.1.2. Nanoparticles targeting glioblastoma stem cells:

Glioblastoma stem cells (GSCs) are involved in GBM tumor recurrence due to their highly infiltrative nature and resistance to chemotherapy/radiotherapy [173]. Nestin and Prominin are putative GSC cell surface markers, with Nestin being highly enriched in invasive GSCs [174]. Several preclinical studies of NP-mediated targeting of GSCs for GBM therapy have been described [175–178]. A recent study reported the development of gold nanorods (AuNRs) functionalized with nestin-specific peptides for selective targeting and eradication of infiltrative GSCs [179]. The AuNRs containing nestin-binding peptides were able to selectively target single invasive GSCs migrating in an ex vivo 3D hydrogel-based invasive brain tumor model and accumulated specifically in invasive nestin-positive GSCs compared to nestin-negative cells. The application of near-infrared radiation resulted in the specific elimination of invasive GSCs via AuNR-mediated photo-thermolysis.

Recent studies highlight that GSCs can migrate along vascular structures and form perivascular niches following initial treatment, where they become elusive therapeutic targets residing across the BBB [180–182]. A recent study described multiplexed RNAi delivery via lipid-based NPs to target invasive GSCs [183]. A lipo-polymeric NP 7C1 (LPNP-7C1) formulation was used to encapsulate siRNAs targeting multiple transcription factors (OLIG2, POU3F2, SOX2 and SALL2) that direct the pro-neural GSC phenotype. The LPNP-7C1 platform showed an excellent safety profile and uniform penetration within the brain tissue after CED and targeted GSCs within vasculature-associated invasive niches. CED of LPNP-7C1 achieved significant tumor growth attenuation and extended the median survival in invasive GBM patient-derived xenograft mouse models.

Although direct targeting of GSCs using putative markers (Nestin, Prominin) as a way to curb GBM invasion has been shown to be effective in preclinical animal models, the heterogeneity of GBM tumor cells and clonal adaptation of GBM towards therapeutic approaches warrants development of interventions that can target the key molecular factors dictating GBM invasion. Recent studies using single cell RNAseq highlight that GSCs can undergo transformation to highly invasive cells using certain molecular pathways which enable them to lose expression of GBM stem cell markers while upregulating gene signatures associated with GBM invasion [184]. This analysis also revealed several molecules including long noncoding RNAs and transcription factors supporting this stem-to-invasion pathway. Specifically, the transcription factor EPAS1 (HLF2A, hypoxia-inducible factor 2A) was implicated in modulating GSCs fate towards highly invasive tumor cells and can be a potential molecular target for NP-mediated therapies targeting GSCs.

3.1.3. Nanoparticles targeting glioma-associated pro-invasive molecules:

Over the past years, RNA interference (RNAi)–mediated gene silencing has emerged as a potential therapeutic approach against many cancer types, including invasive GBM tumors [185,186]. However, short in vivo circulatory system stability, rapid elimination by the renal system, inefficient delivery to tumor sites, and inadequate treatment efficacy have restricted the clinical implementation of many RNAi-based therapeutics [187,188]. Several preclinical studies have demonstrated that some of these limitations can be overcome by using NP delivery systems, including inorganic, lipid-based, and polymer-based NPs for systemic administration of RNAi molecules [189–191].

Bcl2Like12 (Bcl2L12) is an effector caspase that is overexpressed in human GBM tumors when compared to the normal brain. This protein can promote GBM invasiveness via inhibition of the p53 tumor suppressor [192,193]. Recently, Jensen et al. described an RNAi-based NP platform, spherical nucleic acid (SNA) NP complexes, to silence Bcl2L12 expression in GBM [194]. SNA NP complexes are gold NPs surface functionalized with dense packings of well aligned small interfering RNA duplexes. Systemically administered SNA NP conjugates effectively crossed the BBB/BTB, disseminated throughout glioma tissue, and effectively reduced Bcl2L12 mRNA and protein levels within the tumor milieu, without any adverse side effects. SNA NP conjugates also improved survival in a human tumor neurosphere (huTNS)-derived xeno-geneic GBM mouse model, which recapitulates key human GBM phenotypic characteristics such as high necrosis, hypercellularity, microvascular proliferation, and diffuse and extensive brain invasion. Taken together, this preclinical work showed the promising potential of SNA NP conjugates as a systemic RNAi delivery platform suited to traverse the BBB/BTB and achieve silencing of pro-invasive molecular targets. Nevertheless, it should be noted that the SNA NP conjugates’ effectiveness was attributed to a compromised BBB. Hence, its ability to specifically target invasive GBM cells residing within regions of intact BBB is unclear at this time.

The role of small non-coding micro RNAs (miRNAs) in supporting GBM invasion and maintaining GBM stem cell plasticity is emerging [195,196]. For example, miRNA-21 and miRNA-10b have been directly associated with promoting GBM invasiveness, and their inhibition using RNAi has been demonstrated to be effective in curbing GBM invasion [197,198]. NPs encapsulating antisense oligonucleotides targeting various oncogenic miRNAs have been reported [199]. One recent study described the development of Poly (lactic-co-glycolic acid) (PLGA)-based NPs encapsulating antisense oligonucleotides targeting miRNA-21 and miRNA-10b [200]. These PLGA NPs inhibited both miRNAs in U87 GBM cells, which enhanced the cytotoxic effects (~2.9 fold) of TMZ in vitro as well as in a U87 GBM subcutaneous xenograft animal model.

