Enzyme-Instructed Self-Assembly for Cancer Therapy and Imaging

Abstract

Enzymatic reactions and self-assembly are two fundamental attributes of cells. It is not surprise that one can use enzyme-instructed self-assembly (EISA)—the integration of enzymatic transformation and molecular self-assembly—to modulate the emergent properties of supramolecular assemblies for controlling cell behaviors. The exploration of EISA for developing cancer therapy and imaging has made considerable progresses over the last five years. In this topical review, we discuss these exciting results and future promise of EISA. After describing several key studies to illustrate the progress of EISA in developing cancer therapy, we discuss the use of EISA for molecular imaging. Then, we give the outlook of EISA for developing supramolecular anticancer medicine that inhibits multiple hallmark capabilities of cancer.

Graphical Abstract

Introduction

It is now well-established that many major processes in a cell require assemblies of 10 or more protein molecules.¹ Supramolecular interactions, largely controlled by enzymatic reactions,² drive such assembling processes in cells. That is, Nature, combining enzymatic reactions and molecular self-assembly, evolves sophisticated assemblies of biomacromolecules for functions. For example, enzymes, such as spastin and katanin,³ regulate microtubule dynamics; procollagen peptidase controls self-assembly of tropocollagen to form collagen fibrils.⁴ We termed the process that integrates enzymatic reaction and self-assembly as enzyme-instructed self-assembly (EISA).^5–7 While using proteins as the substrates of enzymes for inducing self-assembly processes is well established, going back at least to the 1990s,^8,9 the use of small molecules as the substrates of enzymes for self-assembly only started at 2004.¹⁰ As a key feature of EISA, enzyme converts soluble precursors into the molecular building blocks for constructing assemblies, thus enabling spatiotemporal control of self-assembly. EISA represents a facile and powerful strategy to control the emergent properties of molecular assemblies.^11–13 Employing short peptides or other synthetic molecules, the research of EISA have made considerable of progress over the last one and half decade.^14–19 Various enzymatic reactions have been explored to instruct the self-assembly of peptides or their analogues to form diverse nanostructures.^15,16,18 EISA in cellular environment, as a molecular process to modulate the emergent properties of molecular assemblies for controlling cell behaviors, promises a wide range of biomedical applications, such as cancer therapy, molecular imaging, tissue engineering, and antibacterial therapy. Among these potential applications, the applications of EISA in cancer therapy and molecular imaging have made significant progress recently, which is the focus of this topical review.

The holy grail in cancer therapy is to increase the efficacy without increasing toxicity.^7,20 Conventional chemotherapy and molecular cancer therapy have contributed to cancer treatment for improving human health; nevertheless, it has met with various challenges, including low selectivity, short lifetime, and drug resistance. One of well-established approach to improve the efficacy of drug is prodrug. Prodrug, as a pharmacologically inactive molecule being converted into the active drugs in vivo, enhances the therapeutic efficacy by prolonging the lifetime as well as improving the selectivity of drugs for targeting cancer cells. However, drug resistance in cancer therapy, still remains the biggest challenge in cancer treatment. Sharing a common feature to the prodrug approach, EISA also uses enzymes to convert the inactivated precursors to active agents. But, as shown in Figure 1, EISA differs from the prodrug approach. In EISA, the active agents are the supramolecular assemblies; in the prodrug approach, the active agents are individual molecules. The individual small molecules diffuse much faster than the supramolecular assemblies in cellular environment. This subtle yet important difference not only makes EISA to be more selective to target cancer cells, but also reduces the odd of acquired drug resistance in cancer therapy because of the polymorphorism of the supramolecular assemblies. Moreover, spatiotemporal control of self-assembly via enzymatic reaction facilitates molecular imaging in cellular milieu, promising detections of cellular markers for cancer diagnostics. In this topical review, we discuss several key studies to illustrate the design principle of EISA for cancer therapy and molecular imaging. Particularly, we focus more on the works published over the last five years or so. We apologize to the researchers whose excellent works are unable to be included in this review.

Figure 1.

Comparison of prodrug approach and enzyme-instructed self-assembly (EISA) for cancer therapy.

We arrange this topical review in the following order. First, we focus on the recent progress of EISA for cancer therapy, being categorized by the microenvironment of EISA (i.e., pericellular, intracellular, or subcellular spaces). Then, we describe the application of EISA for molecular imaging, derived from the spatiotemporal control of EISA in cancer cellular milieu. Finally, we provide a brief outlook to discuss the prospective of using EISA to advance cancer therapy and imaging.

