siRNA therapeutics for breast cancer: recent efforts in targeting metastasis, drug resistance, and immune evasion

Abstract

Small interfering RNA (siRNA) has an established and precise mode of action to achieve protein knockdown. With the ability to target any protein, it is very attractive as a potential therapeutic for a plethora of diseases driven by the (over)expression of certain proteins. Utilizing siRNA to understand and treat cancer, a disease largely driven by genetic aberration, is thus actively investigated. However, the main hurdle for the clinical translation of siRNA therapeutics is to achieve effective delivery of siRNA molecules to tumors and the site of action, the cytosol, within cancer cells. Several nanoparticle delivery platforms for siRNA have been developed. In this Review, we describe recent efforts in developing siRNA therapeutics for the treatment of cancer, with particular emphasis on breast cancer. Instead of conventionally targeting proliferation and apoptosis aspects of tumorigenesis, we focus on recent attempts in targeting cancer’s metastasis, drug resistance, and immune evasion, which are considered more challenging and less manageable in clinics with current therapeutic molecules. siRNA can target all proteins, including traditionally undruggable proteins, and is thus poised to address these clinical challenges. Evidence also suggests that siRNA can be superior to antibodies or small molecule inhibitors when inhibiting the same druggable pathway. In addition to cancer cells, the role of the tumor microenvironment has been increasingly appreciated. Components in the tumor microenvironment, particularly immune cells, and thus siRNA-based immunotherapy, are under extensive investigation. Lastly, multiple siRNAs with or without additional drugs can be co-delivered on the same nanoparticle to the same target site of action, maximizing their potential synergy while limiting off-target toxicity.

Introduction

In 2018, roughly two million new cases of breast cancer were diagnosed worldwide [1], and 271,270 are estimated to be diagnosed in the US in 2019 [2]. Current treatment approaches include surgery, chemotherapy, radiation, molecular targeted therapy, and most recently immunotherapy. Trends for research and development of novel therapeutics have shifted from conventional targeting of cancer cells with cytotoxic agents to a more targeted approach that relies on the concept of oncogene addiction [3]. Oncogene addiction is a scenario where cancer cells rely on certain genes/proteins to grow uncontrollably, metastasize, and tolerate drug treatment, among others. These genes are typically mutated, amplified or overexpressed in cancer cells resulting in their oncogenic behaviors. Targeted therapies utilize monoclonal antibodies or small molecule inhibitors to inhibit the activity of such oncoproteins, reversing the cancerous phenotypes.

However, tumors are not merely masses of proliferating cancer cells, but rather complex tissues that also contain a repertoire of resident or recruited ostensibly normal cells that support tumor growth and progression, such as endothelial cells, pericytes, immune inflammatory cells, and fibroblasts [4]. These cells along with secretory factors (e.g., cytokines, chemokines, growth factors) and extracellular matrix (ECM) proteins collectively constitute the Tumor Microenvironment (TME) [5] (Fig. 1). The crosstalk of cancer cells and the components in the TME shapes cancer’s biological capabilities (i.e. hallmarks) [4]. Over the past two decades, the role of the TME has been extensively explored with rapid progress, resulting in the discovery of several impactful therapeutics. Clinical success of the recent immunotherapeutic approaches further emphasizes the role of the immunosuppressive TME, which allows and supports tumors to grow without being cleared by the immune system. A prime example is the discovery and clinical use of immune checkpoint inhibitors that release the brake of certain immune cells, unleashing their attack on cancer cells [6]. At the same time, other components in the immunosuppressive TME also partly account for why current immunotherapies (e.g., checkpoint inhibitors, chimeric antigen receptor (CAR) T-cell therapy) and other anti-cancer therapeutics work in only a subset of patients [6, 7]. Understanding the TME and designing targeted therapeutics to reprogram/re-educate the TME have thus been actively investigated.

Figure 1. Tumor microenvironment (TME).

Cancer cells are surrounded by several cellular and non-cellular components of the tumor microenvironment. Resident or recruited host cells in the tumor include stromal cells (fibroblasts, endothelial cells, and pericytes) and immune cells (T cells, B cells, NK cells, dendritic cells, macrophages, and myeloid-derived suppressor cells (inflammatory monocytes and neutrophils)). Non-cellular components of the TME include soluble factors (e.g., growth factors, cytokines, chemokines) and extracellular matrix (ECM). Together, these several components serve as a complex and dynamic network that drives several hallmarks of cancer. ECM = extracellular matrix; CAF = cancer-associated fibroblast; MDSC = myeloid-derived suppressor cells; NK cell = natural killer cell.