Another recent nanotherapeutic approach highlighted the targeting of two independent aspects of glioma invasion: (a) Chlorotoxin (CTX)-mediated NP-targeting of chloride channels on invasive GBM cells and (b) Simultaneous silencing of miRNA-21 activity within the TME via delivery of NP encapsulated miRNA-21-anti-sense oligonucleotides, which subsequently, enhances the cytotoxic effects of concomitantly administered, targeted therapeutic agents such as the tyrosine kinase inhibitor, sunitinib [201,202]. Intravenously administered CTX functionalized nucleic acid lipid nanoparticles (SNALPs) carrying antisense oligonucleotides against miRNA-21 accumulated within GBM tumors and silenced miRNA-21 activity. CTX-mediated targeting enabled the SNALPs to specifically target invasive glioma cells, while silencing of miRNA-21 resulted in induction of tumor suppressor p53-related activity. In addition, a combination treatment of intravenously administered SNALPs along with oral sunitinib, resulted in the simultaneous activation of tumor suppressor p53-associated activity and suppression of the oncogenic NF-kB pathway that potentiated sunitinib’s cytotoxic effects, resulting in improved survival of tumor-harboring mice in an orthotopic GL261 mouse model of GBM.

Another emerging pro-invasive molecular target for GBM therapeutic development are long noncoding RNAs (LncRNAs), which have been shown as key molecular factors modulating GBM progression and invasion [203]. Recent single cell RNAseq studies have implicated a cluster of dysregulated genes (module 1) and associated LncRNAs dictating GBM invasion. The LncRNAs WWTR1, VIM-AS1, NEAT1 and AS1 were identified as being associated with GBM invasion [204]. Thus, development of NP-mediated therapies that can silence the activities of such LncRNAs can potentially be an effective strategy to target a central molecular node of gene dysregulation in GBM progression and tumor invasion.

3.2. Nanoparticles targeting invasion supporting non-neoplastic cells

3.2.1. Vascular niche targeting

An important node within the glioma TME that supports glioma invasion and proliferation is the vascular niche [205]. Glioma cells frequently localize and invade along extant brain structural components, including nerve tracks, blood vessels and the meninges. Invading glioma cells move adjacent to the vasculature and preventing this association may be a favorable therapeutic strategy to curtail glioma invasion. Preclinical studies on human glioma cell migration demonstrate that upon injection into the brain, the bulk (~>80%) of injected glioma cells home in to establish contiguity with blood vessels. This recruitment is mediated by the chemotactic signaling peptide bradykinin, produced by vascular ECs [206]. The binding between bradykinin and bradykinin 2 receptors on glioma cells causes downstream molecular signaling events controlling cell structure and volume modulations that are indispensable for an invading tumor cell.

Multiple experimental brain tumor models have shown glioma perivascular invasion [207,208] and heightened perivascular invasion has been reported in VEGF-deficient glioma cells [209,210]. This has also been shown in human glioma cell xenografts that were treated with anti-angiogenic therapies using anti-VEGF blocking antibodies [211,212] as well as in tumor tissues of GBM patients who had developed bevacizumab treatment resistance while undergoing antiangiogenic-bevacizumab therapy [213]. Orthotopically implanted glioma cells and GSCs from GBM tumors in mice, rats, and humans have been shown to grow in a preferential perivascular manner throughout tumor progression. Furthermore, perivascular growth has been observed in human biopsies of primary GBM and in genetically produced, de novo formed mouse gliomas [214]. Taken together, these reports underscore the importance of developing therapeutic strategies that can target perivascular niche-specific elements promoting glioma invasiveness.

Nanotherapeutics precisely targeting the brain perivascular region and associated blood vessels - that facilitate glioma proliferation and invasion - could be advantageous in treating invasive glioma by arresting diffuse tumor infiltration all around an interlinked micro-vessel network. NP therapies targeting tumor-supporting micro-vessels may also be efficacious against gliomas taking the perivascular route of invasion by eliminating preexisting brain micro-vessels already enveloped by brain-invading glioma cells.

A recent study described the development of peptide-conjugated drug-loaded polymeric NPs to simultaneously target both glioma cells and the accompanying blood vessel ECs. A CooP peptide targets the mammary-derived growth inhibitor heart fatty acid-binding protein-(H-FABP/FABP3)), which is concurrently overexpressed in glioma as well as ECs [215]. In U87MG tumor-harboring nude mice, intravenously administered peptide (CooP) functionalized, paclitaxel-containing, poly (lactic acid)-poly (ethylene glycol) NPs (CooP-NP-PTX) were able to selectively localize around tumor-associated blood vessels, and exerted cytotoxic effects on ECs, while also penetrating deeper into the tumor tissue. These combined cytotoxic effects led to increased survival time in U87 glioma-harboring mice receiving CooP-NP-PTX treatments (~47.5 days) compared to saline-treated (~20 days), Taxol-treated(~25.5 days), or untargeted NP-PTX treated (~29 days) animals.

Another vascular niche-associated protein promoting GBM invasiveness is the cytokine interleukin-6 (IL-6). IL-6 has been associated with activating the Jak/STAT signaling pathway which is a hallmark of aggressive and invasive GBM [216,217]. In addition, IL-6 augments the expression of several matrix metalloproteinases (MMPs) - MMP2, MMP9, and MMP14 - enhancing GBM invasion via MMP-mediated ECM degradation [218–220]. ECs within the perivascular niche have been identified as the primary source of IL-6 expression in the GBM TME [221]. IL-6 secreted by ECs can induce pro-tumorigenic phenotypes in GBM-associated macrophages. EC-specific IL-6 knockout is reported to inhibit this pro-tumorigenic macrophage activation and improve survival in GBM-bearing mice [221]. In addition, IL-6 within the GBM TME has also been recently shown to activate STAT3 expression in a reciprocal fashion, where glioma-derived IL-6 activated STAT3 expression and enhanced IL-6 expression in astrocytes. The astrocyte-derived IL-6 further enhanced glioma STAT3 activation and promoted glioma growth, apoptosis resistance, migration, and invasion [222]. Given these findings, nanotherapeutic targeting of IL-6 within the glioma TME may offer a selective and efficient therapeutic strategy for curbing GBM invasion.