EISA for cancer therapy

Using EISA for cancer therapy, the active entities are the supramolecular assemblies, not individual molecules. That is, enzymes, being overexpressed by cancer cells, can convert individually innocuous small molecules into cytotoxic supramolecular assemblies in/on the targeted cancer cells. Moreover, supramolecular assemblies, being generated by EISA, promise to minimize acquired drug resistance in cancer therapy.²¹ When the molecules bind minimally with cellular components, the ability of self-assembly determines the efficiency of cancer therapy.²² For instance, by simply modifying the C-terminal of an EISA substrate (made of D-peptides) with N-methylacetamine (-CONHCH₃) or methylacetate (-COOCH₃) groups increases the self-assembling ability and improve the efficiency of the IC₉₀ values of the EISA substrate against Saos-2 cells more than ten times of that of the analog having carboxylic acid terminus.²³ Moreover, the supramolecular assemblies, being accumulative inside the targeted cells, should minimize multidrug resistance due to efflux pump.²⁴

Pericellular EISA.

According to the cellular milieu in which the EISA occurs, EISA for cancer therapy falls into three subtypes: pericellular, intracellular, and subcellular EISA. Pericellular space, being part of cellular microenvironment, houses a variety of ectoreceptors to sense the external environment and facilitate homeostasis so that assemblies of biomolecules in pericellular milieu can modulate cellular signaling and control cell fates.^25,26 Therefore, generating supramolecular assemblies by pericellular EISA promises an effective strategy to inhibit various cancer cells, including drug-resistant cancer cells. For example, alkaline phosphatase (ALP), being highly expressed on cell membrane and secreted from certain cancer cells,²⁷ can act as a trigger for pericellular EISA.²⁸ Specifically, placental alkaline phosphatase (PLAP), overexpressed on cervical adenocarcinoma cells (e.g. HeLa cells), allows the pericellular EISA of the peptide (Nap-ff_py, 1) by dephosphorylating the phosphotyrosine residue, resulting in the formation of nanonets on the cellular surface of HeLa cells (Figure 2a). These nanonets, composed of supramolecular assemblies, significantly reduce the viability of cancer cells.^17,29

Figure 2.

EISA for cancer therapy. (a) Pericellular EISA. (b) Intracellular EISA. (c, d) Subcellular EISA: (c) by using the organelle-targeting motif of EISA substrate, (d) by using the enzymes overexpressed on the targeted organelle. Reproduced with permissions from ref. 30, 39, 45 and 47.

As revealed in the mechanistic elucidation³⁰ of the above observation, the pericellular EISA regulates the emergent properties of supramolecular assemblies to induce the cell death. The precursor (1) and the molecular product (Nap-ffy, 2) of dephosphorylation, as individual molecules, are cell compatible. Only the peptide assemblies of 2, being formed in situ on the surface of HeLa cells, effectively kill the HeLa cells. The co-culture of HeLa and HS-5 cells confirms the selectivity of precellular EISA for cancer cells. Moreover, the assemblies on the HeLa cells induce the apoptosis via the multifaceted mechanisms, including the presentation of autocrine proapoptotic ligands to their cognate receptors in a juxtacrine manner as well as the direct clustering of the death receptors.³⁰ In addition to HeLa cells, a variety of other cancer cells, such as MES-SA/Dx5, T98G, and drug-resistant A2780cis, also express ALP on their membranes so EISA of 1 inhibits those cells as well.³⁰ In addition, the structural analogues of 1 undergo EISA at different rates and reveal the structure-activity relationship of EISA in pericellular cancer therapy.³¹ For pericellular EISA, several parameters are major contributors for inhibiting cancer cells: the self-assembling ability of the peptidic backbone, the isozymes of the triggering enzyme (i.e., ALP), and the number of enzymatic action sites on the peptides. These structure-activity relationships provide a foundation for achieving high specificity against cancer cells.³²

A major concern for using ALP-based EISA for inhibiting tumors was that liver also expresses ALP.³³ This concern has been addressed by designing a peptide substrate (Nap-ff_py-OMe, 3) of ALP and of carboxylesterase (CES). CES, being expressed in mammalian liver tissue,^34,35 accelerates the disassembly of the peptide assemblies that are resulted from the EISA catalyzed by ALP, thus detoxifying in liver cells.³⁶ This advance allows in vivo evaluation of the antitumor efficacy of EISA in an orthotopic osteosarcoma tumor murine model.²⁰ As shown in Figure 3, the precursor (4) undergoes ALP-catalyzed self-assembly to nanofibers on and inside the cancer cells (Saos-lung). Such an EISA process kills the cancer cell, and inhibits the tumor growth in the mice model without harming normal organs. Treating the mice models with 4 significantly improve the survival ratio of metastatic tumor bearing mice. This result, without relying on other drug molecules, confirms that it is feasible to using EISA to generate supramolecular assemblies for cancer therapy. Because ALP contribute to immunosuppression in solid tumors and there is, unfortunately, no suitable inhibitor for ALP, this work illustrates that EISA promises a fundamental new way for developing anticancer therapy when overexpressed ALP causes immunosuppression.