Conventional targeted therapeutics, which comprise monoclonal antibodies and small molecule inhibitors, are not without shortcomings. These compounds can target only a certain set of proteins/pathways, so-called “druggable” targets. Monoclonal antibodies cannot be taken up by cells effectively, so they cannot drug intracellular proteins. Thus, targets for monoclonal antibodies are limited to either proteins on the cell membrane or secreted and extracellular proteins. Various strategies, including uses of nanoparticles, protein-transduction domain, and viral envelopes, have been explored for intracellular targeting of antibodies [8], but none have been in clinical applications to date. Small molecule inhibitors warrant investigation due to the broadened range of proteins that can be explored as therapeutic targets. However, the process of developing new inhibitors is complicated and specificity issues are not uncommon. Moreover, many proteins still cannot be practically targeted by small molecule inhibitors due to their large flat contact areas and a lack of deep hydrophobic pockets where small molecules can insert and perturb their activity [8].

RNA interference with small interfering RNA (siRNA) in contrast has no such limitations. siRNA has been routinely used in research for gene silencing of any protein with known mRNA sequences with high specificity, mediated by Watson-Crick base-pairing interaction between siRNA and the targeted mRNA. Nevertheless, the main hurdle of translating siRNA technology to clinics is to overcome systemic delivery issues. Research advances in this area are progressing, yet there remains no siRNA therapeutics approved by the FDA for cancer treatment. Patisiran (Alnylam) has been recently approved by the FDA to treat hereditary transthyretin amyloidosis, and is the first FDA-approved siRNA therapeutic [9, 10]. This hallmark success validates the pursuit and promise of the clinical feasibility of siRNA technology. With its immense potential to target any oncoprotein, siRNA technology opens doors to countless effective therapeutic possibilities. Utilizing siRNA technology for treating cancers, particularly beyond liver targets, has thus been intensively researched.

This review focuses on recent efforts within the past five years that utilize siRNA technology to therapeutically target the cancer hallmarks in breast cancer. Herein, we focus on three cancer traits that are considered clinically challenging; including metastasis, drug resistance, and immune evasion. These traits cannot be effectively managed by current therapeutic compounds in clinics. With its ability to target the undruggable genome, siRNA technology is poised to make a revolutionary impact in cancer medicine.

Breast cancer clinical subtypes and current treatments

Breast cancer consists of three major clinical subtypes: estrogen receptor (ER)/progesterone receptor (PR)-positive, human epidermal growth factor receptor type 2 (HER2)-positive, and triple negative (ER−/PR−/HER2−) breast cancer.

ER/PR+ (hormone receptor-positive) breast cancer accounts for about 60–70% of breast cancers [11, 12]. Endocrine therapy, either aromatase inhibitors or selective estrogen receptor degraders (e.g., fulvestrant), represents the first-line treatment for this subtype [12]. Ovarian function suppression (by drugs or surgery) is also given with endocrine therapy for pre-menopausal women [11]. This subtype has the best prognosis with a median survival of about 57 months, while the median survival is about 33 months for hormone receptor-negative breast cancer [13]. Still, resistance to the first-line treatment is common and several clinical trials have been evaluating additional targeted agents, such as inhibitors of CDK4/6, mTOR, and immune checkpoints, in combination with endocrine therapy [14].

The HER2+ subtype accounts for about 15–25% of invasive breast cancer [15, 16] and is known for its poor prognosis [17]. Trastuzumab (first HER2-targeted therapy) has substantially improved the prognosis of this subtype [18], but resistance to the treatment is still common. This leads to newer HER2-targeted therapies (e.g., pertuzumab, lapatinib, neratinib) that have replaced or been used with trastuzumab [19–21] to mitigate the observed resistance, but the best outcomes were achieved with a combination of pertuzumab, trastuzumab, and taxane [22]. Thus, this combination serves as a first-line treatment for both early and metastatic HER2+ breast cancer. Nevertheless, it has improved progression-free survival to only 18.5 months (vs. 12.4 months on trastuzumab + docetaxel) [22]. Due to unsatisfactory results in the phase III MARIANNE study [23], the newest drug T-DM1 (antibody-drug conjugate trastuzumab emtansine) remains the second-line therapy and the current first-line treatment will unlikely change soon.

