HYDROPHILIC BIOMATERIALS: FROM CROSSLINKED AND SELF-ASSEMBLED HYDROGELS TO POLYMER-DRUG CONJUGATES AND DRUG-FREE MACROMOLECULAR THERAPEUTICS

Abstract

This “Magnum Opus” accentuates my lifelong belief that the future of science is in the interdisciplinary approach to hypotheses formulation and problem solving. Inspired by the invention of hydrogels and soft contact lenses by my mentors, my six decades of research have continuously proceeded from the synthesis of biocompatible hydrogels to the development of polymer-drug conjugates, then generation of drug-free macromolecular therapeutics (DFMT) and finally to multi-antigen T cell hybridizers (MATCH). This interdisciplinary journey was inspiring; the lifetime feeling that one is a beginner in some aspects of the research is a driving force that keeps the enthusiasm high. Also, I wanted to illustrate that systematic research in one wide area can be a life-time effort without the need to jump to areas that are temporarily en-vogue. In addition to generating general scientific knowledge, hydrogels from my laboratory have been transferred to the clinic, polymer-drug conjugates to clinical trials, and drug-free macromolecular systems have an excellent potential for personalizing patient therapies. There is a limit to life but no limit to imagination. I anticipate that systematic basic research will contribute to the expansion of our knowledge and create a foundation for the design of new paradigms based on the comprehension of mechanisms of physiological processes. The emerging novel platform technologies in biomaterial-based devices and implants as well as in personalized nanomedicines will ultimately impact clinical practice.

Graphical abstract

1. Hydrogels

Traditional hydrogels contain covalent crosslinks, e.g., they result from a copolymerization of a monovinyl compound with a divinylic crosslinker, such as copolymerization of 2-hydroxyethyl methacrylate (HEMA) with ethylene dimethacrylate (Figure 1A) [1]. Self-assembled hydrogels hold physical crosslinks created by biorecognition of complementary motifs, such as peptides or oligonucleotides [1–5]. These hydrogels [6] can be based on one type of macromolecule, e.g., genetically engineered triblock peptides [7,8] or occur as hybrids composed from a synthetic macromolecule and grafted peptide, oligonucleotide, or peptide nucleic acid biorecognition moieties [9–18].

Figure 1.

Hydrogels: from soft contact lenses and implants to self-assembled nanomaterials. (A) Schematic of chemically crosslinked hydrogels. (B) Preparation of hydrogels by crosslinking copolymerization of 2-hydroxyethyl methacrylate (HEMA) and ethylene dimethacrylate (EDMA) [19]. Soft contact lens from HEMA-based hydrogel [20]. (C) The pH dependent NaCl permeability through membranes from stimuli-sensitive HEMA-based hydrogels containing ionogenic groups [33]. (D) Hydrogels containing a triple mutant of adenylate kinase as crosslinks. Translation of substrate (ATP) recognition into conformation change of the peptide with concomitant macroscopic motion expressed as hydrogel deswelling. Inset: De-swelling ratios of Gels 1, 2, 3, 4 prepared with 100%, 50%, 25% of adenylate kinase triple mutant (AKtm) and 0%, 50%, 75%, and 100% dithiothreitol (control) in molar percentage of crosslinking agents, respectively [34]. (E) N-(2-hydroxypropyl)methacrylamide (HPMA) hydrogels containing enzymatically degradable oligopeptide crosslinks. Inset: Second period of degradation following dissolution of the hydrogel. The dependence of the molecular weight ratio (M_w/M_wo) for crosslinks of different structure [35]. (F) Poly(ethylene glycol) (PEG) hydrogels crosslinked with enzymatically degradable GFLG oligopeptide. Inset: Degradation of PEG2100 and PEG8800 based hydrogels by papain. Impact of PEG chain length on rate of degradation [36]. (G) Hydrogels for oral delivery of peptide and proteins contain aromatic azobonds in crosslinks to warrant degradation in the colon. Modification of crosslink structure impacts the rate of degradation [38–44]. (H) Biocompatibility of hydrogels. Different fate of homogeneous and heterogeneous (porous) HEMA hydrogels following subcutaneous implantation into pigs. Homogeneous hydrogels are surrounded by a capsule of fibrous tissue, whereas porous hydrogels, in addition to encapsulation, show substantial calcification [49,50]. (I) Translation of hydrogels into the clinic. HEMA-based hydrogels in rhinoplasty [54]. (J) Schematic of physically crosslinked hydrogels. (K) Self-assembly of genetically engineered triblock copolymers consisting from two terminal coiled-coil forming peptides attached to a central water-soluble random coil block [7]. (L) Hybrid N-(2-hydroxypropyl)methacrylamide (HPMA) copolymers containing coiled-coil forming grafts of peptide from the stalk region of the motor protein kinesin. Temperature induced melting of the peptide results in collapse of its structure and deswelling of the hydrogel [9]. (H) Formation of hydrogels by self-assembly of two HPMA copolymers grafted with complementary coiled-coil forming oppositely charged heptapeptad peptides CCE and CCK (P is the HPMA copolymer backbone) [11].

1.1. Chemical hydrogels: Hydrogels with covalent crosslinks

My research in hydrogels commenced at the Institute of Macromolecular Chemistry of the Czechoslovak Academy of Sciences in Prague when I joined as a graduate student in 1961. This was a place where the study of biomedical materials started in the 1950s. My mentor Drahoslav Lím invented hydrogels [19], first rationally designed biomedical polymers and the Head of the Institute, Otto Wichterle invented hydrogel-based soft contact lenses (Figure 1A,B) [20,21]. These innovations instigated research of biomedical materials worldwide and continue to be an inspiration until today. For history of hydrogel origination see part “Discovery” in ref. [22]; for the development of soft contact lenses, see ref. [23].

At first, the focus was on polymeric biomaterials from hydrophilic esters and N-substituted amides of (meth)acrylic acid. The kinetics of linear and crosslinking polymerizations of these monomers was assessed [24–30], and methods to control their mechanical and photodynamic properties were formed [31,32]. The mechanism of the crosslinking copolymerization revealed the important impact of the structure of the crosslinking agent (distance between two double bonds) on the degree of crosslinking, cyclization, and formation of pendent vinyl groups in the 3D polymerization product - hydrogels [24].

Stimuli-sensitive hydrogels change their properties (swelling, permeability, mechanical properties, etc.) upon a minor change in the environment (e.g., pH, ionic strength, electrical field, biorecognition). Early pH sensitive hydrogels were prepared by crosslinking copolymerization of HEMA with comonomers containing ionogenic groups. Hydrogel membranes containing ionogenic comonomers such as methacrylic acid, 2-(diethylamino)ethyl methacrylate, or both demonstrated solute permeability dependent on pH, crosslinking density and ionic strength (Figure 1C) [33].

An interesting stimulus that impacts hydrogel properties is based on the conformational change associated with protein-ligand recognition (Figure 1D). Enzyme adenylate kinase (Ake) catalyzes the phosphoryl transfer reaction Mg²⁺·ATP + AMP ↔ Mg²⁺·ADP + ADP. Following binding of a substrate, AKe undergoes a large conformational change. The bulky lid domain closes over the active site to shield it from water to avoid substrate hydrolysis and to facilitate the transfer of the phosphate group. To convert this conformational change into macroscopic motion a triple mutant AKtm (C77S, A55C, V169C) was engineered; it contains two attachment points (cysteines) at positions 55 and 169. The distance between C^α-atoms of the residues 55 and 169 decreases from 29.5 Å in the open conformation to 12.4 Å following substrate biorecognition (closed conformation). AKtm was inserted via C55 and C169 attachment points as a crosslink into N-(2-hydroxypropyl)methacrylamide (HPMA) copolymers or 4-arm poly(ethylene glycol) (PEG). The conformational change of the covalently attached AKtm translated into macroscopic motion of the hydrogel with concomitant hydrogel deswelling as shown in the inset of Figure 1D [34].

Enzymatically degradable hydrogels based on HPMA (Figure 1E) were synthesized cotaining oligopeptide sequences in crosslinks that matched the active site of chymotrypsin. Hydrogels containing in equilibrium swollen state < 95% water degraded by surface erosion, whereas highly swollen hydrogels (>95% water) degraded by bulk erosion as the enzyme could penetrate into the hydrogel interior. Time needed for total dissolution of the hydrogels was proportional to crosslinking density and the structure and length of the oligopeptide sequence [35]. Hydrogels synthesized by click reaction of 4-arm azido-terminated PEG and [alkyne]-GFLGK-[alkyne] and ([alkyne]-GFLG)₂K contained crosslinks degradable by cathepsin B and papain (Figure 1F) [36].

HPMA-based hydrogels containing oligopeptide sequences matching the active site of cathepsin B were used as a depo for macromolecules (FITC-labeled dextran of different mol. wt.) and daunomycin. The hydrogels degraded by a mixture of lysosomal enzymes (Tritosomes) isolated from rat liver or chymotrypsin. The rate of release was dependent of the crosslinking density and structure of oligopeptide crosslinks. Simultaneously the hydrogel was degraded and the cargo released [37].

Hydrogels for oral delivery of peptides that break down in the colon contain aromatic azobonds in the crosslinks (Figure 1G). If used as coatings for capsules containing peptides/proteins (insulin, calcitonin) they protect the cargo in the small intestine. In the colon, the aromatic azo bond containing crosslinks are degraded [38] and the peptide with the help of a penetration enhancer can be transferred into the circulation [39–44].

1.1.1. Biocompatibility of hydrogels.

Following biomaterial implantation into a living organism, the healing response involves inflammation, wound healing, and the foreign body reaction. The foreign body reaction to an implanted biomaterial follows a conventional succession of events [45–47]: nonspecific protein adsorption, and adhesion of cells (such as monocytes, leukocytes and platelets) to the biomaterial surface, leading to giant cell formation and cytokine release. Ultimately, the implant becomes encapsulated by a fibrous capsule [48].

Factors that impact the biocompatibility of hydrogels are chemical structure, physical structure (porosity), surface microarchitecture, and mechanical properties. Systematic studies of the biocompatibility of hydrogels based on crosslinked polyHEMA [48–50], poly(N-monosubstituted methacrylamides) [51], poly(N-monosubstituted acrylamides) [52], and poly(N,N-disubstituted acrylamides) [48,52], and their copolymers with ionogenic comonomers [53] revealed that all these structures are well tolerated after subcutaneous implantation into rats and pigs. No significant differences were observed with healing-in of hydrogels of different chemical compositions [48]. In all groups with different chemical structures the implants were encapsulated by a capsule of collagen fibrous tissue; no signs of proliferation were observed. However, significant differences were observed for hydrogels with different morphology.