Figure 3.

EISA for tumor inhibit in immunosuppressive microenvironment that overexpresses ALP. ALP-instructed assembly for inhibiting metastatic osteosarcoma in an orthotopic osteosarcoma mouse model. Reproduced with permission from ref. 20.

Intracellular EISA.

Since the first demonstration of using EISA to inhibit cancer cells,³⁷ intracellular EISA by esterases has received considerable investigation. Cytosolic esterases catalyze the cleavage of the ester bond between the butyric diacid and the peptide, leading to the intracellular EISA in HeLa cells. Simply conjugating the taurine, a natural amino acid,³⁸ has led to intracellular EISA. Taurine conjugation of small D-peptide (NBD-ff-es-tau-(O), 5) drastically enhances the cellular uptake of the peptide by above 10-folds, likely through dynamin-dependent endocytosis and micropinocytosis (Figure 2b).³⁹ As a result, the peptide assemblies in cancer cells disrupt the dynamics of actin filament, causing in the apoptosis or necroptosis.⁴⁰ Moreover, this intracellular EISA also inhibits various drug-resistant cancer cells, such as triple negative breast cancer cell (HCC1937) and platinum-resistant ovarian cancer cells (SKOV3, A2780cis),⁴¹ as well as boosts the activity of cisplatin against the drug-resistant cancer cell.⁴²

Subcellular EISA.

Because cellular organelles play multiple essential roles in cellular functions, disrupting the functions of organelles, such as mitochondria and endoplasmic reticulum (ER), is emerging as a new strategy to reduce drug resistance in cancer therapy.⁴³ For this reason, subcellular EISA for cancer therapy promises high therapeutic efficacy by accumulating the assemblies in subcellular organelles. Two kinds of subcellular EISA has been reported up to date: (i) conjugating organelle-targeting motif with EISA substrates and (ii) using the enzymes overexpressed on the targeted organelle for EISA. In the first case, triphenyl phosphinium (TPP), one of the most-investigated motifs for targeting the mitochondrial matrix,⁴⁴ conjugates with the peptide to form NBD-ff_pyk(^εTPP) (6) (Figure 2c). The ALPs overexpressed by Saos-2 dephosphorylate 6 to form nanoscale assemblies. Then, the assemblies enter the Saos-2 cells via endocytosis, mostly caveolae/raft-dependent pathway.⁴⁵ After escaping from the lysosomes, the assemblies accumulate in the mitochondria matrix due to TPP, leading to the mitochondrial dysfunction for killing the Saos-2 cells. Although most mitochondria-targeting molecules, including the TPP-conjugated peptide, induces mitochondrial dysfunction, they, being lipophilic and cationic, lack cell specificity and become cytotoxic with their accumulation.¹⁴ The combination of EISA with TPP increase the selectivity for targeting the mitochondria of cancer cells.

Without accumulative and unwanted cytotoxicity of TPP or other lipophilic and cationic peptides, a branched peptide (Nap-ffk(^εG-FLAG)-NBD, 7) is able to target mitochondria via EISA on mitochondria (Figure 2d).^46,47 Containing the enterokinase (ENTK)-cleavable FLAG-tag (DYKDDDDK), 7 self-assembles to form micelles, which become nanofibers upon ENTK cleaving the FLAG-tag. In cell culture, the micelles, being taken up by the cells, turn into the nanofibers upon proteolytic cleavage of FLAG-tag catalyzed by intracellular ENTK, resulting in the accumulation of the nanofibers at the mitochondria.⁴⁷ The micelles of 7, acting as carriers, are able to deliver various cargos, such as small molecule drugs, proteins, or nucleic acids, into mitochondria. The subcellular EISA for cancer therapy is applicable for other organelles, such as ER. For example, the peptide assemblies, containing the L-homoarginine, accumulate at ER.⁴⁸ Via the EISA catalyzed by ALP, the L-homoarginine-containing peptides form the crescent-shaped assemblies, which adhere and disrupt the cell membrane. Because the ER membranes account for more than 50% of total cell membrane,⁴⁹ most of the peptide assemblies enter the cells through the disrupting plasma membrane, eventually accumulating at ER for selectively inhibiting cancer cells.