About 15% of all breast cancer is triple negative (TNBC, ER-/PR-/HER2-) [24], which can be further classified into non-basal and basal-like TNBC at about 50% each by two additional histochemical markers, epidermal growth factor receptor (EGFR) and cytokeratin 5/6 (CK5/6). Basal-like TNBC is EGFR+ and/or CK5/6+, while non-basal TNBC is negative for all five markers [25]. From a study that monitored 3,726 early stage breast cancer patients for 15 years [26], basal-like TNBC was diagnosed in the youngest women (53 (median), 42 to 65 (range)) and was the most aggressive. Treatment for TNBC relies on new chemotherapeutic drugs (not previously used in the primary setting), or re-challenges with a combination of the same drugs (e.g., doxorubicin + paclitaxel). Despite such treatments, the median overall survival from metastasis to death is 6 months for basal-like TNBC vs. 11 months for non-basal TNBC or up to 2.2 years for the least aggressive breast cancer (ER/PR-positive) [26]. Most recently, the FDA approved atezolizumab (PD-L1 immune checkpoint inhibitor) plus nab-paclitaxel for patients with locally advanced or metastatic triple-negative breast tumors expressing PD-L1 [27]. However, in treatment-naive patients with metastatic TNBC, progression free survival in the PD-L1-positive population is still only 7.5 months with atezolizumab/nab-paclitaxel, compared to 5.0 months with nab-paclitaxel alone [28].

The suboptimal treatment regimens for breast cancer call for a novel technology to address such unmet medical needs.

Small interfering RNA (siRNA) for cancer therapeutics

The Nobel Prize in Physiology or Medicine in 2006 was jointly awarded to Andrew Z. Fire and Craig C. Mello for their discovery of RNA interference [29]. Since its discovery in 1998, siRNA has become a very valuable tool to knock down targeted genes in cells. Most (>80%) of the attractive therapeutic targets found by the Cancer Genome Atlas project [30] are currently not druggable by conventional means (e.g., monoclonal antibodies and small molecule inhibitors) [31], and hence RNA interference with siRNA holds great promise in the field of cancer therapy. siRNA can be designed to silence any gene responsible for different cancer hallmarks [4], such as angiogenesis, invasion, and metastasis.

Unfortunately, in vivo delivery of siRNA to tumors has proven challenging due to poor cellular uptake and the extremely short blood half-life of siRNA. siRNA, like other oligonucleotides, does not readily cross the negatively charged cell membrane due to its high molecular weight and anionic charge of its phosphodiester backbone [32]. Furthermore, siRNA is readily degraded by serum endonucleases and is rapidly cleared by glomerular filtration because siRNA is smaller than the renal filtration threshold (~5–10 nm) [32, 33]. Several strategies have been developed to prolong the fate of siRNA systemically and deliver siRNA to cancer cells [34].

Nanoparticle platforms for siRNA delivery to tumors

Nanoparticles are considered the most promising carriers for siRNA delivery. Viral-based carriers have been explored for this application; however, major concerns include an unwarranted immunogenic response and insertional mutagenesis [35]. Nanoparticles have advantages over siRNA conjugates because of (1) higher number of siRNAs delivered per uptake event (e.g., thousands of siRNA molecules per a single nanoparticle vs. 1–10 per conjugate) and (2) the ability to incorporate other components, such as homing targets, endosomal escape units, drugs for synergistic treatment benefits, and imaging agents. The first siRNA delivery platform that entered clinical trials for cancer treatment is the cyclodextrin-based material for treating metastatic melanoma [36], but further development was halted due to the short half-life and poor stability in systemic circulation [37, 38]. Since then, most of the siRNA delivery platforms that have entered cancer clinical trials are lipid-based and success has been found when targeting cancers with liver involvement due to the tendency of lipid-based nanoparticles to accumulate in the liver [39–41]. Delivering siRNA to other solid tumors is still a challenge and under continuing investigation.