Comparison of biocompatibility of homogeneous (transparent) hydrogels with microporous hydrogels containing interconnecting channels revealed considerable differences. The intensity of the response was greater with higher hydrogel porosity. An investigation of calcium deposits using von Kóssa staining revealed the dependence of the extent and localization of calcium deposits on hydrogel porosity. There was only infrequent calcification in the margin of the implant following implantation of homogeneous or microporous hydrogels; however, with an increase in porosity the position of calcium deposition moved from the margin of the implant toward its center (Figure 1H) [49,50].

1.1.2. Clinical application of hydrogels.

Biocompatibility results obtained from detailed studies in animal models [48–53] were validated in the clinics. Implantation of homogeneous HEMA-based hydrogels to treat nasal malformations (saddle, scoliosis, and nasal dorsum) showed excellent biocompatibility [54]. Evaluation of patients after 3–10 years revealed minor calcification at the margin of the implant (about 50% of patients evaluated) [55]. Apparently, with scalpel damaged surface (due to surgeons modifying the size of the hydrogel implants in the operation room) connective tissue accumulated and initiated calcium deposition. Minor calcium deposition did not affect the biocompatibility or the final cosmetic effect. None of the patients had any complains in connection with long-term implantation of the hydrogel (Figure 1I).

1.2. Physical Hydrogels: Hydrogels self-assembled from graft and block copolymers

We hypothesized that the self-assembly of soluble block and graft copolymers will be mediated and the structure of the hydrogel controlled by the properties of the biorecognition motif. Coiled-coil forming peptides were selected as the first biorecognition motif [56,57]. The biorecognition results in crosslinking of polymer chains and formation of hydrogels (3D networks) (Figure 1J).

1.2.1. Hydrogel assembly of recombinant triblock polypeptides.

Coiled-coil peptides were employed as biorecognition moieties creating physical crosslinks. The coiled-coil is a supercoil formed by two or more strands of α-helices [56]. We genetically engineered a series of recombinant triblock (ABA, ABC, CBA, CBC) polypeptides, consisting of two coiled-coil forming blocks (A or C) flanking a central water-soluble random coil segment (block B) [7,8,58]. In addition to structure, protein concentration was an important factor in the self-assembly of these block copolymers into hydrogels. Self-assembly occurs as an outcome of the balance between the oligomerization of the coiled-coil forming blocks and the swelling of the central water-soluble polyelectrolyte segment (Figure 1K). The gelation concentrations correlated well with the oligomerization state (dimers or tetramers) of the coiled-coil domains as determined by analytical ultracentrifugation. The self-assembly of hydrogels was responsive to the changes in temperature or pH, and was reversible after denaturation, indicating that the self-association, thermal stability, and pH-sensitivity of the coiled-coils facilitated the self-assembly of the hydrogels [7]. Hydrogel properties could be manipulated by changing the structure and the length of the coiled-coil blocks [8].

1.2.2. Hybrid hydrogels self-assembled from graft copolymers.

First coiled-coil based hybrid hydrogels were designed by self-assembly of HPMA copolymers grafted with genetically engineered protein domains. The coiled-coil consist of two (or more) α-helices that wind together forming a super helix. HPMA copolymers grafted with CC1 (a segment of the stalk region (amino acids 336 to 590) of the motor protein, kinesin) or CC2 (de novo designed pentaheptad coiled-coil sequence EK42 separated from His tag by a non-coiled region of 30 amino acids). The physical crosslinks were formed by self-assembly of the coiled-coil domains. A temperature-induced hydrogel collapse of the CC1 occurred that was related to the structural transition of the coiled-coil domains from an elongated helix to an unfolded state (Figure 1L). CC2 peptide possessed high thermal stability; consequently, no temperature dependent hydrogel collapse was observed. The results validated our hypothesis that the properties of a well-defined coiled-coil protein motif can be imposed onto a hybrid hydrogel containing synthetic polymer-based primary chains [9]. Tailoring material properties with genetically engineered proteins adds a new feature to the field of “smart” hydrogel-based biomaterials [4,9,18].

An improved design of self-assembling hybrid hydrogels involved self-assembly of HPMA copolymers grafted with complementary peptides forming antiparallel coiled-coil heterodimers [11–13]. The primary structure of coiled-coil forming peptides is characterized by a heptad repeating sequence (abcdefg), where a and d positions are usually hydrophobic amino acid residues, e and g are charged residues, and b, c, and f residues are chosen for their helical propensity or contribution to solubility. Two distinct oppositely charged pentaheptad peptides CCE (CYGG E VSALEKE VSALEKK NSALEKE VSALEKE VSALEK) and CCK (CYGG K VSALKEK VSALKEE VSANKEK VSALKEK VSALKE) were designed based on three stabilizing interactions: the hydrophobic interaction in the core, electrostatic interactions across the interface, and helical propensity effects. The peptides (synthesized by solid phase peptide synthesis) achieved an antiparallel heterodimeric conformation and thus generated a dimerization motif and acted as biorecognition moieties forming physical crosslinks [11]. Two graft copolymers, P-(CCE)_x (P is the HPMA copolymer backbone) and P-(CCK)_y, self-assembled into hybrid hydrogels in situ; the process was modulated by the formation of antiparallel heterodimeric coiled-coils (Figure 1M). The antiparallel arrangement decreased the steric hindrance of the polymer backbone on the formation of peptide dimers and enabled the “in-register” alignment of the peptide grafts. Equimolar mixtures of the graft copolymers, P-(CCE)_x/P-(CCK)_y, self-assembled into hydrogels in PBS (phosphate buffer) solution at neutral pH at concentrations as low as 0.1 wt.%. Denaturation of the coiled coils with guanidine hydrochloride (GdnHCl) led to a reversible disassembly of the hydrogels. Removal of GdnHCl by dialysis initiated coiled-coil refolding and hydrogel reassembly [11].

1.2.3. β-Sheet peptides as biorecognition moieties and physical crosslinks.

HPMA hybrid graft copolymers containing β-sheet peptides as biorecognition/crosslinking moieties were made by attachment of N-terminus CGG modified β-sheet peptide (CGGTTRFTWTFTTT) to HPMA copolymer precursor, which contained pendent maleimide groups. Circular dichroism (CD) spectra demonstrated that the strong tendency of the peptide to self-assemble into β-sheets was preserved in the graft copolymers. In addition, β-sheet response to temperature and pH variations decreased due to PHPMA shielding effect [14]. Finally, we investigated the capacity of a hybrid hydrogel self-assembled from HPMA copolymers and complementary β-sheet grafts (TTRFTWTFTTT-NH₂ and TTEFTWTFETT-NH₂) to act as scaffolds for bone tissue engineering [15]. The hydrogels displayed anisotropic porosity, were biocompatible with preosteoblast cells and presented surfaces characterized by epitaxy that favored template-driven mineralization of hydroxyapatite. Energy dispersive spectroscopy analysis of the surface of the hydrogels incubated in simulated body fluid revealed that the Ca/P (phosphorus) ratio was 1.7 [15] close to the value 1.67 found in hydroxyapatite.

1.2.4. Peptide nucleic acids/DNA complexation mediated crosslink formation.

An alternate design of oligonucleotide containing hybrid hydrogels employed peptide nucleic acid (PNA)/DNA complexation as crosslinking origination. HPMA copolymers grafted with multiple oligomers of PNA were crosslinked upon addition of the linker DNA. Two types of complex formation were studied: a) PNA/DNA duplex hybridization via Watson-Crick base pairing; b) via P-form oligonucleotide triple helix formation composed of two polypyrimidine PNAs and one polypurine DNA; complexation was achieved via combination of Watson–Crick and Hoogsteen base interactions [16].

1.2.5. Potential of hybrid biomaterials.

Conjugation of peptide domains to synthetic polymers may lead to novel materials with properties superior to those of individual components. Compared to synthetic polymers, proteins and protein modules have well-defined and homogeneous structures, consistent mechanical properties, and cooperative folding/unfolding transitions. The peptide domain may impose a level of control over the structure formation at the nanometer level; the synthetic part may contribute to the biocompatibility of the hybrid material (e.g., decreasing the immunogenicity of the protein component). The synergistic combination of two types of structures may produce new materials that possess unprecedented levels of structural organization and novel properties.

Lessons learned from (traditional) hydrogel research were applied in the design of water-soluble polymers as drug carriers:

Crosslinking copolymerization of methacrylic acid monoesters of glycol [24,25], diglycol [24,26] and triglycol [27]) generated hydrogels with excellent translational potential as demonstrated by their clinical applications as implants [54] and soft contact lenses [20]. However, problems with the preparation of pure monomers without minute amounts of diesters created by transesterification directed us to concentrate on N-substituted amides of (meth)acrylic acid. In particular, we have chosen N-substituted methacrylamides as our target because the α-carbon substitution and the N-substituted amide bond guaranteed hydrolytic stability of the side-chains [48,59]. We synthesized various compounds of this family trying to discover a crystalline monomer for straightforward purification and reproducible synthesis. The first crystalline N-substituted methacrylamide we achieved to synthesize, N-(2-hydroxypropyl)methacrylamide (HPMA), was selected for further development. This warranted a reproducible synthesis of water-soluble polymeric drug carriers [60–62]).

2. Water-soluble polymers

After establishing the biocompatibility of hydrogels based on esters and substituted amides of (meth)acrylic acid, we concentrated on water-soluble polymers. First, we deemed the biocompatibility more challenging – after intravenous administration the polymer can reach numerous sites in the organism. Second, biocompatible soluble polymers could be used for protein modification and as drug carriers. As mentioned above, a new hydrophilic polymer, poly[N-(2-hydroxypropyl)methacrylamide] (PHPMA) [60–63] was chosen as a candidate for a soluble polymeric drug carrier as well as protein and nanoparticle surface modification.

2.1. Modification of proteins and vesicular carriers with PHPMA

PHPMA can replace poly(ethylene glycol) (PEG) for surface modification of enzymes or vesicular carriers. Either semitelechelic PHPMA (ST-PHPMA) or HPMA copolymers containing multiple reactive side-chains can be used for nanoparticle [64], protein [65], or virus [66] modification (Figure 2A). Modification of nanospheres with ST-PHPMA resulted in decreased protein adsorption in vitro, increased intravascular half-life, as well as decreased accumulation in the liver after intravenous administration into rats. The higher the molecular weight of ST-PHPMA, the more pronounced the changes in these properties [64]. Similarly, carboxyl and amino group modification of chymotrypsin with ST-PHPMA-CONHNH₂ and ST-PHPMA-COOSu (N-hydroxysuccinimide ester) produced conjugates [67] with comparable properties to PEG-modified chymotrypsin [68].

