TABLE 1

Selected key structural classes

This table describes the key structural platforms used in brain cancer clinical trials and their core advantages and limitations.

Superclass	Subclasses	Advantages	Limitations	Useful Resources
Small molecules	Small molecules Small molecules + delivery system	Low molecular weight is favorable for penetrating the BBB Many available screening approaches Easy to use drug formulations can be available Targeting is not limited to membrane targets	On-target and off-target toxicities Clearance might be rapid, requiring an additional administration of an agent The delivery system might possess additional toxicities	(Ferguson and Gray, 2018; Macarron et al., 2011; Ovacik and Lin, 2018; Scherrmann, 2002)
Antibody products and derivatives	Whole antibody	Phage-display libraries allow the relatively easy design of a wide spectrum of antibodies (all antibody products) Long blood circulation, hence administration is less frequent	Tumor antigen escape (all antibody products) Distribution is typically limited to blood and interstitial fluids The target spectrum is usually limited to the membrane and free-floating proteins (all antibody products) On-target off-tissue toxicity (all antibody products) Immunogenicity that might lead to anaphylaxis and cytokine release syndrome (CRS) (all antibody products)	(Ovacik and Lin, 2018)
	Antibody-drug conjugate	Combine the potency of small molecules with the selectivity of antibodies Reduced off-target toxicity of the cytotoxic molecule Some versions possess bystander killing activity Linker design might help to overcome drug resistance	Deconjugation could cause toxicities The quantity of cytotoxic agent delivered is typically low	(Beck et al., 2017)
	Radio-antibody conjugate	Therapeutic and diagnostic tool The radioactive exposure is limited to a tumor site	Tumor exposure can be low The amount of exposure is not easy to quantify Radionuclides could be uptaken by a nontarget tissue/organ	(Bourgeois et al., 2017; Vivier et al., 2018)
	Antibody fusion product	Depending on the attached component this molecule could act as a vaccine, targeted toxin, or immune modulator	Molecular weight is higher than those of antibodies, and hence tissue penetration could be challenging	(Balza et al., 2010; Bao et al., 2016; Dhodapkar et al., 2014)
	Bispecific T-cell engager (BiTE)	Tumor antigen does not require MHC presentation to elicit the immune response Lower molecular weight favors tumor penetration	Short-term serum half-life that requires relatively frequent administration of an agent Immunogenicity that might lead to anaphylaxis and cytokine release syndrome	(Baeuerle and Reinhardt, 2009; Labrijn et al., 2019)
	Peptibody	Peptide could possess various pharmacological functions Fc domain improves the pharmacokinetic properties of a peptide and provides immune effector function	Immunogenicity is higher compared with peptides and may be undesired	(Cavaco et al., 2017; Shimamoto et al., 2012)
Protein and peptide molecules	Peptide	Peptide antigens are easier to produce than full proteins Peptide vaccines induce better responses than full protein vaccines	Pharmacokinetic properties are mediocre; renal excretion is an issue Stability is limited The use of adjuvants is required for proper immunogenicity (for vaccines)	(Cavaco et al., 2017; Hos et al., 2018; Mahmood and Green, 2005)
	Recombinant human protein	Immunogenicity is lower compared with nonhuman proteins	Pharmacokinetic properties may be an issue, thus modifications may be required	(Mahmood and Green, 2005)
	Protein fusion product	If a protein is conjugated with an Fc domain of an antibody, its pharmacokinetic properties (half-life) are improved One of the protein components could be used as a targeting moiety to deliver a toxin	Molecular weight can be high, hence cell permeability is limited Immunogenicity may be an issue	(Kawakami et al., 2003; Sperinde et al., 2020; Strohl, 2015)
	Protein-drug conjugate	Protein component is used for targeted delivery of a small molecule Conjugation may help delivery through BBB	Similar to antibody-drug conjugates	(Thomas et al., 2009; Venepalli et al., 2019; Vhora et al., 2015)
Cell products	Dendritic cells	Multiple antigen coverage The immune response is generally stronger compared with peptide vaccines Formulations could be personalized Different options to load tumor antigens are available	Antigens are MHC-restricted Production is complex, expensive, and hard to automate and unify Tumor microenvironment still could diminish the efficacy	(Sabado et al., 2017)
	CAR T cells	Complete response rates in B cell malignancies were as high as 90% Primary brain tumors may be a promising area for CAR T therapy given a relatively low mutation burden of these tumors Targets are not MHC-restricted Phage-display libraries allow the relatively easy design of a wide spectrum of recognition domains for CAR T cells Recognition domains could be also composed using other molecules than ScFv	Production is complex, expensive, and hard to automate and unify Tumor antigen escape renders CAR T cells completely ineffective if occurs Severe and even lethal systemic toxicities have been observed On-target off-tumor toxicity Lack of efficacy in solid tumors Targeting is limited to membrane tumor antigens	(Aijaz et al., 2018; Chandran and Klebanoff, 2019; Jackson et al., 2016; Levine et al., 2016; Rafiq et al., 2020; Sadelain et al., 2017)