EISA for molecular imaging

Molecular imaging, which reports biological processes at the molecular and cellular levels within intact living organisms,⁵⁰ provides not only the significant information for understanding the underlying cellular dynamics, but also facilitates the early diagnosis of disease, by extension, the optimization of drug doses and the individualized therapeutic treatment.⁵¹ Being able to control the formation of the assemblies in accordance with enzymatic activity, EISA of molecular imaging agents becomes one of the strategies for spatiotemporal profiling of the enzymatic activity of living cells. For cell imaging, the fluorophores, such as dansyl group⁵² and 4-nitro-2,1,3-benzoxadiazole (NBD),^53,54 have been conjugated at the ε-amine site of the lysine of the peptide (Nap-FFK_pY, 8) for EISA. The NBD-conjugated peptides (Nap-FFK(^εNBD)_pY, 9) self-assemble upon the dephosphorylation catalyzed by ALP, and then the quantum yield of NBD is enhanced due to the hydrophobic environment of the peptide assemblies.⁵³ Therefore, the intracellular assembly of NBD-conjugated peptides leads to concentrate the NBD at the locations of phosphatases. This feature allows the investigation of dynamics and localization of phosphatases on/in living cells. For example, the treatment of 9 to HeLa cells results in bright fluorescence at ER, which agrees with that the protein tyrosine phosphatases localize at the cytoplasmic face of ER. Moreover, another NBD-conjugated peptide (10) can report the activities of intracellular/pericellular ALPs (Figure 4a).⁵⁴ As a reaction-diffusion controlled process (i.e., reaction makes the assemblies non-diffusive), EISA of 10 provides the opportunity to screen the expression levels and activities of ALP on a variety of living cells, including MCF-7, A2780, and Saos-2, with high spatiotemporal resolution. Recently, the profiling of intracellular (excluding pericellular) phosphatase activities becomes feasible by using an inhibitor of ectophosphatases.⁵⁵

Figure 4.

EISA for molecular imaging. ALP expressions are spatiotemporally detectable via (a) fluorescence imaging (FL), (b) magnetic resonance imaging (MRI), (c) photoacoustic imaging (PA), and (d) synergetic combination of near-infrared (NIR) FL and MRI. Reproduced with permissions from ref. 54, 58, 61, and 63.

Fluorescence imaging, delivering high sensitivity and resolution and requiring simple instruments, is, however, less useful for clinical diagnosis. The rapid decrease of the spatial resolution of optical imaging as a function of the imaging depth remains a challenge.⁵⁶ To address this issue, magnetic resonance or ultrasound becomes a modality for developing the EISA-based molecular imaging agents. Because magnetic resonance imaging (MRI) is an effective non-invasive molecular imaging technique and allows the excellent tissue penetration with the high spatial resolution, it has already found extensive usage in the clinic.⁵⁷ However, due to the intrinsic poor sensitivity of MRI, it used to be virtually impossible to detect molecular targets via MRI without contrast agents. Combining EISA with gadolinium (Gd)-based chelates, the most common MRI contrast agents, promises a new alternative for addressing technical limitation and for increasing sensitivity. The attachment of Gd (III)-tatraazacyclododecanetetraacetic acid (Gd-DOTA) at the C-terminal of the short peptide (Nap-FFF_pY, 11) by an ethylenediamine (EDA) linker allows to develop the advanced MRI contrast agent (Nap-FFF_pY-EDA-DOTA(Gd), 12), which enables spatiotemporally detection of the ALP activity (Figure 4b).⁵⁸ The molecules of 12, treated in the ALP-overexpressed HeLa tumor, undergo the dephosphorylation catalyzed by ALP; and then self-assemble into Gd nanofibers, leading to condensation of MRI contrast agent (i.e. Gd-DOTA) on the surface of the nanofiber. As a result, T₂-weighted MR imaging of the tumor improves remarkably with the increased rotational correlation time.⁵⁹