Figure 2.

Water-soluble polymers for modification of proteins and surfaces and as drug carriers. (A) Design of water-soluble polymers for modification of proteins and vesicular carriers: HPMA copolymers with multiple reactive side chains and semitelechelic HPMA copolymers suitable for one-point attachment. (B) First generation HPMA copolymer-drug conjugates. Structure and characterization of (nondegradable) HPMA copolymer-drug conjugates [133]. (Ca) Internalization of HPMA copolymer-drug conjugates into cells. (Cb) Internalization of HPMA copolymer-doxorubicin (P-GFLG-DOX) conjugates in OVCAR-3 cells mediated by clathrin-coated and caveolin-coated organelles as illustrated by confocal fluorescence microscopy [145]. (Da) Structure of HPMA copolymer – 9-aminocamptothecin (9-AC) conjugate (P-9-AC). 9-AC was bound to the HPMA copolymer carrier via a spacer composed from an aromatic azo group and 1,6-elimination spacer. (Db) Survival curves of nu/nu nude mice bearing HT29 human colon carcinoma xenografts, treated with 9-AC and P-9-AC at a dose 3 mg/kg of (-AC or 9-AC equivalent [147]. (Ea) Structure of HPMA copolymer-DOX conjugate. (Eb) Repeated exposure of A2780 human ovarian carcinoma cells to 10 or 20% of IC50 dose of free DOX and P-GFLG-DOX. Repeated exposure to DOX resulted in a formation of a new phenotype – resistant cell expressing the MDR gene encoding P-glycoprotein. On the contrary, following exposure to P-GFLG-DOX no significant changes were found [159]. (Ec) Anticancer activity of free DOX and P-GFLG-DOX in solid tumor mice models of sensitive (A2780) and DOX-resistant (A2780/AD) human ovarian carcinoma xenografts in female athymic nu/nu mice [144].

First HPMA copolymer modified protein was prepared by the reaction of the copolymer of HPMA with N-methacryloylglycylglycine p-nitrophenyl ester with insulin. The HPMA copolymer – insulin conjugate exhibited a slower onset and a slight prolongation of hypoglycemic effect in rats when compared to free insulin [65]. Similar modification was achieved with chymotrypsin and the impact of HPMA copolymer attachment on the kinetics of (modified) chymotrypsin catalyzed substrate cleavage was evaluated [69].

Different chemistry was used for the modification of cobra venom acetylcholinesterase [70]. The secondary OH groups of PHPMA (Mw 25–30 kDa) were activated with 4-nitrophenyl chloroformate in dimethylformamide followed by attachment to acetylcholinesterase in borate buffer. The PHPMA-modified acetylcholinesterase demonstrated a 70-fold prolongation of enzyme activity in blood after intravenous injection into mice when compared to unmodified enzyme. The thermoinactivation rate of the PHPMA-acetylcholinesterase conjugate was 74 times smaller than that of native enzyme [70].

The advantage of PHPMA modification of proteins or vesicular carriers when compared to PEG modification is the absence of antibody formation and of accelerated blood clearance [71].

2.2. First-generation (nondegradable) HPMA copolymer-drug conjugates

The biocompatibility of HPMA copolymers as drug carriers was appraised by tests designed for the evaluation of blood plasma expanders [72–75]. After validating the biocompatibility of HPMA copolymers we started their evaluation as drug carriers (Figure 2B).

Macromolecules are internalized by cells via endocytosis and ultimately localize in the (enzyme rich) lysosomal compartment (Figure 2Ca,b) [76]. To model intracellular drug release, we performed fundamental studies of the rates of enzymatic degradation of polymer-bound oligopeptide side-chains [69,77–86]. Using the S-P interaction scheme of Schechter and Berger [87] and data of X-ray diffraction studies of enzyme-inhibitor complexes [88], we evaluated relationship between the detailed structure of the oligopeptide sequences and the rate of enzymatically catalyzed release of a leaving group (p-nitroaniline). First, model enzymes, chymotrypsin [69,77], trypsin [78], and papain [79] were evaluated, followed by intracellular (lysosomal) enzymes. A study with a mixture of lysosomal enzymes revealed that cysteine (thiol) proteinases were responsible for the cleavage of these polymer-drug conjugates [80]. Subsequently, polymers were designed to match the specificity of individual cysteine proteinases: cathepsins B [84], L, H, and artificial mixtures of lysosomal enzymes [81]. The stability of the oligopeptide side-chains in blood plasma and serum was verified [82]. Based on these results it was possible to control the degradability of side-chains in HPMA copolymers by a particular enzyme, as well as in the in vivo system [83]. Based on these studies the sequence GFLG was selected for the design of polymer-drug conjugates [84]. Notably, very successful FDA approved ENHERTU^® (Transtuzumab deruxtecan), an antibody drug conjugate, contains a tetrapeptide spacer GGFG, validating the concept of cathepsin B sensitive peptide spacers (Daiichi Sankyo, Tokyo, Japan, [89]).

The final design of polymer-drug conjugates was based on additional basic studies:

• Relationship between polymer structure and rates of pinocytic uptake of HPMA-based copolymers into cells was established [90–92].

• Polymers capable of specific intralysosomal release of chemically bound therapeutic agents were designed and evaluated [82,93–96].

• Fundamental studies were commenced of tissue localization of polymeric carriers bearing ligands for receptors expressed on the surface of specific target cells [97–104].

• Immunogenicity of oligopeptide spacers and HPMA copolymer-drug conjugates was assessed [105–108].

• Basic studies of the relationship between the structure of HPMA copolymer conjugates on one hand and their conformation in aqueous environment and biorecognition and rates of enzymatically catalyzed drug release on the other hand were carried out [109,110].

The results of these studies suggested the composition of polymer – drug conjugates (Figure 2B). Their structure was composed from the HPMA copolymer backbone, drug bound via a lysosomally cleavable sequence (GFLG) designed to fit into the active site of cathepsin B [84] and potentially a targeting moiety - antibody [97,101,111–115], antibody fragment [116,117], peptide [118–121], saccharide [98,99,122], lectin [123] or alendronate [124,125] (Figure 2B).

Colon-specific targeting requires specific design. For targeting to colon interaction antigens with lectins attached to polymer carriers is one of the options. The Thomsen-Friedenreich (TF) antigen is a core carbohydrate structure defined by a D-galactose-ß1,3-N-acetyl-D-galactosamine glycoprotein sequence. TF antigen has been identified in human carcinomas and inflammatory bowel disease. HPMA copolymer-peanut agglutinin (PNA) conjugates bind to TF antigen and are suitable for targeted colon or esophagus delivery [123].

A special approach is subcellular targeting where the distribution into organelles is manipulated. It is known that subcellular location influences the efficacy of drugs. Mitochondrial targeting can be facilitated by utilizing the negative mitochondrial potential and positively charged triphenylphosphonium groups as targeting moieties [126,127]. Triphenylphosphonium containing HPMA copolymer-bound Mce₆ augmented cytotoxicity toward human ovarian SKOV3 cells as compared to non-targeted HPMA copolymer-Mce₆ conjugates [127].

Basic studies on the passive [126] and chaperoned [128] transport through the nucleopore complex assist the design of conjugates for nuclear targeting. Nuclear targeting can be negotiated by steroid hormone receptors. For example, when glucocorticoid receptor (GR) binds to its steroid ligand, it dimerizes and transfers into the nucleus. Rebuffat et al. enhanced the transport of transfected DNA into the nucleus of GR positive cells using this approach [129]. This methodology can be also utilized for HPMA copolymer conjugates. Incubation of 1471.1 cells (transfected with GR labeled with green fluorescent protein, GFP) with HPMA copolymer containing mesochlorin e₆ (Mce₆)-cortisol bound to the polymer via a lysosomally degradable GFLG spacer attained the localization of the GR-GFP receptor in the nucleus. Apparently, the copolymer was internalized by endocytosis, Mce₆-cortisol released by enzyme catalyzed hydrolysis and translocated to the cytoplasm where it bound to GR (mediated by cortisol). This initiated dimerization of the GR – cortisol-Mce₆ complex and its nuclear transport [130].

Studies of HPMA copolymer-drug conjugates were reviewed numerous times [74,131–138]. The first generation of HPMA copolymer-drug conjugates with nondegradable backbone was active on numerous animal models of cancer [139–148]. The drugs (doxorubicin, mesochlorin e₆, docetaxel, 9-aminocamptothecin) were attached to the polymer backbone via lysosomally degradable oligopeptide spacers [75], hydrolytically cleavable (e.g., hydrazone) bonds [149] or a combination of lysosomally (GFLG) and matrix metalloprotease-2 (PLGLAG) degradable spacers [150]. For the treatment of colon cancer a combination of aromatic azo bond with a 1,6-elimination group has been used (Figure 2Da,b) [147,148,151–154].

The advantages of polymer-bound drugs (when compared to low-molecular weight drugs) are: a) active uptake by fluid-phase pinocytosis (non-targeted polymer-bound drug) or receptor-mediated endocytosis (targeted polymer-bound drug) [155], b) increased passive accumulation of the drug at the tumor site by the enhanced permeability and retention (EPR) effect [156,157], c) increased active accumulation of the drug at the tumor site by targeting [97], d) extended circulation in the bloodstream, e) decreased non-specific toxicity of the conjugated drugs [158], f) potential to overcome multidrug resistance [144,159,160], g) decreased immunogenicity of the targeting moiety, h) immunoprotecting and immunomobilizing activities [161], and i) modulation of the cell signaling and apoptotic pathways [144,159,160,162].

An important aspect of HPMA copolymer-drug conjugates is their capability to overcome efflux pump-mediated multidrug resistance. Polymer-drug conjugates are internalized in membrane limited organelles and thus are not recognized by P-glycoprotein. Repeated exposure of sensitive human ovarian carcinoma A2780 cells to HPMA copolymer-DOX conjugate does not induce multidrug resistance in contrast to exposure of cells to free DOX which resulted in a new phenotype expressing the MDR gene (Figure 2Eb) [159]. The drug is released from the carrier in the lysosomal compartment located in the perinuclear region far from the location of the efflux protein embedded in the plasma membrane. Consequently, the activity of HPMA copolymer-DOX conjugate (Figure 2Ea) in experimental resistant tumors is superior to the treatment with free drug (Figure 2Ec) [144].