A recent study described the application of doxorubicin-loaded polyglycerol-nano-diamond conjugates (Nano-DOX) that can target glioma cells and interrupt the IL-6/STAT3 activation loop within the TME [222]. Nano-DOX was loaded onto allograft mouse bone marrow-derived tumor-associated macrophages (Nano-Dox-TAMs). Systemic administration of Nano-Dox-TAMs in orthotopic U87 MG glioblastoma xenografts led to impedance of the STAT3/IL-6 governed mutual activation loop amongst glioma cells and astrocytes and modulated the glioma TME towards a tumor-suppressive type.

3.2.2. Nanoparticles for genetic programming of macrophages

Tumor-associated macrophages (TAMs) represent a considerable part of the non-neoplastic cell population within the glioma tumor microenvironment. TAMs generally exhibit an M2-like phenotype in GBM, which confers immunosuppressive functionalities on TAMs, enabling them to orchestrate tumor growth and invasiveness [223]. Clinical trials of small molecule drugs inhibiting the recruitment and association of TAM-precursor cells to tumors via targeted blocking of cellular recruitment and proliferation pathways (e.g., CSF-1/CSF-1R inhibitors [224,225]) showed low therapeutic responses unless supplemented with cytoreductive therapies or immune checkpoint blockers [226,227]. Emerging therapeutic approaches such as administration of CD40 agonists, interferon (IFN)-γ, interleukin-12 (IL-12), and Toll-like receptor (TLR) agonists aimed towards reprogramming TAMs to an M1-like phenotype to attenuate their pro-tumor functionalities and enhance anti-tumor immunity are in development [228–230]. However, these immunomodulatory agents can affect multiple cell types and hence are therapeutically limited due to dose-limiting adverse effects and associated systemic toxicities [231–233].

A recent study described the development of targeted biodegradable polymeric NPs that enable delivery of in vitro-transcribed (IVT) mRNAs - encoding M1-polarizing molecules towards reprograming M2 type TAMs - while obviating any non-specific systemic toxicities caused by disruptions in immune hemostasis [234]. Poly(β- amino ester) (PbAE)-based polymeric mRNA NPs (~100 nm) comprising IVT-mRNAs encoding for master regulators of macrophage polarization - the Interferon Regulatory Factor 5 (IRF5) and IKKβ (protein kinase activating IRF5) - were condensed into self-assembled NPs via charged based interactions between cationic PbAE and anionic mRNAs. The NP surfaces were further modified with Di-mannose moieties to enable TAM targeting. Systemic administration of these IRF5/IKKβ mRNA-PbAE NPs in PDGFβ-driven transgenic mouse glioma models resulted in BBB crossing and intracellular delivery of mRNA via ester bond hydrolytic release of mRNA from the mRNA-PbAE complex. When combined with radiation therapy, mRNA-PbAE NPs substantially reduced tumor growth and improved median survival compared to PBS controls (~52 days vs ~25 days, respectively). Taken together, these preclinical results showed the clinical potential of systemic NP-mediated mRNA gene delivery for TAM repolarization, which can remodel the TME and halt glioma progression without causing systemic toxicities.

Another recent study described the preclinical development of a rabies virus glycoprotein (RVG) peptide-conjugated polymeric core-mixed lipid shell-based NP platform for targeted paclitaxel (PTX) delivery to TAMs [235]. This NP platform is comprised of a biodegradable PTX loaded PLGA-based polymeric core along with a PEGylated mixed lipid monolayer shell that has a linked RVG peptide (RVG-PTX NPs). The RVG ligand enabled trans-vascular transport across the BBB and subsequent TAM targeting via its specific affinity to nicotinic acetylcholine receptors (AChR), which are present on BBB ECs as well as macrophages and microglial cells. Upon systemic administration in SCID mice bearing intracranial U87 gliomas, RVG-PTX NPs (~140 nm) were able to traverse the BBB, specifically target and enter into TAMs, and release PTX in a sustained manner. RVG-PTX NPs resulted in more significant brain accumulation than unconjugated PTX-loaded NPs. Notably, the RVG-PTX NPs avoided RVG-AChR binding in the neurons and subsequent uptake in non-phagocytic neurons, thus attenuating any non-specific neurotoxicity due to PTX. The sustained release of PTX within TAMs augmented the expression of pro-inflammatory cytokines (IL-6, TNF-α), that re-polarized TAMs from a tumor-supportive M2 phenotype to an anti-tumor M1 phenotype. These results support the clinical potential of such TAM-targeted nanotherapeutics for curbing the immunosuppressive glioma microenvironment and curtailing glioma invasiveness.

3.2.3. Nanoparticles targeting tumor-supporting astrocytes:

During GBM progression, glioma cells can activate astrocytes and coerce them to initiate expression of survival genes in tumor cells that confer chemotherapy resistance [236–238]. Glioma-associated, reactive astrocytes have been implicated in creating a tumor permissive microenvironment by the production of various proteins and release of chemokines that act in a paracrine manner to support glioma cells [219]. Within the perivascular niche, astrocytes are in close proximity to glioma cells, regulating and enhancing glioma cell invasiveness via paracrine factors, including, chondroitin sulfate proteoglycans, MMPs, and connective tissue growth factors.