To be effective, the nanoparticle platform has to overcome several barriers (Fig. 2). First, upon introduction to systemic circulation, nanoparticles must delay the fast renal clearance of free siRNA. Also, nanoparticles of 20–200 nm can preferentially accumulate in the tumor due to the enhanced permeability and retention (EPR) effect, which describes tumors with leaky vasculature and poor lymphatic drainage [42, 43]. Secondly, once siRNA-nanoparticles arrive in the tumor, they have to be effectively taken up by cancer cells. The use of targeting moieties, such as antibodies, aptamer, peptides, etc., can promote uptake of the nanoparticle into target cells, and the common route of nanoparticle uptake is endocytosis. Lastly, siRNA has to escape the endosome to the cytosol, which is its site of action, in a timely manner before potential degradation within the lysosomal compartment. We refer the readers to our prior review on design concepts of nanoparticles for siRNA delivery and several nanoparticle platforms under preclinical and clinical development [44].

Figure 2. Schematic illustration of nanoparticle-mediated siRNA delivery to tumors upon systemic administration.

(A) Nanoparticles protect siRNAs from blood degradation and prolong their blood circulation time. This enhances the prospect of siRNA-nanoparticles to accumulate in tumors. (B) Endocytosis is the main route of nanoparticle uptake, which can be mediated by targeting ligands or cationic components on the nanoparticles. (C) Upon endosomal escape into the cytosol, siRNA is processed by intracellular machinery (RNA-induced silencing complex, RISC) and degrades its target complementary mRNA. Reproduced with permission from Elsevier[44].

In the following sections, we describe the recent works that utilize siRNA technology to attack the various clinically challenging characteristics of breast cancer and briefly describe the platform of choice in each work. Summary of all cited works in the following sections is also found in Table 1.

Table 1:

Summary of the recent attempts in developing siRNA-nanoparticles to target different hallmarks of breast cancer.