Combinations of polymer-bound drugs have been evaluated for several decades [141]. The factors that were studied include combination of two polymer conjugates vs. polymer conjugate that contains both drugs covalently attached; sequence of administration (which drug first, or simultaneous administration), dosing in synergistic concentrations. Another approach is the combination of polymer-drug conjugates (chemotherapy) with physical stimuli – focused ultrasound, photothermal, and photodynamic therapy (activation by light).

An interesting combination treatment is the combination of two conjugates that target different cell subpopulations, cancer stem cells (CSC) and differentiated cells. A problem to eradicate CSCs in the clinics is in their inherent resistance to established chemo- or radiotherapies, which target differentiated cancer cells (Figure 3A). We hypothesized that prostate cancer stem cells could be downregulated by blocking the hedgehog pathway and the differentiated cells should respond to the action of clinically used docetaxel (DTX). Consequently, the combination of two HPMA copolymer conjugates, one containing cyclopamine (CYP), a hedgehog pathway inhibitor, the second containing DTX was evaluated (Figure 3B). Notably, the HPMA copolymer-cyclopamine conjugate (P-CYP), like free cyclopamine, showed selective inhibitory effect on an enriched subset of CD133+ cells within a prostate cancer epithelial cell line RC-92a/hTERT, whereas the HPMA copolymer-DTX (P-DTX) conjugate showed preferential cytotoxicity to bulk cancer cells [163,164].

Figure 3.

Combination therapy with HPMA copolymer-drug conjugates. (A) Principle of targeting cancer stem cells and differentiated cells in cancer [134]. (B) Structure of HPMA copolymer conjugates: cyclopamine conjugate targeting stem cells and docetaxel conjugate targeting differentiated prostate cancer cells. (C) CD133 expression (stem cell marker) and cytotoxicity in PC-3 cells following CYP, DTX, P-CYP and P-DTX treatments. Black columns: CD133 expression level (%); gray columns: cell viability (%). *, p<0.05 [164]. (D) Tumor growth inhibition by P-CYP (CYP equivalent 40 mg/kg), P-DTX (DTX equivalent 10 mg/kg), and combination of P-DTX (DTX equivalent 10 mg/kg) and P-CYP (CYP equivalent 40 mg/kg) in PC-3 prostate tumor-bearing nude mice. *statistically significant difference between P-DTX or P-CYP single treatments and combination treatment (P-DTX + P-CYP) group (p<0.05) [164]. (E,F) Combination chemotherapy and photodynamic therapy of human ovarian carcinoma OVCAR-3 xenografts in nude mice. Synthesis of OV-TL-16 antibody targeted HPMA copolymer-DOX and -Mce₆ conjugates (E) and treatment efficacy of targeted and non-targeted conjugates (F) [112]. (Ga) Design of a bone-targeting (via D-Asp octapeptide) HPMA copolymer-prostaglandin E1 (PGE1) conjugate. PGE1 is bound to the polymer backbone via a cathepsin K sensitive spacer (Gly-Gly-Pro-Nle) and a self-eliminating 4-aminobenzyl alcohol structure [166]. (Gb,c) Efficacy of the conjugate to promote bone formation in aged, estrogen-deficient rats. Single iv injection of 10 mg of the conjugate (0.6 mg PGE1) was administered. (b) Bone formation rate (μm² / μm / day) for control (untreated) and treated (10 mg) rats. (c) Four weeks after the administration of a single injection of the P-Asp₈-FITC-PGE1 conjugate new bone formation was indicated by tetracycline label (yellow, double arrows) [167].

In PC-3 prostate tumor-bearing nude mice, the P-CYP conjugate preferentially targeted the cancer stem/progenitor cell population at doses that do not reduce bulk tumor cells. P-CYP inhibited the self-renew ability and induced apoptosis of CD133+ prostate cancer stem/progenitor cells (Figure 3A,B,C) [164]. The combination of P-CYP and an effective bulk cancer cell-killing conjugate P-DTX showed more effective long-term tumor growth inhibition in PC-3 prostate cancer xenografts in mice than single agent therapies (Figure 3D) [164]. Similarly, a combination of a DTX conjugate with a conjugate containing GDC-0980, a dual PI3K/mTOR inhibitor, was effective in the PC-3 prostate cancer model [165].

Combination of photodynamic therapy (PDT) with chemotherapy using polymer-bound drugs is efficient in cancer treatment. Photosensitizers (PSs) are not active in dark, but when irradiated with a characteristic wavelength the PS gets into its first excited singlet state. Intersystem crossing converts the singlet state PS into its triplet state. Exchange of energy with ground state triplet state oxygen produces reactive singlet oxygen (¹O₂). PSs are active both when covalently bound to the polymer carrier or when released inside the cell. However, the quantum yield of singlet oxygen formation is substantially higher when released from the carrier [141]. Thus, enzymatically degradable spacers are favored for PS attachment and activation with near infrared wavelengths that are more efficient due to deeper penetration of light into the tissue. Combination of HPMA copolymer-DOX and HPMA copolymer-Mce₆ conjugates was more active in the treatment of Neuro 2A neuroblastoma in A/J mice [141] and in treatment of human ovarian OVCAR-3 xenografts in female athymic mice than individual therapies [142]. OV-TL-16 antibody targeted HPMA copolymer-DOX and -Mce₆ conjugates (Figure 3E) were more active in the OVCAR-3 model than non-targeted conjugates (Figure 3F) [112].

2.2.1. Clinical trials.

Several conjugates were evaluated in clinical trials [158,168–173]; the results clearly demonstrated a significant decrease of non-specific toxicity – the MTD (maximum tolerated dose) of PK1 (HPMA copolymer-DOX conjugate) was 320 mg/m² of DOX equivalent (MTD of free DOX is 60–80 mg/m²) and no congestive heart failure was observed despite doses up to 1680 mg/m² [158]. In phase II of PK1 (dose 280 mg/m²) the enhancement of PK1 efficacy [169] when compared to free DOX was smaller than in animal models [101,112,139–146]. Similar results were obtained in other clinical trials [174]. The main reason was the relatively short intravascular half-life of conjugates with a nondegradable polymer backbone with a molecular weight below the renal threshold. For detailed analysis of first-generation conjugates and the rationale for the design of second-generation conjugates, see ref. [137] for ideas how to improve the translation of nanomedicines into the clinics, see [175–177].

2.2.2. HPMA copolymer conjugates – beyond cancer.

HPMA copolymer-drug conjugates are fitting for the treatment of non-cancerous diseases, such as musculoskeletal and infectious diseases. Bone and bone diseases can be targeted by nanomedicines containing acidic oligopeptides or bisphosphonates [178]. The location of binding to bone depends on the crystallinity of hydroxyapatite. D-aspartic acid octapeptide (D-Asp₈) would recognize resorption sites in skeletal tissues, whereas alendronate would direct the bound cargo to both bone resorption and formation sites. This selectivity relates to the strength of the binding force between targeting moiety and bone as determined by atomic force microscopy [179].

HPMA copolymer – prostaglandin E₁ (PGE₁) conjugates targeted by D-Asp₈ were designed for the treatment of osteoporosis. The PGE₁ was bound to the carrier via a self-immolating spacer containing a cathepsin K sensitive Gly-Gly-Pro-Nle tetrapeptide (Figure 3Ga) [166]. When the conjugate attaches to bone, the anabolic agent PGE₁ is released at sites of higher osteoclast activity by cathepsin K mediated cleavage. The resulting bone formation is facilitated by PGE1 mediated activation of EP receptors on bone cell surfaces [180]. Pharmacokinetic and biodistribution studies as well as evaluation of bone formation in ovariectomized rats (Figure 3Gb,c) demonstrated efficacy of the PGE1 conjugates [167,181].

Lessons learned from clinical trials with nondegradable (first generation) HPMA copolymer-drug conjugates were the basis for the design of backbone degradable long circulating HPMA carriers

We hypothesized that the reason for lower efficacy in humans when compared to animal models was the short blood circulation time of the low molecular weight first-generation HPMA copolymer-drug conjugates. To ensure biocompatibility of nondegradable polymer carriers, molecular weights below the renal threshold were used. The intravascular half-life was sufficient in animal models where the tumor is up to 10% of body weight. In humans, a prolonged concentration gradient between vasculature and tumor is needed to ensure abundant extravasation of the conjugate into the tumor and efficacy [137,138]. This hypothesis was supported by the evaluation of branched HPMA copolymer conjugates of varying molecular weight containing a degradable GFLGOGLFG degradable sequence (Figure 4A,B,C) [156]. This premise initiated the design of backbone degradable, 2^nd generation conjugates described in the next paragraph.

Figure 4.