Reactive astrocytes are critical promoters of GBM invasion, especially in the peritumoral invasive edge. Such reactive astrocytes overexpressing glial fibrillar acidic protein (GFAP) are found in substantial numbers and colocalize with invasive glioma cells. In addition, reactive astrocytes as well as aggressively migrating glioma cells overexpress connexin 43 (Cx43) [239]. Indeed, Cx43, a gap junction protein, is highly expressed within glioma-associated astrocytes, with highest expression within the peri-tumoral invasive region. A study looking at astrocyte-glioma cross talk with the TME in a syngeneic GL261 mouse glioma model, found that intercellular communication among astrocytes, governed by Cx43, played a significant role in facilitating glioma cell invasion into the brain parenchyma, especially in the early phases of glioma development. Abrogation of Cx43 from the TME reduced the brain-invading propensities of glioma cells [240]. Collectively, these indications highlight Cx43 as a promising molecular target for curbing glioma invasiveness and possible tumor recurrence. A NP-mediated targeted approach to Cx43 silencing could potentially alter astrocyte signaling and normalize a glioma permissive TME. However, more importantly, it could also attenuate Cx43 expression in residual glioma cells that may prognostically migrate and cause distant tumor recurrence.

Recently, GFAP and Cx43 dual targeting liposomes have been developed and tested in an invasive C6-intracranial GBM rat model. PEGylated immunoliposomes (~130 nm) surface-functionalized with both anti-GFAP and anti- MAbE2Cx43 monoclonal antibodies accumulated within the peritumoral invasive zone after systemic administration and colocalized with reactive astrocytes and invasive glioma cells [241].

In another study, poly(ethylene-b-PMAA) diblock copolymer-based nano gels conjugated with anti-Cx43 and anti-BSAT1 (a brain-specific anion transporter) monoclonal antibodies were shown to successfully target the peritumoral invasive zone in an intracranial 101/8 glioma model [242]. Systemically-administered, cisplatin-loaded Cx43 and BSAT1 targeted nano gels (~130 nm) accumulated within the peritumoral zone and significantly reduced tumor volumes and increased survival rates compared to cisplatin-loaded non-targeted nano gels.

3.3. Nanoparticles targeting glioma cell-associated pro-invasive factors in the tumor microenvironment

Targeting the glioma TME to suppress pro-invasive biological factors is a promising therapeutic strategy to curtail GBM invasiveness and recurrence. Preclinical and clinical data indicate that GBM invasion is dictated by various factors facilitating tumor spread along with different anatomic and molecular structures. The key migratory pathways taken by brain-invading glioma cells can be topographically categorized to perivascular spaces and parenchyma. These anatomical regions of GBM invasion are intrinsically different in their mechanical and physical constraints for glioma invasion [243,244]. These regional differences are due to the unique composition of ECM molecules within these two compartments. ECM components such as proteoglycans from the lectican group and their binding substrates, tenascins and hyaluronan, predominate the interstitial spaces of the brain parenchyma [245].

Glioma cells orchestrate their invasion through the extracellular space (ECS) via an organized process of attachment of the migrating cell edge, affixing to the ECM, followed by uncoupling of the trailing end. The contractile force needed to sustain glioma cell motility is provided by actin-myosin molecular motors. Myosin II has been notably implicated in enabling glioma invasion through the narrow spaces of the ECS [246]. Tumor microenvironmental cues mediate glioma cell attachment through intercellular and cell-ECM communication via certain receptors, including integrins, neural cell adhesion molecules and cadherins. In addition to utilizing basement membrane-based ECM molecules for attachment and movement, glioma cells overcome the physical constraint of the dense matrix of ECS via the expression of many secreted proteases, including MMPs [247,248]. In addition, glioma cells can coerce ECs, microglia, and astrocytes to produce MMPs. The concerted activity of these multiple cellular proteases reconditions the ECM to enable tumor cell migration and invasion. The application of MMP inhibitors to arrest glioma invasion is a potential therapeutic strategy; however, clinical trials using MMP inhibitors haven’t succeeded yet in treating GBM [243].

3.3.1. Integrin targeting nanoparticles:

The ECM is a key component of the GBM TME. It plays a supporting role in mediating various cell surface receptor-based signaling events amongst tumor cells and non-neoplastic tumor supporting cells. Among the various cell surface receptors, those within the integrin family have been reported to be involved in regulating GBM cell invasiveness, growth, and conferring therapeutic resistance via intercellular as well as cell-ECM interactions [249–251]. Integrins are heterodimers composed of one of the 18 α- subunits along with one of the 8 β-subunits, where the combination dictates specificity to a substrate as well as the signaling modality. The increased expression of several integrins (α3β1, αvβ3, αvβ5, α5β1, α6β1, α9β1) in glioma cells has been linked with increasingly invasive phenotypes [252]. The β1 subunit, when upregulated, is associated with heightened glioma invasion, while the αvβ3 and αvβ5 integrins facilitate tumor-induced angiogenesis [253–255]. Integrin-derived signaling also modulates the enzymatic activities and spatial organization of ECM degradative proteases at the invasive glioma front, and expression of integrin αvβ3 has been shown to correlate with expression of MMPs (MMP-2, MMP-9) in brain- invading glioma cells [256,257].