Targeted hallmarks	siRNA gene targets	Other therapeutics co-delivered on the same NP	Targeting agents (beyond EPR, and charge driven uptake)	Carriers	Intended target cells/ route of administration in vivo (if applicable)	(Tumor) Models	Key results	Ref
Metastasis	PTPN22	none	hyaluronic acid (CD44 targeting)	lipid-substituted PEI	cancer cells	In vitro (MDA-MB-231, SUM149PT, MCF7)	Among six phosphatases tested, siPTPN22 inhibited the highest percentage of cell migration in vitro.	[54]
	p65	cisplatin produg	none	PEG-b-PAGA-b-PDPA micelle	cancer cells (i.v.)	Orthotopic tumor model with spontaneous lung metastasis (4T1)	Cisplatin prodrug + sip65 effectively inhibited both primary tumor and its metastasis.	[57]
	Twist	paclitaxel	none	PEI-p- PDHA/PEG- PDHA micelle	cancer cells (i.v.)		NP uptake was mediated by MMP-9 cleavage; Co-delivery inhibited growth of 4T1 tumor and its metastasis.	[58]
	VEGF	CXCR4-binding cyclam	none	cyclam-PEI polyplex	cancer cells (i.t.)		siVEGF reduced primary tumor growth, while CXCR4 antagonist reduced metastasis to lung.	[61]
	Lcn2	CXCR4 antibody	CXCR4 antibody	DOPC/DODAP /N-dod-PE liposome	cancer cells	In vitro (HCC1500, MDA-MB-231,..)	Effectively inhibited migration effect in vitro without affecting viability.	[62]
	DANCR IncRNA	none	RGD peptide	Amino-lipid NP	Cancer cells (i.v.)	Orthotopic tumor (MDA-MB-231, BT549)	Effectively inhibited migration, invasion, and 3D growth in vitro; and tumor growth in vivo. Downregulated multiple cancer-driven pathways.	[63]
	PLK1	antioxidant nanoparticle	trastuzumab (anti-HER2)	MSNP-PEI-PEG	cancer cells (i.v.)	i.v. injected metastasis (LM2–4/H2N)	siPLK1 reduced overall tumor burden, while antioxidant nanoparticles decreased tumor spread to other organs.	[64]
Drug Resistance	Pgp	doxorubicin	none	MSNP-PEI-PEG	cancer cells (i.v.)	Ectopic breast tumor (MCF-7/MDR)	siPgp increased drug uptake to cancer cells; codelivery of siPgp + dox inhibited tumor growth effectively.	[67]
	MRP-1	doxorubicin	hyaluronic acid (CD44 targeting)	liposome coated lbl with PLA	cancer cells (i.v.)	Ectopic breast tumor (MDA-MB-468)	siMRP1 had no effect on tumor growth, but significantly improved the efficacy of doxorubicin upon co-delivery by nP.	[68]
	ATG7	docetaxel	none	Lipoic acid- peptide micelle	cancer cells (i.v.)	Ectopic breast tumor (MCF-7)	siATG7 had no therapeutic benefit, but improved the efficacy of chemodrug.	[69]
	Ca2+channel	doxorubicin	none	MSNC-amine-PEG	cancer cells (i.t.)	Orthtopic breast tumor (MCF-7/ADR)	siRNA against calcium channels on the NP decreased intracellular calcium level, synergizing with doxorubicin treatment. Intratumoral injection inhibited growth of the injected tumor.	[70]
	CXCR4	none	none	chitosan NP	cancer cells	In vitro (MCF-7)	CXCR4 knockdown had no effect on viability per se, but sensitized cell to cisplatin treatment in vitro.	[71]
	MTDH	paclitaxel	none	PEI-PLGA NP	cancer cells (i.v.)	Orthotopic tumor (MCF-7)	siMTDH had no effect on tumor growth, but significantly improved the efficacy of paclitaxel upon co-delivery by NP.	[72]
	HER2	trastuzumab	trastuzumab (anti-HER2)	MSNP-PEI-PEG	cancer cells (i.v.)	Orthotopic tumor (HCC1954)	siHER2 inhibited growth of tumor that is resistant to antibody-based HER2- targeted therapy.	[73]
Immune Evasion	VEGF, PIGF	none	mannose chitosan/PAH- Cit,PC NP	PEG- TAM in the TME ML	cancer cells + M2	Orthotopic, and i.v. injected met. (4T1)	Two siRNAs worked better than one in inhibiting both primary tumor and its lung mets; induced higher T cell activity.	[78]
	MIF	none	none	glucan NP	cancer cells + M2 TAM in the TME (i.t.)	Orthotopic tumor model with spontaneous lung metastasis (4T1)	Local treatment inhibited both primary and lung metastasis.	[84]
	CCR2	none	none	PEG-PLA NP	Inflammatory monocytes throughout the body (i.v.)		Decreased CCR2 in monocytes in blood, spleen, and bone marrow; Inhibited growth of primary/met tumors.	[79]
	PITPNM 3	none	CD4 aptamer	Conjugate	CD4+ T cells throughout the body (i.p.)	Orthotopic tumor model (MDA-MB-231) with lung metastasis, humanized mice	Successfully decreased the number of intratumoral regulatory T cells; inhibited growth of primary/met tumors.	[80]
	PD-1 or PD-L1	none	none	lipid-coated CaP NP	cancer cells (PD-L1) and T cells (PD-1)	in vitro (MCF7)	Knocking down both PD-L1 in cancer cells and PD-1 in T cells provided the most effective killing.	[81]
	CD73	none	none	Chitosan-lactate nanoparticle	cancer cells (i.v.)	Ectopic tumor model (4T1) with lung metastasis	Knocking down CD73 in tumors inhibited tumor growth and also synergized with DC vaccine activity.	[83]

Cancer metastasis

Metastasis or the spread of breast cancer cells to other organs remains the underlying cause of death in the majority of breast cancer patients who succumb to the disease, and thus the central clinical challenge of breast cancer treatment [45, 46]. The metastatic cascade is a complex process encompassing many steps leading to cancer cell dissemination, including: 1) loss of cellular adhesion, 2) increased motility and invasiveness of cancer cells through ECM, 3) intravasation and entry into the circulation, 4) exit into a distant tissue (extravasation), and 5) colonization in a new foreign environment [47]. Disseminated cancer cells get released since early stages of tumors. However, the rate-limiting step is survival and outgrowth of cancer cells in different tissues. Metastatic traits can be driven over time by clonal evolution (e.g., by selective pressure and genetic diversification) of both cancer cells and stromal cells, that together enable successful extravasation and colonization of ‘seed’ cancer cells in distant organs, as reviewed elsewhere [48, 49]. Common metastatic sites for breast cancer include bone, brain, liver, lung, distant nodal, and pleural effusion [26], and different subtypes of breast cancer have distinct tropisms. Bone is the most common metastatic site for all subtypes, except basal-like TNBC. In basal-like TNBC, the more common metastatic sites are brain, lung, and distant nodal, while the less common sites are liver and bone. HER2+ breast cancer also has a higher rate of brain, liver, and lung metastases than ER/PR+ breast cancer. In recent years, several research groups have sought to develop siRNA therapeutics to intercept several biological pathways responsible for metastasis.