Second generation backbone degradable, long-circulating HPMA copolymer-drug conjugates. (A) Chemical structure of soluble branched long-circulating P-DOX conjugate containing glycylphenylalanylleucylglycine side-chains and N²,N⁵-bis(N-methacryloylglycylphenylalanylleucylglycyl)ornithine crosslinker. (B) Growth inhibition of s.c. human ovarian OVCAR-3 carcinoma xenografts in nu/nu mice by long-circulating P-DOX conjugates of varying Mw. The mice received i.v. injection of 2.2 mg/kg DOX equivalent (n=6). (C) Concentration of DOX in OVCAR-3 carcinoma xenografts in nu/nu mice after i.v. bolus of free DOX or P-DOX of different Mw [156]. (D) Design and synthesis of 2^nd generation backbone degradable HPMA copolymer-DOX conjugates by RAFT polymerization mediated by a new bifunctional chain transfer agent (CTA) that contains an enzymatically degradable oligopeptide sequence (GFLGKGLFG) flanked by two dithiobenzoate groups. Post-polymerization chain-end modification and chain-extension via click reaction converts diblock copolymers into tetrablocks and hexablocks [182]. (E) Degradation of diblock HPMA copolymer by cathepsin B or papain results in two macromolecules of half size. This indicates that the monomers incorporate in the two dithiobenzoate groups of CTA with the same rate [182]. (F) Impact of Mw of HPMA copolymer-gemcitabine (GEM) conjugates on activity toward A2780 human ovarian carcinoma xenografts in nude mice. Comparison of free GEM, first generation (nondegradable) conjugate (P-GEM; ~50 kDa) and second-generation conjugates: diblock (2P-GEM; ~100 kDa), tetrablock (mP-GEM₂₀₀; ~200 kDa), hexablock (mP-GEM₃₀₀; ~300 kDa), and combination therapy of 2P-GEM + 2P-PTX (paclitaxel) [183]. (G) Impact of Mw of HPMA copolymer-PTX conjugates on activity toward A2780 human ovarian carcinoma xenografts. Comparison of free PTX, first generation (nondegradable) conjugate (P-PTX; ~30 kDa) and second-generation conjugates: diblock (2P-PTX; ~100 kDa), tetrablock (mP-PTX₂₀₀; ~200 kDa), and combination therapy of 2P-GEM + 2P-PTX [183]. (H) Comparison of first generation HPMA copolymer-epirubicin conjugate (P-EPI; ~40 kDa) with second-generation backbone degradable HPMA copolymer-EPI conjugate (2P-EPI; ~100 kDa) in the treatment of A2780 human ovarian carcinoma xenografts in nude mice. P-EPI and 2P-EPI were administered via tail vein with dose of 5 mg/kg EPI equivalent on days 0, 4, and 8 [184]. (I,J,K) Sequential combination therapy of A2780 human ovarian carcinoma xenografts in nude mice with backbone degradable HPMA copolymer-PTX and -GEM conjugates. (I) Tumor growth in mice treated with different formulation combinations (n=5). Animals were administered single dose of PTX/PTX conjugates on day 0 (20 mg/kg) and three doses of GEM/GEM conjugates on days 1, 7, and 14 (5 mg/kg). (J) SPECT/CT images of mice bearing subcutaneous A2780 human ovarian carcinoma in right flank after i.v. injection of ¹²⁵I-labeled conjugates (2P-PTX, 2P-GEM). The representative images were acquired 24 h, 48 h, and 7 d after administration of conjugates. T, tumor. (K) Dual labeled ¹²⁵I-Tyr-P-DTPA-¹¹¹In conjugate. Structure, terminal half-life (Table) and tumor uptake of ¹¹¹In-DTPA and ¹²⁵I-Tyr-P in mice bearing subcutaneous A2780 tumor 48 h after injection of 2^nd generation ¹²⁵I-Tyr-P-DTPA-¹¹¹In or 1^st generation model conjugates. Data obtained using the radioactivity count method plotted as percentage of injected dose per gram of tissue (%ID/g) [185].

2.3. Second generation (backbone degradable, long circulating) HPMA copolymer-drug conjugates

Advances in polymer chemistry permitted to design and synthesize new, backbone degradable, long-circulating HPMA copolymers with narrow molecular weight distribution. Corresponding polymer-drug conjugates contain an enzymatically degradable sequence both in the main chain and in the spacer between polymer backbone and drug. We designed a new RAFT (reversible addition-fragmentation chain transfer) polymerization chain transfer agent, Peptide2CTA, (N^α,N^ε-bis(4-cyano-4-(phenylcarbonothioylthio)pentanoylglycylphenylalanylleucylglycyl)-lysine); it is composed from an oligopeptide (GFLGKGLFG) degradable sequence flanked by two dithiobenzoate groups. During RAFT copolymerization, HPMA and comonomers insert at both dithiobenzoate groups with the same rate, producing a diblock copolymer with a very narrow molecular weight distribution in one scalable synthetic step [182]. Click reactions may be used to convert the diblock copolymer (Mw~100 kDa) into tetrablock (Mw~200 kDa), and hexablock (Mw~300 kDa) HPMA copolymers (Figure 4D,E). Evaluation of the relationship between the molecular weight of HPMA copolymer-gemcitabine and -paclitaxel conjugates on human ovarian carcinoma A2780 xenografts in mice revealed that diblock conjugates were more efficient in tumor growth inhibition than tetrablock and hexablock conjugates (Figure 4F,G) [183].

Numerous data have validated the hypothesis on circulation time importance when comparing the efficacy of second-generation conjugates with first-generation conjugates. These are results of markedly improved pharmacokinetics and tumor accumulation. Comparison of HPMA copolymer-epirubicin (EPI) conjugates (non-degradable ~40,000 Da P-EPI vs. degradable diblock ~100,000 Da 2P-EPI) on human ovarian carcinoma A2780 xenografts (5 mg/kg EPI equivalent doses administered i.v. on days 0, 4, and 8) showed that total tumor regression was detected for 100 days in all five mice treated with 2P-EPI, whereas four mice (out of five) treated with P-EPI commenced fast tumor regrowth from day 35 (Figure 4H) [184]. Contrasted with the 1^st generation conjugate (P-EPI, Mw~40kDa), 2P-EPI (Mw~100 kDa) demonstrated staggeringly enhanced pharmacokinetics such as four-fold increase in terminal half-life (33.22±3.18 h for 2P-EPI vs 7.55±3.18 h for P-EPI) [184], which is primarily due to the increased molecular weight of the polymer carrier. Backbone degradable HPMA copolymer-EPI conjugate 2P-EPI (also called KT-1), was selected by Nanotechnology Characterization Laboratory (NCL) for evaluation with respect to physicochemical characterization, in vitro cytotoxicity, immunotoxicity, and cancer treatment efficacy. 2P-EPI displayed outstanding antitumor activities over free EPI in several preclinical models including colon, lung, pancreatic, and breast cancers (unpublished). Comparable data were obtained with DOX conjugates in a human ovarian carcinoma A2780 animal model [186] and paclitaxel (PTX) conjugates in orthotopic A2780 ovarian tumors in nude mice [187].

Cancer treatment in the clinics frequently uses combination of drugs that hold different mechanisms of action. One such combination is gemcitabine (GEM) + PTX. Synergistic doses were selected and mice bearing A2780 human ovarian carcinoma xenografts were treated with a combination of 2P-GEM and 2P-PTX. As controls combination of 1^st generation conjugates (P-GEM + P-PTX) and combination of free drugs (GEM + PTX) were employed. The data demonstrated that 2^nd generation backbone degradable conjugates acquired prolonged blood circulation time, enhanced tumor accumulation, and enhanced anti-tumor efficacy as compared to 1^st generation low-Mw conjugates and free drugs (Figure 4 I,J,K). The new 2^nd generation conjugates were degradable in vivo and lacked noticeable systemic toxicity [185].

Interestingly, second generation, long-circulating, backbone degradable HPMA copolymer carriers of PGE1 were more efficient in stimulating bone growth in ovariectomized rats than first generation nondegradable conjugates [181].

2.3.1. Combination chemotherapy and immunotherapy with macromolecular therapeutics – design of multivalent polymer-peptide PD-L1 antagonists.

The blockade of the interaction of programmed cell death ligand 1 (PD-L1 on cancer cells) with PD-1 on T cells employing anti-PD-1 or anti-PD-L1 antibodies is an effective cancer treatment approach [188]. Effective checkpoint inhibitors include antibodies, small molecules, peptides, and macrocycles [189]. However, in various tumors the immune checkpoint blockade (ICB) is not efficient due to hostile tumor microenvironment. One way to boost the efficacy of ICB is to augment effector T cell recruitment in the tumor microenvironment by immunogenic cell death (ICD). Cells that undergo ICD are characterized by translocation of calreticulin to the cell surface and post-apoptotic release of high-mobility group box 1 (HMGB1) protein and ATP into the extracellular environment [190]. Backbone degradable HPMA copolymer-EPI conjugate (2P-EPI; KT-1) efficiently induces ICD in BALB/c mice bearing 4T1 tumors (Figure 5A,B) [191].

Figure 5.

Combination chemotherapy and immunotherapy with HPMA copolymer conjugates. (A) EPI and 2P-EPI (KT-1) induce immunogenic cell death of 4T1 breast cancer cells in Balb/c mice. (Aa) Dosing schedule of KT-1 and EPI in Balb/c mice bearing 4T1 tumor. (Ab) Time dependent calreticulin expression on the surfaces of 4T1 cells in Balb/c mice treated with KT-1 or EPI. (Ac) Time dependent intra-tumoral release of high mobility group box 1 protein. (Ad) Activation of dendritic cells within 4T1 tumors in Balb/c mice. (Ae) CD8+ T cell infiltration into 4T1 tumors in Balb/c mice [191]. (B) Mean tumor growth curves following treatment of Balb/c mice bearing 4T1 breast cancer (left panel) and adaptive enrichment of PD-L1 expression in 4T1 tumors in Balb/c mice (right panel). (C) Schematic illustration of polymer-enhanced combination of chemotherapy and immunotherapy [191]. (D) Schematic of PD-1/PD-L1 interaction blockade. (E) Structure of the Multivalent Polymer-Peptide PD-L1 Antagonist (MPPA) [191]. (F) Crosslinking of PD-L1 with MPPA changes its subcellular fate; it biases the recycling to the lysosomal route and degradation. (G) Time-dependent recovery of surface PD-L1 after treatments with α-PD-L1, PPA (PD-L1 binding peptide; NYSKPTDRQYHF [193]), or MPPA [191]. (H) Monotherapy and combination therapy of Balb/c mice bearing 4T1 breast carcinoma. Combination of KT-1 and MPPA demonstrated remarkable tumor regression and achieved 100% long-term survivors [191]. (I) Therapeutic efficacy (tumor growth and survival rate) of KT-1+MPPA combination treatment in advanced-stage MMTV-PyMT mouse model of breast cancer. Monotreatment of early and advanced tumor were used as controls [194].

We designed a new, effective PD-1/PD-L1 inhibitor based on multivalency and receptor crosslinking - multivalent polymer-peptide PD-L1 antagonist (MPPA) [191]. Receptor crosslinking manipulates subcellular trafficking of receptor-bound ligands to the endosomal/lysosomal route [192]. Thus, we hypothesized that PD-L1 crosslinking will change its subcellular fate and direct PD-L1 into the lysosomes for degradation. Indeed, a PD-L1 antagonist peptide (PPA-1, sequence NYSKPTDRQYHF [193]) attached in multiple copies to HPMA copolymer hyper-crosslinked the cell surface PD-L1 receptors and efficiently trafficked the complex to the lysosomes for degradation resulting in decreased PD-L1 expression at cell surface and enhanced antitumor activity (Figure 5C–G) [191]. Combination chemotherapy and immunotherapy with 2P-EPI and MPPA was effective in the treatment of triple negative 4T1 breast cancer (10/10 complete tumor regressions) (Figure 5H). The hypothesis on the impact of receptor crosslinking on the efficacy of immunotherapy was validated by decreased PD-L1 expression following treatment and by excellent tumor growth inhibition effect of the combination of KT-1 with MPPA [191].