By replacing the contrast agents, other imaging modalities, such as photoacoustic (PA) imaging, can employ EISA for molecular imaging. PA, a rapidly growing non-invasive whole-body imaging modality, integrates near-infrared (NIR) excitation with ultrasonic detection.⁶⁰ The conjugation of NIR dye (IR775) to the N-terminal of a phosphotripeptide permits the formation of ALP-activatable probes (IR775-FF_pY, 13) for tumor-targeted PA imaging (Figure 4c).⁶¹ Because the nanoparticles, formed by the ALP-triggered self-assembly of 13, exhibit a 6.4-fold increase of the PA signal of 13, it is possible to detect the ALP-overexpressed HeLa tumors by means of the enhanced PA signals. Because condensation reactions of appropriate contrast agents can occur at target positions that overexpress ALPs, the applications of EISA for molecular imaging are recently receiving increasing attentions from the research community of molecular imaging.⁶²

Recently, Ye et al. reported an innovative, ESIA-based dual imaging probe for detecting tumor in a murine model.⁶³ They directly phosphorylated a near infrared (NIR) dye and attached it to the N-terminal of a self-assembling peptide, and added to a paramagnetic DOTA-Gd chelate at the C-terminal of the peptide. ALP dephosphorylates the resulting probe (P-CyFF-Gd, 14) to enable the probe to fluoresce and to self-assemble (Figure 4d). This EISA process simultaneously enhances NIR fluorescence (>70-fold at 710 nm) and MRI r₁ relaxivity (~2.3-fold). Using the 14 and its control molecule (P-Cy-Gd, 15), the authors confirmed that the EISA probe is superior to reveal tumors in the murine model. A notable result is that the probe accumulates more in tumor than in liver, confirming the selectivity of EISA to cancer cells, despite that liver expresses ALP.

Outlook

The development of EISA, especially for cancer therapy and molecular imaging, has progressed rapidly over the last five years. Although these works have confirmed the advantages of EISA, including high selectivity and minimal drug resistance, for potential cancer treatment; there are still much room for improvement. One typical disadvantage of EISA in cancer therapy is anticancer efficiency. Despite the IC₅₀ of EISA agents against cancer cells has decreased from submillimolar to several micromolar, it needs further improvement to reduce the IC₅₀ values to the nanomolar (or submicromolar) range. One approach toward this goal would be conjugate other anticancer drugs to the substrates of EISA, as reported by several labs.^24,64–73 Along this direction, Yang et al. recently reported a highly active EISA agent that inhibits HepG2 and A549 cancer cells with the IC₅₀ of 220 nM and 260 nM, respectively, as well as in vivo activity against HepG2 tumor in a murine model.⁷⁴ Considering the exceptional selectivity of EISA against cancer cells, it is also plausible to use relatively high dosages for therapy, as long as therapeutic index permits. Moreover, the emergent functions of peptide assemblies could enhance the curative effects because peptide assemblies can exhibit the advanced properties, which may be insignificant at the single molecular level.⁷⁵ In addition, combining physical approaches, such as radiation, with EISA would improve killing cancer cells.^76,77 Currently, most explored EISA imaging agents are fluorescent probes. It would be more clinically relevant to combine EISA with other imaging modalities, such as MRI, positron emission tomography (PET), and computed tomography (CT), and exciting progresses are being made.^78–81

While various re enzymes are already explored for EISA in the cell-free environment,^18,82–84 only a few have been tested for targeting cancer cells (Table 1). The lack of information about the activities of enzymes in different cancer cells, however, demands more studies. EISA, as an effective approach for profiling enzyme activities in cells,⁵⁴ may help this task. Although it is possible that EISA of peptides in vivo stimulates immune responses, this subject remains exploration.⁸⁵ Along the same concept of combination therapy, it would be fruitful to combine ESIA with another therapeutics for synthetic lethality, which is just at the beginning.⁸⁶ EISA, inhibiting cancer cells by enzymatic reaction (not enzyme inhibition) and by assemblies of molecules (not individual molecules), represents a first-in-kind approach for developing cancer therapy, and promises exciting opportunities for scientists to explore the emergent properties of supramolecular assemblies.

Table 1.

Enzyme-instructed self-assembly (EISA) by different enzymes and their applications.

Purpose	Enzyme	Reference
Cancer therapy	Alkaline phosphatase (ALP)	17–23, 28–32, 36, 45, 48, 69–71, 74, 77, 85, 86
	Esterase	20, 36–42
	Enterokinase (ENTK)	47
	Matrix metalloprotease (MMP)	16
	Furin	24
Molecular imaging	Alkaline phosphatase (ALP)	52–55, 58, 61, 63, 70, 81
	Furin	78, 79
	Caspase-3/7	80
Cell-free environment	Alkaline phosphatase (ALP)	10, 18
	Thermolysin	14, 15
	Enterokinase (ENTK)	46
	Matrix metalloprotease (MMP)	83
	β-Galactosidase	82
	β-Lactamase	84