Kinases and phosphatases are opposing enzymes that catalyze protein phosphorylation and dephosphorylation, respectively. Signal transduction via these phosphorylation/dephosphorylation cycles is deemed a hallmark of cell signaling [50], including metastasis. While kinases are considered druggable and among the most prevalent drug targets, phosphatases have been considered undruggable [51]. Their undruggable nature is due to the highly conserved active site across phosphatases, compromising the inhibitor’s selectivity among closely related family members [52]. Furthermore, the positively charged active site calls for negatively charged inhibitor molecules, which usually lack cell permeability. siRNA-based therapeutics can overcome these limitations. For example, Pavan et al. performed an siRNA screen on 714 kinases and 272 phosphatases to identify potential therapeutic targets important for the epithelial-to-mesenchymal transition (EMT) process required for the metastasis of breast cancer cells [53]. They monitored major cellular changes (e.g., formation of focal adhesion and stress fiber, and fibronectin expression) of siRNA-treated breast cancer cells upon an EMT triggered by TGF-p. They identified the MEK5-ERK5 axis as an important effector of the TGF-β-mediated EMT. In particular, stably knocking down MEK5 and ERK5 in 4T1 cells with short hairpin RNA (shRNA) significantly decreased migration and invasion in vitro. shMEK5- and shERK5–4T1 cells were also orthotopically implanted into the mammary fat pad of mice. While there was no difference in the growth of primary tumors when compared to 4T1 containing non-targeting shRNA, mice with shMEK5- and shERK5–4T1 tumors had a substantially lower number of lung metastases. Parmar et al. have utilized a polymeric nanoparticle based on lipid-substituted polyethylenimine (PEI) and hyaluronic acid (HA) to screen six phosphatases (PPP1R7, PTPN1, PTPN22, LHPP, PPP1R12A and DUPD1), which have been reported in prior works to be involved in the metastasis of several types of cancer [54]. siRNA against PTPN22 appeared to have the strongest anti-migratory effects in vitro without inhibiting cell growth.

Transcription factors regulate gene expression (at transcription level) in cancer cells and have been of interest owing to their roles in regulating production of several proteins involved in metastasis [55]. With few recent exceptions, transcription factors are generally undruggable [56], highlighting the potential of siRNA technology. Yu et al. reported that blocking NF-kB pathway inhibited expression of matrix metalloproteinase-9 (MMP-9) [57], which can degrade basement membranes and extracellular matrices, facilitating intravasation and extravasation of cancer cells [47]. Blocking NF-κB pathway thus inhibited migration and invasion of breast cancer cells (4T1 cells) [57]. They utilized a micelle based on a poly(ethylene glycol)-block-poly(aminolated glycidyl methacrylate)-block-poly(2-(diisopropyl amino) ethyl methacrylate) (PEG-b-PAGA-b-PDPA) triblock copolymer to deliver siRNA against NF-kB subunit p65 (sip65) [57]. sip65 delivered by this micelle was found to effectively inhibit lung metastasis of the orthotopicallly implanted 4T1 tumors. In another work, Tang et. al. developed a micelle based on polyethylenimine-block-poly[(1,4-butanediol)-diacrylate-β−5-hydroxyamylamine] (PEI-PDHA) and polyethylene glycol-block-poly[(1,4-butanediol)-diacrylate-β−5-hydroxyamylamine] (PEG-PDHA) [58]. The micelle co-delivered paclitaxel and siRNA against Twist (siTwist), another transcription factor that regulates the EMT. The material is enzyme-pH dual sensitive. In particular, PEG can be cleaved from the material when exposed to matrix metalloproteinase enzymes, which are overexpressed in tumors with high metastatic potential. DePEGylation yields a positively charged PEI-nanoparticle that has enhanced cellular uptake. Upon uptake, PDHA is protonated in endo/lysosomes, mediating rapid drug release. Tang et al. showed that the micelle co-delivering paclitaxel and siTwist effectively inhibited the growth of 4T1 mammary tumor in vivo [58]. However, since scrambled siRNA-loaded nanoparticle was not used as a control, it is unclear to what extent the control of primary tumor growth and metastasis in vivo was contributed by siTwist or the paclitaxel on the nanoparticle.