This combination (KT-1 + MPPA) was also efficient in a clinically relevant model of transgenic MMTV-PyMT tumor that closely mimics the development of human breast cancer in an immunocompetent background. At early stage MMTV-PyMT tumors, treatment with ICD-inducing KT-1 significantly remodeled the immune status of the tumor microenvironment and effectively eliminated tumors by augmenting a profound activation of tumor-reactive CD8⁺ T cells and depletion of MDSCs (myeloid-derived suppressor cells). However, advanced tumors treated with KT-1 only appeared to remain under partial control of immune activation, and adapted the immune suppressive mechanism of elevating PD-L1 expression on both cancer cells and myeloid cells to enable continuous tumor growth. The addition of MPPA to KT-1 not only depleted the PD-L1 expression from the two disparate cellular sources, but also directly reduced the number of immunosuppressive myeloid cells, thereby leading to a shrinkage of larger tumors (Figure 5I). The combinatory strategy of ICD-inducing and PD-L1 blocking approaches on transgenic MMTV-PyMT tumor model suggests a rational design of polymer-based cancer immunotherapy [194].

The field of polymer-drug conjugates advanced immensely and principles of design are fitted with the knowledge of biological and immunological effects and delivery barriers these conjugates are exposed in vivo. Nanomedicine clinical trials exposed a complex situation and a large interpatient heterogeneity [195,196]. The most important issue is the development of methods for patient stratification [177,197]. In contrast to the advancement of antibodies or other targeted therapies, where patients have been selected based on the expression of the target antigen on cancer cells, no patient selection has been made in clinical trials with nanomedicines based on tumor susceptibility to the enhanced and permeability (EPR) effect [175,176,198,199]. Also, basic appreciation of differences between water-soluble polymer-drug conjugates (nanomedicines) and nanoparticles needs to be recognized. In the literature both are usually used interchangeably without recognizing the advantages of polymer-drug conjugates over nanoparticles (flexibility, diffusibility, enhanced extravasation, decreased nonspecific uptake (phagocytosis)).

The enhanced intravascular half-life and increased intratumoral drug accumulation of the second-generation, backbone degradable HPMA copolymer-drug conjugates (compared to first generation conjugates) and the efficacy of combinations with MPPA or another macromolecular therapeutic demonstrates their high translational potential. Bringing new nanomedicines to the market is still a challenge for manufacturers in the pharmaceutical industry due to complexity in the chemistry, manufacturing and controls and good manufacturing practice requirements. Universities are not structured to translate research data into the clinics. This is the task of the industry. The interest of large companies to translate nanomedicines would increase dramatically following the FDA approval of the first polymer-drug conjugate. Enormous activities of start-up companies keep the pipeline strong. Concentrated efforts are needed to help start-up companies to overcome the “valley-of-the-death”.

Lessons learned from hydrogels self-assembled from graft or block copolymers (Section 1.2) inspired the design of drug-free nanomedicines

The outcomes of self-assembly of graft copolymers into 3D hydrogels (Section 1.2) offered diverse routes for translation. First, to generate a therapeutic depo by subcutaneously injecting a mixture of two graft copolymer solutions, one holding a therapeutic protein. The pharmacokinetics of release can be controlled by the structure of graft copolymers. However, this was discarded because similar results can be accomplished using less expensive polymers.

Notably, we were able to apply our experience with the design of self-assembling hydrogels to receptor crosslinking by soluble nanomedicines resulting in apoptosis initiation. We created a new paradigm in nanomedicines, Drug-free Macromolecular Therapeutics (DFMT), based on one of the mechanisms of apoptosis induction in non-Hodgkin lymphoma [200]. When two Rituximab (RTX) antibodies bind to a tetramer of CD20 they can be crosslinked via their Fc fragments by immunocompetent cells. This results in the translocation of CD20/RTX complexes to lipid rafts, mitochondrial depolarization, calcium influx, caspases activation and apoptosis initiation. We hypothesized that when biorecognition of complementary coiled-coil forming peptides results in self-assembly of graft copolymers into crosslinked hydrogels then biorecognition of these peptides at cell surface should result in crosslinking of receptors and initiation of apoptosis. As demonstrated in the next sections, we have validated this hypothesis in vitro and in vivo.

3. Drug-Free Macromolecular Therapeutics (DFMT)

3.1. DFMT design based on self-assembly of coiled-coil forming peptides

Based on our experience with self-assembling hydrogels we intended to use the CCE/CCK peptide pair as biorecognition moieties for receptor crosslinking and apoptosis initiation in B cell malignancies (Figure 6A) [201]. We selected CD20 as our first target based on its importance in B cell malignancies and the fact that it is very slowly internalizing. As mentioned above, the crosslinking of two RTX antibodies bound to a CD20 tetramer by immunocompetent cells via Fc receptors is one of the apoptosis induction stimuli. We hypothesized that the two complementary pentaheptad peptides, CCE and CCK (see 1.2.2 for structure), that form antiparallel coiled-coil heterodimers with high fidelity would form coiled-coils at cell surface when included in the following design of two nanoconjugates: a) bispecific engager, CCE attached to an anti-CD20 Fab’ antibody fragment (Fab’-CCE); b) crosslinking effector component, HPMA copolymer grafted with multiple copies of CCK, P-(CCK)_x, where P is HPMA copolymer backbone. Exposure of Raji cells to anti-CD20 Fab’-CCE decorates the cells with CCE based on antigen – antibody fragment biorecognition. Further exposure of decorated cells with P-(CCK)_x results in the formation of coiled-coils at cell surface based on CCE/CCK biorecognition. The multivalency of P-(CCK)_x ensures crosslinking of receptors (Figure 6B). High levels of apoptosis were achieved both in vitro [202] and in a Non-Hodgkin Lymphoma (NHL) animal model in vivo (Figure 6C) [203]. Three nanomolar doses of nanoconjugates produced long-term (100 days) survivors [203]. The use of two nanoconjugates permits pretargeting; it is frequently used in cancer radiotherapy [204]. In DFMT, Fab’-CCE is used as a pretargeting agent; after a time-lag to clear out nonspecific binding the multivalent effector P-(CCK)_x delivers the therapeutically active dose. The time lag between administration of the two conjugates is based on the pharmacokinetics and biodistribution of Fab’-CCE [205]. When the optimized pretargeting time lag (5 h) was used in disseminated non-Hodgkin lymphoma model in SCID mice, the therapeutic efficacy was significantly better than that of identical experimental conditions but with a 1 h interval [206]. Notably, the mechanisms controlling therapeutic efficacy of anti-CD20 mAbs (ADCC (antibody-dependent cellular cytotoxicity), CDC (complement-dependent cytotoxicity), and CD20-mediated apoptosis) require immune effector cells to function [207,208]. In contrast, DFMT triggers direct and specific apoptosis of B-cell malignancies without the help of effector cells. This is achieved by the design of synthetic effectors that reproduce the crosslinking function of immune effector cells [205,209].

Figure 6.

Drug-free macromolecular therapeutics (DFMT). (A) Design of DFMT was inspired by self-assembly of HPMA copolymers grafted with multiple copies of coiled-coil forming peptides into hydrogels. Exposing B cells to anti-CD20 Fab’-motif 1 (peptide or oligonucleotide), followed after a time lag with a multivalent macromolecular conjugate of motif 2 results in receptor crosslinking and apoptosis induction. The mechanism of apoptosis induction – calcium influx mitochondrial depolarization and caspase 3 activation is shown. (B) Structures of two component DFMT system and biorecognition at the surface of Raji cells. Confocal images show the biorecognition of both conjugates. Raji cells were exposed to Rhodamine Red-X labeled Fab’-CCE and FITC-labeled P-(CCK)x (P is HPMA copolymer backbone). CD20 is a very slowly internalizing receptor, so colocalization can be detected after 4 h [202]. (C) Therapeutic efficacy of DFMT against systemically disseminated Raji B cell lymphoma in C.B.-17 SCID mice (n=7). Premixed (Prem) administration: the two conjugates were mixed together 1 h before injection via the tail vein. Consecutive (Cons) administration involved the i.v. injection of 50 μg/20 g Fab’-CCE first and 1 h later the i.v. administration of 324 μg/20 g P-(CCK)x conjugate [203]. (D) Morpholino oligonucleotide (MORF) based DMFT, composed of Fab’-MORF1 and P-(MORF2)x. Therapeutic efficacy of the nanomedicine against systemic lymphoma in SCID mice. Four million Raji B-cells were injected via tail vein on day 0; incidence of hind-limb paralysis or survival of mice was monitored until day 125. One-dose treatment on day 1; three-dose treatment on days 1, 3, and 5. The paralysis-free survival of mice is presented in a Kaplan-Meier plot. Numbers of long-term survivors in each group are indicated. Statistics was performed with log-rank test (*: p < 0.05, ***: p < 0.0001, n.s.: no significant difference) [213]. (E) Incubation of resistant Raji 4RH cells with gemcitabine (GEM) upregulates CD20 expression. The upregulation was dependent on the time of incubation, not on the GEM concentration [219]. (F) In vivo validation of therapeutic efficacy of DFMT in C.B.-17 SCID mice bearing Raji 4RH lymphoma. Comparison of pretreatment with GEM and 2P-GEM (backbone degradable HPMA copolymer-GEM conjugate) revealed that 2P-GEM pretreatment was more effective than free GEM due to long intravascular half-life [219]. (G,H,I) Human serum albumin (HSA)-based DFMT. (G) Synthesis of multivalent HSA-(MORF2)x conjugate. (H) Dual-target (CD20 + CD38) DFMT efficacy on xenograft Non-Hodgkin’s lymphoma in SCID C.B-17 mice. Paralysis-free survival of single-dose DFMT-treated mice. Single dose treatment groups received a Fab’-MORF1 (1 nmol) i.v. injection followed by an HSA-(MORF2)₁₀ i.v. injection 5 h later on Day 1. Triple dose treatment groups received the same dose schedule on Days 1, 3, and 5. Saline control mice (black) are compared with Fab’_DARA + HSA-(MORF2)₁₀ treated mice (green), + HSA-(MORF2)₁₀ treated mice (blue), and Fab’_DARA + Fab’_RTX combination + HSA-(MORF2)₁₀ treated mice (red) (upper panel). Paralysis-free survival of triple-dose DFMT-treated mice (lower panel). Cohorts consisted of six randomly distributed mice per group (α=0.05, power=0.80). **p<0.01, *p<0.05, n.s. not significant by logrank (Mantel-Cox) test [230]. (I) Type II antibody-based DFMT. Type II Abs (Obinutuzumab, OBN) binds within one CD20 tetramer with the conformation compatible with homotypic adhesion, leading to actin cytoskeleton remodeling and lysosome disruption. Modification of OBN with one MORF1 does not impact the binding of OBN-MORF1 to CD20 and following binding Type II effects occur. Further exposure to multivalent effector HAS-(MORF2)_X results in clustering the OBN-MORF1-CD20 complexes into lipid rafts and Type I effects occur. This new approach combines effects of both antibody types resulting in very high apoptotic levels [235].