Cilengitide (CGT) is a cyclic arginine-glycine-aspartic acid (RGD) pentapeptide antagonist of the integrins avβ3 and avβ5, which are abundantly expressed on invasive glioma cells as well as on ECs [258]. CGT has been shown to block the integrin αvβ5 selectively and has shown modest anti-tumor activities in a phase II clinical trial in patients with recurrent GBM (NCT00093964) [259]. However, this therapy eventually failed to be efficacious when evaluated in a randomized Phase III trial involving patients with recently diagnosed GBM tumors which show methylation in the MGMT promoter (NCT00689221) [260]. The limitations to clinical translation of CGT therapy were centered around drug delivery issues, such as rapid clearance of CGT from blood, nonspecific accumulation in kidney and liver, and extremely low BBB permeability [261]. Recent studies demonstrated the effectiveness of focused ultrasound-mediated BBB opening (FUS-BBBO) in augmenting the transport and delivery of CGT nanotherapeutics to the brain [262]. Systemically administered gelatin/poloxamer 188 grafted heparin, copolymer-based CGT nanoparticles (CGT-NP) showed enhanced transport across the BBB upon application of FUS-BBBO in an orthotropic C6-rat glioma model. This increased transport of CGT-NPs across the BBB resulted in 3-fold higher delivery of CGT, along with significantly reduced kidney accumulation compared to free CGT administration with or without FUS-BBBO. Furthermore, even at a low CGT dosing (2 mg/kg ~ every 4 days), CGT-NPs + FUS-BBBO improved the median survival of glioma-bearing animals (from 20 days in untreated controls to 80 days). This finding indicated that a combination of CGT-NPs and FUS-BBBO can overcome the BBB transport barriers that limit standard CGT treatment for gliomas.

Studies have shown that NPs surface-functionalized with RGD peptides are able to deliver cytotoxic payloads to glioma cells through the blood-brain-tumor barrier (BBTB) [263]. The RGD peptides are ligands that target αvβ3/αvβ5 integrins overexpressed on glioma-associated vasculature and glioma cells themselves [264]. Systemically administered cyclic RGD (cRGD) peptide-functionalized polymeric micelles (~30 nm) incorporating the chemotherapeutic agent oxaliplatin (cRGD-OXA-NPs), showed active receptor-mediated transcytosis across the BBTB and delivery into the tumor tissue in orthotopic U87MG brain tumor models [265]. In addition, cRGD-OXA-NPs produced significant tumor efficacy in an orthotopic mouse U87MG GBM model, relative to polymeric micelles functionalized with the non-targeted ligand “cyclic-Arg-Ala-Asp” (cRAD-OXA-NPs) or free oxaliplatin drug.

Recent studies have highlighted integrin alpha-2 (ITGA2) as a potential molecular target for GBM therapy. ITGA2 is substantially overexpressed in human GBM cells relative to normal glial cells [266]. Furthermore, increased ITGA2 expression negatively correlates with GBM patient survival [266]. Antibody-based blockade of ITGA2 significantly impeded migration of GBM cells but not their proliferation, making ITGA2 an attractive molecular target for NP-based therapies against curtailing invasive GBM [266]. Indeed, a recent study demonstrated that ITGA2 antibody-functionalized doxorubicin-loaded liposomal NPs (ITGA2-Dox-LP) could cross a disrupted BBB and selectively target GBM cells for cytotoxic doxorubicin delivery, while also inhibiting GBM cell invasion [266].

3.3.2. ECM molecule targeting nanoparticles

Tenascin-C (TNC) is a glycoprotein that governs intercellular as well as cell-ECM interactions, and its expression is upregulated during cancer progression. TNC is highly expressed within the GBM TME and its expression is positively correlated with glioma grade, GBM invasiveness, and poor patient prognosis [267,268]. Invading glioma cells are a predominant source of ECM-associated TNC within the TME. TNC along with Tenascin W (TNW) localizes to glioma-associated blood vessels, stimulating vessel sprouting (angiogenesis) [269]. The basement membranes surrounding the vasculature in the brain are rich in other ECM molecules like fibronectin and vitronectin, which have also been implicated in increasing glioma cell motility [270–272].

Several ECM targeting strategies have been developed using NPs functionalized with high-affinity peptides against ECM components. However, such single target-based strategies can lead to higher ECM localization of NPs, but with lower affinity to tumor cells [169,273–275]. Recently, peptide-functionalized PEG-PLA NPs with dual affinity to TNC and glioma-associated antigens were developed to overcome this limitation. PEG-PLA NPs were conjugated with a dual affinity fusion peptide (Ft) containing binding sequences from FHK and tLyp-1 for simultaneous targeting of TNC and the glioma-associated vasculature protein neurolipin-1 (NRP-1) [276]. Ft-functionalized paclitaxel-containing PEG-PLA nanoparticles (Ft-NP-PTX) bound to the ECM components along the invasive glioma edge by TNC targeting. In contrast, concurrent NRP-1 targeting facilitated enhanced penetration of Ft-NP-PTX across the BBB. This enhanced penetration of Ft-NP-PTX within the TME improved survival of mice bearing intracranial U87MG tumors compared to FHK-functionalized NPs, tLyp-1-functionalized NPs or Taxol^®.

Another study highlighted the dual targeting of different molecular substrates within the tumor-associated ECM via a unique bispecific peptide (PL1: PPRRGLIKLKTS), which recognizes two clinically-relevant glioma microenvironment ECM proteins – TNC and fibronectin Extra Domain-B (FN-EDB) [277]. Systemically administered PL1-conjugated iron oxide nanoworms (NWs) and metallic silver NPs were localized within the invasive glioma microenvironment and delivered proapoptotic peptides that attenuated tumor cell proliferation and increased survival in orthotopic GBM mouse models. In a subsequent study, the same group demonstrated the efficacy of a next-generation PL3 peptide (AGRGRLVR)-functionalized NPs with dual affinity to TNC and NRP-1 [278]. Systemically administered PL3-functionalized NWs and metallic silver NPs showed increased tropism towards GBM lesions in orthotopic GBM mouse models and in areas of clinical tumor samples overexpressing TNC and NRP-1, suggesting potential for clinical translation. In addition, NWs containing proapoptotic D(KLAKLAK)2 peptides enhanced anticancer efficacy in a subcutaneous U87-MG glioma cell mouse model, compared to non-targeted NWs.