	Stem cells	Many possible applications, which include drug delivery, chemoprotection, hematopoiesis, chemosensitization, immune modulation Some cell subsets can possess tumor tropism	Limited availability Production is complex Invasive administration Safety may be an issue	(Aijaz et al., 2018; Bexell et al., 2013; Mount et al., 2015; Parker Kerrigan et al., 2018)
	T cells	Multiple antigen coverage Formulations could be personalized Depending on amplification and modifications these cells could acquire additional properties, such as chemotherapy resistance	Additional stimulation (IL-2) is usually needed Antigens are MHC-restricted Production is complex, expensive, and hard to automate and unify Severe systemic toxicities may occur T cells may not be isolated from the initial tumor Tumor microenvironment could diminish the efficacy	NCT04165941, (Met et al., 2019; Stroncek et al., 2019)
	Modified bacterial cells	Relatively high immunogenicity Manufacturing is scalable	Toxicities may be high Risk of undesired infections Antivector immune response	(Lopes et al., 2019)
	NK cells	In contrast to T cells, preimmunization is not required Tumor recognition is different compared with T cells Immunomodulatory function	Additional stimulation is usually needed Production is complex, expensive, and hard to automate and unify Invasive administration Tumor microenvironment could diminish the efficacy	(Fang et al., 2017; Hodgins et al., 2019)
	Tumor lysate	Multiple antigen coverage Formulations could be personalized Production is relatively simple (compared with other cell-based products)	Immunogenicity could be low These formulations can lead to an autoimmune response Allogeneic vaccine sources might have different antigen composition than the patient’s tumor Tumor may not be surgically available to extract antigens Lysates may have immune-suppressive molecules from tumor cells	(Gonzalez et al., 2014; Olin et al., 2014; Rojas-Sepulveda et al., 2018)
	TCR T cells	Compared with CAR T cells, TCR T cells can be targeted against intracellular targets Less antigen density is required to trigger the immune response Downstream receptor signaling may be more persistent compared with CAR	Antigens are MHC-restricted Production is complex, expensive, and hard to automate and unify Severe and even lethal systemic toxicities have been observed Tumor microenvironment could diminish the efficacy	(Aijaz et al., 2018; Chandran and Klebanoff, 2019; D'Ippolito et al., 2019)
	CAR NK cells	Because of the biologic nature of NK cells, the allogeneic application might be safer for CAR NK cells rather than CAR T cells Primary brain tumors may be a promising area for CAR NK therapy given a relatively low mutation burden of these tumors Targets are not MHC-restricted Phage-display libraries allow the relatively easy design of a wide spectrum of recognition domains for CAR NK cells Recognition domains could be also composed using other molecules than ScFv	Common CAR NK cell line NK-92 is derived from non-Hodgkin lymphoma, thus raises safety concerns CAR NK cells are more sensitive to cryopreservation than T cells Other limitations similar to CAR T cells	(Burger et al., 2019; Wang et al., 2020)
	Bi-armed T cells	Bispecific antibodies could be designed against numerous antigens Antigens are not MHC-restricted	Severe systemic toxicities have been observed On-target off-tumor toxicity Targeting is limited to membrane tumor antigens	(Lum and Thakur, 2011; Zitron et al., 2013)
	NKT cells	Immunomodulatory function Tumor recognition is different compared with classic T cells and NK cells	Limited availability of NKT cells in cancer patients Tumor microenvironment could diminish the efficacy	(Chen et al., 2018; Nair and Dhodapkar, 2017; Terabe and Berzofsky, 2018)
	Tumor cells	Extracellular vesicles released by a vaccine are a good source of tumor antigens Formulation is personalized Immunogenicity of exosomes is considered stronger than those of peptides and lysates	Tumor-derived vesicles may be immunosuppressive This vaccine is implanted compared with other formulations	(Harshyne et al., 2015; Robbins and Morelli, 2014; Tarasov et al., 2019a)
Viral therapeutics	Adenovirus HSV Measles virus Parvovirus Poliovirus Reovirus Vaccinia virus	Many viruses can initiate the antitumor immune response Adenoviral, HSV genomes rarely integrate into the host genome HSV has many receptors for cell binding and entry The HSV genome is large and allows to incorporate large genes The HSV can be effectively controlled by common antiviral drugs Usually, HSV faster degrades cancer cells than adenoviruses Measles virus, Parvovirus, and Reovirus possess a natural tumor tropism Viruses could be targeted at specific cells using engineered proteins and peptides	Lack of efficacy is common Antivector immune response could limit the viral efficacy Safety may be an issue Many patients have antibodies against adenoviral vectors Adenoviruses are sequestrated by nontarget cells Manipulations with the genome sizes can decrease the stability of the virus Adenoviral genome is relatively small (36 kb) HSV genetic modification is difficult Some viruses integrate their genome into the host cells	(Bretscher and Marchini, 2019; Foreman et al., 2017; Gromeier and Nair, 2018; Hajeri et al., 2020; Msaouel et al., 2009; Saha et al., 2014; Watanabe and Goshima, 2018)
Other	DNA plasmid	Multiple antigen coverage No MHC restriction	Stability is a concern Relatively weak immunogenicity The delivery vehicle may be required	(Lopes et al., 2019)
	Aptamer	Many possible targets Easy and cheap large-scale synthesis Specificity Can be generated to cross the BBB	Limited stability and short half-life Delivery vehicle is required Antiformulation immune response	(Cesarini et al., 2020; Zhu and Chen, 2018)