The biorecognition of nanoconjugates at the cell surface was validated by imaging techniques including 2-channel FMT (fluorescence molecular tomography), 3D confocal microscopy, 4-color FACS [210] and dSTORM (direct stochastic optical reconstruction microscopy) [211]. Notably, the immunogenicity of peptides and their conjugates was evaluated and proven satisfactory for translation into the clinics [212].

3.2. DFMT design based on hybridization of morpholino oligonucleotides

The biorecognition of the coiled-coil forming oligopeptides, CCE and CCK, in the DFMT system worked well both in vitro [202] and in vivo [203]. However, to achieve a strong anticancer effect (produce tumor-free long-term survivors), we used a 1:25 molar ratio of CCE equivalent (in Fab’-CCE) to CCK equivalent (in P-(CCK)₉) [203]. This is because the individual peptide sequences (CCE and CCK) do not have a pronounced secondary structure at pH 7 and are in a random coil conformation [11]. Attachment of oligopeptides to macromolecules increases their secondary structure only slightly [10,11]. Consequently, Fab’-CCE and P-(CCK)_y interact first via hydrophobic and electrostatic interactions, and then the oligopeptides fold fast into a strong antiparallel coiled-coil heterodimer [12]. Such relatively complex binding pattern likely results in inadequate interaction of polymer conjugates with Fab’ conjugates when administered at the 1:1 molar ratio condition. Therefore, we tried to identify a biorecognition pair that would bind efficiently at the 1:1 molar ratio. Morpholino oligonucleotides have been selected due to their fast hybridization, excellent binding affinity, stability in plasma, and water-solubility.

We designed a pair of phosphorodiamidate morpholino (MORF) oligomers, MORF1 (5’-GAGTAAGCCAAGGAGAATCAATATA-3’) and MORF2 (5’-TATATTGATTCTCCTTGGCTTACTC-3’), as the biorecognition motifs for the “drug-free” therapeutic system [213]. The MORF oligos are charge neutral, resulting in stronger binding than natural DNA and RNA. Hybridization of the MORF pair has distinct binding specificity, which avoids potential off-target effects. In addition, MORF oligos have good aqueous solubility and advantageous pharmacokinetics. The sequences of MORF1 and MORF2 were designed to achieve optimal binding efficiency. The A/T/C/G content was selected to achieve optimal binding efficacy and specificity (GC = 35–65% [214]), maintain aqueous solubility (G < 36% [214]), and potentially provide favorable pharmacokinetics (number of C < 7 to avoid rapid kidney uptake [215]). After the base composition was determined, the sequences were generated by a scrambling software to minimize off-target binding with human and murine mRNA and further optimized to prevent self-complementarity [213]. This new therapeutic system was composed of two hybrid conjugates, similar to peptide design: (1) anti-CD20 Fab’ linked to MORF1 (Fab’-MORF1), and (2) HPMA copolymers grafted with multiple MORF2 (P-(MORF2)_x). The two conjugates self-assembled via MORF1-MORF2 hybridization at the surface of CD20⁺ B-cells, crosslinked CD20 and initiated apoptosis in vitro and in vivo [206,213,216]. Three-dose treatment of systemically disseminated Raji B cell lymphoma in C.B.-17 SCID mice resulted in long term (125 days) survivors without detection of residual disease (Figure 6D) [213]. The system was optimized based on the study of the relationship between the detailed structure of the nanoconjugates and apoptosis induction in Raji cells: (i) An improvement of apoptotic activity was observed for a 28 base pair MORF sequence when compared to MORFs composed of 20 and 25 base pairs; (ii) The higher the valence of P-(MORF2)_x the higher the levels of apoptosis (up to a saturation valency); (iii) Higher molecular weight of P-(MORF2)_x induced higher levels of apoptosis; and iv) A miniPEG₈ spacer was effective in enhancing apoptotic levels in contrast to a miniPEG₂ spacer [216].

3.2.1. Mechanism of apoptosis induction.

The study of apoptotic mechanism in Raji B cells revealed that, following biorecognition of two nanoconjugates and CD20 crosslinking at cell surface, the crosslinked complex CD20-Fab’-MORF1 : MORF2-P relocates to lipid rafts and initiates signaling pathways resulting in calcium influx, mitochondrial depolarization, cytochrome c release and caspase 3 activation. Bcl-2 inhibition contributes to mitochondrial depolarization and to the apoptotic process [217].

3.2.2. Amplification of CD20 crosslinking enhances apoptosis induction in RTX-resistant cells.

RTX resistance is a common challenge in patient management. Detailed studies of three pairs of RTX-sensitive/RTX-resistant cell lines, Raji/Raji 4RH, RL/RL 4RH, and U-2932/U-2932 4RH [218] demonstrated the potential of DFMT to overcome RTX resistance. In resistant cells with low level of CD20 expression (Raji 4RH and RL 4RH), preincubation with GEM or HPMA copolymer-GEM conjugates augmented CD20 expression and restored DFMT susceptibility to induce high levels of apoptosis. Notably, the combinatory treatment with backbone degradable HPMA copolymer-GEM conjugate (2P-GEM) and DFMT significantly extended the survival time in NOD/SCID IL2R γ^null (NSG) mice bearing systemically disseminated Raji 4RH RTX-resistant lymphomas and considerably reduced the cancer spreading in bone marrow (Figure 6F) [219].

Sensitive U-2932 and resistant U-2932 4RH cells had comparable levels of CD20 expression. It was shown that U-2932 4RH cells rapidly take up RTX-CD20 complexes before CD20 is sufficiently crosslinked, which is an Fc-dependent pattern [219]. DFMT has in-built advantages to evade this deficiency due to Fc deletion in Fab’-MORF1. In U-2932 4RH cells DFMT overcomes RTX resistance due to two characteristics. First, the deletion of Fc-region in Fab’-MORF1 decreases the Fc-assisted CD20 internalization as well as trogocytosis and increases CD20 availability at cell surface to be crosslinked. Second, the multivalence of P-(MORF2)_x intensifies CD20 ligation and maximizes the induction of apoptosis [219].

These results offered first demonstration of using CD20 crosslinking as augmenting strategy to overcome RTX resistance.

3.2.3. Other receptor crosslinking approaches.

Alternative drug-free strategy is to use multivalent macromolecule - Fab’ conjugates for direct crosslinking of receptors and apoptosis initiation. HPMA copolymers grafted with multiple Fab’_CD20 fragments effectively induced apoptosis in Raji, Ramos, Daudi, and Namalwa cells. Both valency and molecular weight of the carrier influenced the biological activity [220,221]. Similarly, HPMA copolymers containing multiple grafts of DR5 (death receptor) binding peptides induce apoptosis in human colon carcinoma COLO205 cells [222]. It appears that apoptosis initiated by receptor crosslinking is a general phenomenon. Alternatively, the premix of two nanoconjugates (Fab’-MORF1 + HSA-(MORF2)_x or Fab’-MORF1 + P-(MORF2)_x) can effectively induce apoptosis in vitro and in vivo [202,203,213,216] with similar outcomes as the consecutive administration. Liposomes containing numerous covalently bound anti-DR5 TRA-8 antibodies induced apoptosis in several cell lines including COLO205, A375, and Jurkat [223].

3.2.4. Patient samples – PHPMA based DFMT.

The efficacy of DMFT was evaluated in cells isolated from 44 patients with various B cell malignancies, chronic lymphocytic leukemia (CLL), diffuse large B cell lymphoma (DLBCL), marginal zone lymphoma (MZL), follicular lymphoma (FL), mantle cell lymphoma (MCL), and Burkitt’s lymphoma (BL). DFMT induced apoptosis in 65.9% of patient samples [224]. High-risk mutations such as11q22, 13q14 and 17p13 deletions, which are judged deficient prognostic indications in CLL, did not impede the therapeutic efficacy of DFMT treatment [201]. On the other hand, in patient samples with poor response to DFMT treatment, low CD20 expression level was observed. Pre-treatment with GEM [225] or HPMA copolymer-GEM conjugates enhanced surface CD20 expression and reestablished the cell responsiveness to DFMT [219]. DFMT effectively initiated apoptosis of tumor cells from patients with a variety of B cell malignancies, regardless of genomic aberrations. The outcomes, in agreement with our previous patient cell data [206,226], indicate that DFMT is an efficient approach for the treatment of diverse B cell malignancies.

3.3. Human serum albumin (HSA)-based DFMT

HSA, a natural transport protein, was employed in drug delivery due to its inherent properties such as long circulatory half-life, natural biocompatibility and multiple amino groups suitable as ligand binding sites. It is an effective carrier for drugs, peptides, and antibodies in clinical applications. The FcRn receptor plays a key role in maintaining high levels of HSA in the circulation. FcRn prevents HSA from degradation by recycling the FcRn-HSA complex back from endosomes to the cell surface [227]. Recently, we have been focusing on HSA-(MORF2)x effector structure as a part of the DFMT system (Figure 6G).

The results of apoptosis induction when employing HSA-based effector (crosslinking) component (HAS-(MORF2)_x) were similar to the HPMA copolymer-based effector (P-(MORF2)_x) [209]. Interestingly, HAS-based coiled-coil peptide system is also effective in initiating high level of apoptosis in Raji cells [228]. Recently, we focused on the HSA-based morpholino DFMT system as we assume it will be easier to translate the system into the clinics [206,229–233].

3.3.1. Simultaneous crosslinking of CD38 and CD20 receptors enhances apoptosis.

Raji (Burkitt’s lymphoma) cells are CD20 and CD38 positive. Simultaneous crosslinking of both receptors on the surface of NHL and multiple myeloma (MM) cells significantly increased apoptosis when compared to crosslinking of individual receptors. Fab’_RTX-MORF1 and Fab’_DARA-MORF1 (DARA is daratumumab) were administered as a targeting motif cocktail. Subsequent administration of the crosslink-inducer, HAS-(MORF2)₁₀, crosslinks CD20 and CD38 on the cell surface [230,231]. The monitored effects of CD20/CD38 crosslinking demonstrate enhanced apoptosis induction in the CD20⁺/CD38⁺ Raji cell line and significantly improved survival in an in vivo xenograft mouse model over single-target DFMT controls (Figure 6H).