3.3.3. STAT3 targeting nanoparticles

The Signal transducer and activator of transcription 3 (STAT3) transcription factor is a critical node for GBM pathophysiology [279]. Most invasive mesenchymal GBM subtypes exhibit persistent STAT3 activation, and STAT3 is implicated in (i) regulating epithelial-to-mesenchymal transition (in concert with NF-κB) and (ii) promoting glioma stem cell invasiveness [217]. In addition, elevated expression of pro-invasive factors including MMP-2, MMP-9, focal adhesion kinase (FAK), and fascin-1 is linked to STAT3 activation [219]. Also, recent studies indicate that continued STAT3 activation in the non-malignant cells within the glioma TME promotes an immune tolerogenic condition (such as defective antigen presentation of dendritic cells) and promotes oncogenic activities of myeloid-derived suppressor cells and tumor-associated macrophages [216,217].

Emerging reports have also implicated STAT3 upregulation in the glioma TME with tumor recurrence and resistance to chemotherapeutic and radiation therapies [280–282]. This work identifies STAT3 as a promising molecular target to suppress GBM invasion, increase the glioma cell sensitivity to chemotherapy and radiation, and illicit anti-tumor immune responses without inducing cytotoxicity on non-neoplastic cells. Numerous clinical trials testing immunotherapeutic approaches targeting canonical STAT3 phosphorylation (at the Tyr-705 residue) in GBM patients have shown promising results [283,284]. However, they were unable to overcome STAT3-mediated immunosuppression caused by continual STAT3 expression in the glioma TME; possibly because delivery of anti-STAT3 therapeutics to the brain and subsequent sustained spatiotemporal STAT3 inhibition within the glioma microenvironment remain unresolved drug delivery challenges [283–285].

Recent studies describe the development of a human serum albumin-based NP platform capable of brain delivery across BBB for siRNA-mediated silencing of STAT3 activity within the TME [286]. Human serum albumin NPs (~200 nm) encapsulating siRNA against STAT3 and surface modified with a tumor tissue targeting/-penetrating peptide (iRGD) (STAT3i-NPs), demonstrated transit across the BBB and distributed within the tumor after systemic administration, facilitating the diffusion of STAT3 inhibiting siRNA within the TME. The iRGD targeting enabled the transit of NPs across the BBB via engaging endothelial αvβ3 and αvβ5 integrins and subsequent proteolytic cleavage, enabling eventual binding to NRP-1, resulting in endocytotic/exocytotic transport across the BBB. The ability to transport across the BBB and enable siRNA diffusion within the GBM microenvironment to inhibit STAT3 activity resulted in extended median survival (>10 days) of GBM-bearing mice relative to those treated with systemic STAT3 siRNA only. Furthermore, a combination of STAT3i-NP treatment with ionizing radiation was efficacious, leading to tumor inhibition and enhanced survival in ~ 88% of GBM-bearing mice, which further developed long-term immunologic memory as GBM tumor rechallenges in the contralateral hemisphere resulted in 100% survival without any intervening treatments. Collectively, these results demonstrate the promising clinical potential of STAT3i-NPs in combination with standard-of-care ionizing radiation towards overcoming activated STAT3 persistence within the GBM microenvironment, resulting in tumor regression and maintenance of immunologic memory.

4. Conclusions and future perspectives

4.1. Conclusions

The treatment challenges related to eradicating the residual brain-invading GBM cells left behind after tumor core resection is a central issue in GBM patient management. The high frequency and rapidity of recurrence, proximity of cancer cells to functional neural components, and the relatively intact nature of the BBB create major obstacles to effectively treating these cells. NP-mediated drug delivery approaches have the potential to eradicate glioma cells left behind after surgery as they can either be engineered to traverse the BBB via RMT, be directly delivered into the brain, or be used in conjunction with FUS-assisted BBB opening methodologies. In addition, local delivery methodologies that bypass the BBB, such as CED and hydrogel-based depot systems for the local release of therapeutics along the resection cavity, offer new opportunities for treating invasive components of the disease. The examples of preclinical drug delivery work covered in this review are considerable efforts to target brain-invading glioma cells directly or components of the glioma-brain microenvironment to counteract invasion (Fig. 3) (Table 2). Although a number of molecular processes have been implicated in glioma cell invasion that can potentially be therapeutically targeted (Table 3), specific NP-mediated targeting along with novel modes of enhancing NP transport across the BBB (via RMT, FUS mediated BBB opening, or local delivery) and further into the brain parenchyma to reach brain-invading cells are still important areas of investigation.

Fig. 3. Nanoparticle-mediated targeted drug delivery can be leveraged to therapeutically affect molecular nodes implicated in GBM invasion.

Molecular factors intrinsic to invading tumor cells (Fn14, miRNAs, Nestin) can be exploited to engage nanoparticles enabling direct targeting of invasive GBM cells. Nanoparticles can also target non-neoplastic cells supporting invasion (endothelial cells, tumor-associated macrophages, astrocytes), as well as pro-invasive factors within the TME (Integrins, Tenascins, STAT3). GBM: Glioblastoma, Fn14: Fibroblast growth factor inducible 14. miRNA: microRNA. TME: Tumor microenvironment. STAT3: Signal and Transducer of Activation 3. Some graphics in this figure were adapted from Servier Medical Art (CC BY 3.0 Unported).

Table 2.

Summary of nanoparticles targeting invasive rim region of GBM.

Targeting Class	Molecular Target/Cell Type	Nanotherapeutic	Mode of Delivery/BBB status	Reference
NPs targeting invasive GBM cells	Fn14	PLGA-PEG based polymeric DART NPs	Systemic Delivery/Disrupted BBB	[168–170]
	Nestin/GSCs	Gold nanorods (AuNRs)	Ex vivo model of intact BBB	[179]
	RNAi delivery using siRNA’s to target multiple transcription factors (OLIG2, POU3F2, SOX2 and SALL2) driving invasive GSCs	Lipid-based lipo-polymeric 7C1 (LPNP-7C1) Nanoparticles	Convection Enhanced Delivery/BBB by pass via direct delivery into the brain.	[183]
NPs targeting Glioma associated pro-invasive molecules	silence Bcl2L12 expression in GBM	Spherical nucleic acid (SNA) NP complexes functionalized with siRNA	Systemically administered/delivery across intact BBB	[194]
	miRNA-21 and miRNA-10b	(PLGA)-based NPs encapsulating antisense oligonucleotides targeting miRNA-21 and miRNA-10b	Systemically administered/delivery across a disrupted BBB	[200]
	Chlorotoxin (CTX)-mediated NP-targeting of chloride channels on invasive GBM cells and simultaneous silencing of miRNA-21 activity within the TME	CTX functionalized nucleic acid lipid nanoparticles (SNALPs) carrying antisense oligonucleotides against miRNA-21	Systemically administered/delivery across a disrupted BBB	[201,202]
NPs targeting invasion associated non neoplastic cells	CooP peptide mediated targeting of mammary-derived growth inhibitor heart fatty acid-binding protein-(H-FABP/FABP3)), concurrently overexpressed in glioma as well as ECs	CooP-functionalized, paclitaxel-containing, poly (lactic acid)-poly(ethylene glycol) NPs (CooP-NP-PTX)	Systemically administered/delivery across a disrupted BBB	[215]
	Targeting glioma cells and interrupting the IL-6/STAT3 activation loop within the TME	Doxorubicin-loaded polyglycerol-nano-diamond conjugates (Nano-DOX) loaded onto TAMs	Systemically administered/delivery across a disrupted BBB	[222]
	Delivery of in vitro-transcribed mRNAs (IVTs) - encoding M1 polarizing molecules (mRNAs encoding for master regulators of macrophage polarization - the Interferon Regulatory Factor 5 (IRF5) and IKKβ (protein kinase activating IRF5) towards reprogramming M2 type TAMs	Poly(β- amino ester) (PbAE)-based polymeric mRNA NPs (~100 nm) comprising IVT-mRNAs	Systemically administered/Delivery across an intact BBB	[234]
	Targeted paclitaxel (PTX) delivery to TAMs	Rabies virus glycoprotein (RVG) peptide-conjugated PLGA-based polymeric core along with a PEGylated mixed lipid monolayer shell, PTX loaded NP platform	Systemically administered/Delivery across an intact BBB	[235]
	GFAP and Cx43 dual functionalized liposomes targeting astrocytes and invasive glioma cells	PEGylated immunoliposomes (~130 nm) surface-functionalized with both anti-GFAP and anti-MAbE2Cx43 monoclonal antibodies	Systemically administered/delivery across a disrupted BBB	[241]
	Cx43 Astrocyte targeting and invasive glioma cell targeting	Cisplatin-loaded poly(ethylene-b-PMAA) diblock copolymer-based nano gels conjugated with anti-Cx43 and anti-BSAT1 (a brain-specific anion transporter) monoclonal antibodies	Systemically administered/delivery across a disrupted BBB	[242]
NPs targeting glioma cell-associated pro-invasive factors in the TME	Integrin targeting via the peptide: Cilengitide	Gelatin/poloxamer 188 grafted heparin, copolymer-based CGT nanoparticles	Systemically administered along with FUS-BBBD across intact BBB	[262]
	Integrin targeting via the peptide: RGD	RGD (cRGD) peptide-functionalized polymeric micelles (~30 nm) incorporating the chemotherapeutic agent oxaliplatin (cRGD-OXA-NPs),	Systemically administered/intact BBB	[265]
	Integrin targeting via the ITGA2 monoclonal antibody	ITGA2 antibody-functionalized doxorubicin-loaded liposomal NPs (ITGA2-Dox-LP)	Systemically administered/delivery across a disrupted BBB	[266]
	Dual affinity fusion peptide (Ft) mediated targeting of TNC and the glioma-associated vasculature protein neurolipin-1 (NRP-1) for targeting invasive GBM	Peptide-functionalized PEG-PLA polymeric NPs with dual affinity to TNC and glioma-associated antigens. Ft-functionalized paclitaxel-containing PEG-PLA nanoparticles	Systemically administered/delivery across an intact BBB	[276]
	Bispecific peptide (PL1: PPRRGLIKLKTS), recognizing two clinically-relevant glioma microenvironment ECM proteins – TNC and fibronectin Extra Domain-B (FN-EDB)	PL1-conjugated iron oxide nanoworms (NWs) and metallic silver NPs	Systemically administered/BBB status not reported	[277,278]
	siRNA-mediated silencing of STAT3 activity within the TME	Human serum albumin NPs (~200 nm) encapsulating siRNA against STAT3 and surface modified with a tumor tissue targeting/-penetrating peptide (iRGD) (STAT3i-NPs	Systemic delivery/transport across intact BBB	[286]

NPs: Nanoparticles; Fn14: Fibroblast growth factor inducible 14; PLGA-PEG: Poly (lactic-co-glycolic acid)-Poly (Ethylene Glycol); siRNA: Small interfering RNA; RNAi: Ribonucleic acid interference; BBB: blood brain barrier; FUS: Focused ultrasound; miRNA: microRNA; TAMs: Tumor associated macrophages; PTX: paclitaxel; TME: tumor microenvironment; Cx43: Connexin43; IL-6: interlukin-6, STAT-3: Signal transducer and activator of transcription 3.