The dual-target receptor crosslinking using DFMT could improve the efficacy of immunotherapy. In the two-step approach the targeting step is separated from the crosslink-inducing, therapeutic step. Consequently, different targeting motifs aiming various receptors can be interchanged, but the crosslink-inducing therapeutic step remains unchanged. A library of dual-target DFMT systems can be prepared and evaluated easily due to the high customizability in the targeting step. No burdensome synthesis or characterization is needed to produce a library of bispecific systems, where generation of an analogous library of BsAbs would require transfection, expression, purification and characterization of each individual IgG [234].

3.3.2. DFMT based on Type II antibodies.

RTX and other anti-CD20 antibodies (ofatumumab and OBN) dramatically improved treatment of NHL and CLL. Individually or in combination with chemotherapy (e.g., R-CHOP, a combination with cyclophosphamide, DOX, vincristine, and prednisone) they produced better clinical outcomes. Anti-CD20 antibodies are divided into Type I such as RTX and Type II such as OBN; they have different patterns of binding to CD20 receptor. RTX binds between CD20 tetramers resulting in accumulation in lipid rafts, calcium influx and caspase activation. OBN binds within one tetramer with the conformation consistent with homotypic adhesion zones, resulting in actin cytoskeleton remodeling and lysosome disruption.

We designed, synthesized, and evaluated the DFMT for the treatment of B-cell malignancies using whole antibodies for the synthesis of the bispecific engager. The design boosts the activity of Type II OBN by triggering the apoptosis activation pathways of both types of antibodies. This new system is composed of two nanoconjugates: a) bispecific engager, OBN-MORF1 (OBN conjugated to one morpholino oligonucleotide MORF1); and b) a crosslinking (effector) component HSA-(MORF2)_X. Attachment of one MORF1 to OBN does not impact the binding of OBN-MORF1 to CD20 and following binding to CD20 Type II effects occur. Additional exposure to multivalent effector HSA-(MORF2)_X results in clustering the OBN-MORF1-CD20 complexes into lipid rafts and Type I effects occur. This new approach, called “clustered OBN (cOBN)” combines effects of both antibody types resulting in very high apoptotic levels (Figure 6I) [235].

3.3.3. Combination of DFMT with low molecular weight chemotherapeutics.

As mentioned above, OBN-based DFMT is composed of OBN-MORF1 and HSA-(MORF2)_x. Gambles et al. evaluated the combination of OBN-based DFMT with three different classes of therapeutics: DNA alkylating agents, proliferation pathway inhibitors, and DNA replication inhibitors. The combination with proliferation pathway inhibitors (idelalisib, ibrutinib, apitolisib) was antagonistic (CI > 1), whereas the combinations with DNA alkylating agent (chlorambucil), and with DNA replication inhibitors (GEM, DTX, etoposide, DOX) were synergistic (CI < 1). The synergism of the combination of DFMT with etoposide was validated in vivo on a disseminated human NHL xenograft mouse model [232].

3.3.4. Patient cells – HSA based DFMT

Similarly, to the HPMA-based DFMT, HSA-based DFMT was efficient in inducing apoptosis in cells isolated from patients diagnosed with chronic lymphocytic leukemia (CLL). A novel strategy, employing crosslinking of CD20 and/or CD38 receptors was employed and cells from 56 patient were evaluated. Fab’ fragments from OBN and Isatuximab (ISA) were used in the synthesis of anti-CD20 (Fab’_OBN-MORF1) and anti-CD38 (Fab’_ISA-MORF1) bispecific engagers; HSA-(MORF2)₁₀ served as the multivalent effector. The efficacy of DFMT was significantly influenced by the expression levels of CD20 and CD38 receptors. Dual-targeting DFMT strategies (CD20 + CD38) were more effective than single-target approaches, particularly in samples with elevated receptor expression. Pretreatment of patient cells with gemcitabine or ricolinostat markedly increased cell surface CD20 and CD38 expression, respectively. Apoptosis was effectively initiated in 62.5% of CD20-targeted samples and in 42.9% of CD38-targeted samples. DFMT induced effectively apoptosis in cells from patients with genetic aberrations [233].

Lessons learned from DFMT motivated the design of Multi-Antigen T Cell Hybridizers.

We learned that complementary 25 base pair morpholino oligonucleotides can efficiently hybridize in vivo, resulting in linking two macromolecules, receptor crosslinking and apoptosis initiation. This instigated the idea for the design of split bispecific antibodies composed from a combination of Fab’_x-MORF1 and Fab’_y-MORF2 that could link T and B cells. It did not escape our attention that, in addition to easy preparation, the new approach is suitable for the generation of targeting motif libraries to fit individual patient’s receptor profiles.

4. Multi-Antigen T Cell Hybridizers (MATCH)

Bispecific antibodies (BsAbs) are antibody (or antibody-derived) molecules that can bind two distinct epitopes. BsAbs can exist as whole antibody molecules with two distinct binding regions (Fab’) (Figure 7A) or as truncated antibody analogues composed of monomeric or multimeric binding domains derived from the antibody small chain variable fragment (scFv). Most designs simultaneously engage cancer cells (CD19, CD20, CD38) and cytotoxic T cells via CD3 [236]. The first FDA approved bispecific T cell engager (BiTE) was Blinatumomab – a fusion protein comprised of an anti-CD19 short chain variable fragment (scFv) and an anti-CD3 scFv joined and expressed as a single molecule. Blinatumomab can bind to CD19 on the surface of target malignant B cells, and simultaneously bind to CD3 on cytotoxic T cells in the T Cell Receptor (TCR) complex [237].

Figure 7.

Multi-antigen T cell hybridizers (MATCH). (A) Schematic of bispecific antibody binding to B and T cells. (B) Schematic of MATCH – a split-antibody-like T cell engaging technology. (C) Confocal microscopy of co-cultured Raji B cells and Jurkat T cells. Cells were treated with Cy5-labeled Fab’_CD3-MORF2 (red) and Cy3-labeled Fab’_RTX-MORF1 (green) for 1 h. Overlaid fluorescence shows co-localization (yellow), designated by arrows, at the cell-to-cell interface. (D) Schematic depiction of the B cell targeting motif library composed of CD20, CD38, BCMA, and SLAMF7-targeting Fab’_x-MORF1 conjugates capable of hybridizing with complementary oligonucleotide strands on T cell engager molecules Fab’_CD3-MORF2 [238,239].

We designed a T cell recruiting technology based on antibody Fab’ fragments and biorecognition of complementary single stranded 25 base pair morpholino oligonucleotides. Heterodimerization of a cancer-targeting Fab’_x-MORF1 with a T cell directed Fab’_CD3-MORF2 creates a construct that resembles a bispecific antibody. In this innovative technology hybridization of MORFs mediates T cell recruitment and activation against malignant B cells (Figure 7B). For example, combination of Fab’_CD20-MORF1 and Fab’_CD3-MORF2 can bridge the interface between the target cancer B cell and the effector cytotoxic T cell (Figure 7C). This is a split antibody technology; its advantage is the possibility to create a targeting motif library based on the expression of receptors on patient cancer cells. Thus, a combination of Fab’_DARA-MORF1, Fab’_RTX-MORF1, Fab’_BCMA-MORF1, Fab’_CS1-MORF1 with Fab’_CD3-MORF2 can be generated and T cells activated against leukemia, lymphoma, and myeloma cells (Figure 7D). We coined the name Multi-Antigen T Cell Hybridizers (MATCH) for this technology [238–241].

To display the flexibility of the T cell activating system, the same well of naïve T cells was exposed to three different cancer cell lines (Raji lymphoma, MM.1S myeloma, and HL-60 leukemia) within a 72 h period. Every 24 h, a new cancer cell was added to the T cells along with a dose of MATCH specific to each cancer. No remaining cancer cells were detected after each of the three treatments [213,214]. The impact of MATCH on T cell exhaustion was also investigated and experiments are underway to optimize the T cell engager dose in vitro and in vivo.

The unique to MATCH is the capability to interchange the cancer cell targeting moieties contingent on the expression of receptors on patient cells. Moreover, the dose of the T cell engaging motif can be optimized independently from the target cell binding dose [241].

5. Conclusions

I commend (and am grateful to) the ones that have made it through the text. I tried to demonstrate that systematic basic research has a valued place in the scientific community. It contributes to the expansion of our knowledge and creates a foundation for the design of new paradigms based on the understanding of mechanisms of physiological processes. The novel platform technologies will ultimately impact clinical practice. My long-term interdisciplinary research has continuously proceeded from hydrogels to water-soluble polymer drug carriers, design of drug-free macromolecular therapeutics and finally to multi-antigen T cell hybridizers. In addition to scientific progress that enhanced our knowledge about hydrophilic biomaterials and therapeutics, numerous students and postdocs were trained and have been in leading positions worldwide.

Mentoring is really a rewarding activity. To witness the growth of graduate students/postdoctoral fellows into admirable scientists and colleagues is the highlight of academic life. I was fortunate to be associated with excellent students and postdoctoral fellows. My main aim in mentoring is to instill in students and postdoctoral associates the love for science.

Finally, I would like to make some suggestions to early-career scientists: The most important point is to enjoy your work. Choose an interdisciplinary area that is significant, challenging and fits your scientific background. Be open to learning new subjects. Remain unchanged after success. Lead by example, not force. Do not chase high-impact journals but focus on producing excellent and meaningful research. Science is a great career. Recently, I was asked what award I consider the most rewarding. I did not have to think about the answer. The fact, that after more than 60 years in science I still love to go to office every morning is the best reward. No award can compete.

Highlights.

Inspired by the invention of hydrogels and soft contact lenses by my mentors, my six decades of research have continuously proceeded from the synthesis of biocompatible hydrogels to the development of polymer-drug conjugates, then generation of drug-free macromolecular therapeutics (DFMT) and finally to multi-antigen T cell hybridizers (MATCH). In addition to generating general scientific knowledge, hydrogels from my laboratory have been transferred to the clinic, polymer-drug conjugates to clinical trials, and drug-free macromolecular systems have an excellent potential for personalizing patient therapies. The emerging novel platform technologies in biomaterial-based devices and implants as well as in personalized nanomedicines will ultimately impact clinical practice.