Table 3.

Summary of clinical trials investigating the use of small molecules and biologics for halting glioma invasion.

Drug/References	Clinical Phase/NCT#	Composition	Molecular Target	Clinical goals/Outcomes
Cilengitide	Phase II/III (NCT00006093) (NCT00679354) (NCT00093964)	RGD peptide	Integrin avp5	Evaluation of antitumor activity of Cliengitide in patients with recurrent GBM. Moderate antitumor activity in Phase II studies, however failure to enhance overall survival in randomized Phase III trials
Prinomastat	Phase II (NCT00004200)	Small molecule drug	MMP inhibitor	Evaluating the effectiveness of prinomastat plus TMZ treatments in GBM patients.
Sulfasalazine	Pilot study (NCT01577966)	Small molecule drug	Inhibition of system xc, preventing glutamate release	Pilot non randomized trial in GBM patients, evaluating the ability oforally dosed sulfasalazine to alter glutamate concentrations in glioma cells monitored via magnetic resonance spectroscopy.
Chlorotoxin(TM-601).	Phase I/II clinical trial (NCT00040573)	36–aminoacid peptide	Cl – channel (ClC3) inhibitor	Open label single dose trial in 18 adult GBM patients, evaluating safety and tolerability of intracavitary application of chlorotoxin upon tumor resection.

4.2. Future perspectives

NP-based drug delivery to invasive GBM has immense potential to advance the clinical standard of care for GBM patients. In particular, with its potential to decrease free drug-mediated toxicities and increase maximal tolerable doses of systemically-administered chemotherapies, advanced nanotherapeutics may offer significant benefit to elderly GBM patient cohorts, who often have low compliance to the standard Stupp regimen of conventional fractionated radiotherapy and adjuvant TMZ therapy [287]. In addition, emerging studies also support that NP-mediated sustained drug delivery within the glioma TME can synergistically enhance the effects of chemo-radiotherapies on GBM tumors as well as have the potential to overcome TMZ resistance and or radio-sensitize GBM tumors at low therapeutic doses [288,289].

As new advances are made in elucidating novel molecular pathways underpinning GBM invasion, a better understanding of the invasive GBM TME will enable elucidation of novel molecular targets that can benefit from NP-mediated targeted drug delivery. Emerging studies leveraging the resolutive power of single-cell RNA sequencing techniques in dissecting the TME cross talk between non-neoplastic cells (endothelial cells, astrocytes, pericytes, glial cells, TAMs) and the glioma cells within the invasive glioma TME will facilitate the development of novel targeted NP drug delivery strategies that can therapeutically curb glioma invasion and recurrence.

At present, much of the benefit of nanotherapeutics for GBM treatment has been shown in in vitro settings and in preclinical GBM animal models. The subsequent failure to translate these efficacy findings to GBM patients in the clinic is in part due to the poor recapitulation of human GBM pathology in most current preclinical animal models of GBM [290]. While advanced animal models recapitulating human GBM pathology are emerging [291,292], NP BBB transport in rodents may not translate well into humans due to molecular and cellular differences between the human and rodent BBB [293]. To circumvent this limitation, brain organoid models of GBM and lab-on-chip based BBB models utilizing human-induced pluripotent stem cells are emerging as powerful alternatives to GBM animal models. These new models can faithfully recapitulate clinical GBM TME, human BBB properties, and model GBM invasiveness, which can accurately characterize response to therapeutic interventions [294,295]. The recent application of patient-derived GBM organoids for ex vivo screening and subsequent development of personalized therapies illustrates the significant potential of such technologies [296].

Novel nanotherapeutic strategies that synergize with emerging surgical technologies or immunomodulation of the GBM TME are currently emerging. FUS-mediated BBB opening is a promising modality for enhancing NP delivery across the BBB to regions of the brain harboring invasive GBM cells. As FUS-mediated BBB opening technologies continue to develop, new methodologies that can accurately assess and quantify NP drug delivery following FUS-mediated BBB opening and quantify NP trafficking within the invasive TME will further advance nanotherapeutics for treating GBM invasion. Another promising GBM patient treatment development is the use of immunotherapeutic strategies such as peptide, dendritic cell, oncolytic viruses-based vaccination, immune check point inhibitors and chimeric antigen receptor expressing T cells, which aim at overcoming the immune suppressive TME of GBM. However, most of these approaches have not demonstrated robust efficacy in GBM clinical trials and the use of combinatorial approaches combining one or more of the aforementioned modalities is suggested to improved immunotherapy outcomes for GBM [297]. NP-mediated drug delivery systems with their potential for delivery of multiple drugs and immunomodulatory agents may emerge as the technology that bridges this gap for GBM treatment [298,299].

Finally, to advance nanotherapeutics into the clinic for treatment of invasive GBM, the long-term safety profiles and clearance pathways of NPs must be characterized and delineated. A recent study has revealed the paravascular glymphatic pathway as the major route for the clearance of intra-parenchymally delivered organic NPs [300]. Future studies looking at the trafficking and clearance of various types of NPs delivered to the brain via RMT, FUS-mediated BBB opening, or CED-based local delivery will be critical for advancing nanotherapeutics into the clinic for invasive GBM treatment.