C-X-C chemokine receptor type 4 (CXCR4), a transmembrane receptor protein on cancer cells, also has an established role in regulating metastasis [59]. In particular, CXCL12 chemokine (a CXCR4 ligand) interacts with CXCR4 to mediate chemotactic invasion of cancer cells to secondary metastatic sites [60]. CXCR4 is druggable by antibodies or other molecules. Recent works have incorporated CXCR4-targeted compounds on the nanoparticle that also simultaneously co-delivers siRNA. For example, Zhou et al. have combined siRNA against vascular endothelial growth factor (VEGF) with CXCR4-binding cyclam on the same polyplex to inhibit cancer cell (4T1 cell) invasion [61]. This design is based on the rationale that anti-VEGF therapies typically upregulate expression of the CXCR4 chemokine. In vivo, 4T1 cells implanted orthotopically in the breast of mice also metastasize to lung. Zhou et al. have reported that the delivery of the CXCR4 antagonist by their polyplex dramatically reduced the number of lung metastasis, compared to the saline control. While siVEGF further reduced growth of the primary tumor, it did not appear to further enhance the efficacy of CXCR4 inhibitor in controlling metastasis. Similarly, Guo et al. utilized a liposome conjugated with CXCR4 antibody to deliver siRNA against lipocalin-2 (Lcn2), a glycoprotein involved in migration and metastasis [62]. They have shown that both CXCR4 antibody and siLcn2 delivered by liposome was superior at inhibiting cancer cell migration in vitro than liposome containing only CXCR4 antibody or siLcn2.

In addition to targeting the specific oncogene, attempts have been made to therapeutically target non-coding RNAs that regulate several oncogenic pathways. For example, Vaidya et al. have reported use of nanoparticle based on amino-lipid, PEG, and RGD targeting peptide to deliver siRNA against Differentiation Antagonizing Non-Coding RNA (DANCR) for TNBC treatment [63]. DANCR knockdown by siDANCR-NP epigenetically repressed expression of several cancer-driven pathways, such as Wnt signaling, EMT, and phosphorylation of several kinases. Indeed, siDANCR-NP effectively inhibited migration and invasion of cancer cells in vitro and tumor growth in vivo. Interestingly, they reported heterogenous roles and broad functions of DANCR in different TNBC cell lines, warranting further mechanistic investigation (e.g., in tumor samples from patients). The oncogenic long non-coding RNAs are undruggable by conventional molecules and thus another promising avenue that can take advantage of siRNA technology.

Beyond siRNA cargo itself, we have reported that mesoporous silica nanoparticles (MSNPs) have an antioxidant (reactive oxygen species (ROS) scavenging) property and can intrinsically control cancer migration, invasion, and metastasis [64]. In particular, MSNPs delivering scrambled siRNA was shown to inhibit migration and invasion of TNBC cells in vitro, not attributable to the cytotoxicity of the nanoparticle or the siRNA against polo-like kinase 1 (siPLK1) it delivered. Besides, we found that lipid-based delivery of siPLK1 did not affect cancer cell migration, while still eliciting cytotoxicity to cancer cells. When we used our MSNP to deliver siPLK1 systemically in mice, we achieved effective control of both tumor burden and tumor spread. ROS are prevalent in tumors and can often be further increased by radiation or chemotherapy. While ROS can be useful in killing cancer cells, ROS and treatments that induce ROS were found to also promote cancer metastasis [65, 66]. For a more detailed overview of mechanisms of ROS in cancer (e.g., their roles in regulating MMPs, NOX, etc.), the readers are referred to our prior review article [47].

Drug Resistance

Drug resistance is another major challenge in breast cancer treatment, resulting in shorter progression free survival or recurrence. Utilizing siRNA to overcome drug resistance has been widely researched. Co-delivery of drug and siRNA against several gene targets that mediate drug resistance mechanisms on the same nanoparticle is a reasonable and extensively explored strategy.

Meng et al. have utilized MSNPs coated with PEI-PEG to co-deliver siRNA against the P-glycoprotein (Pgp) drug exporter and doxorubicin in a multidrug resistant (MDR) human breast cancer xenograft (i.e. MCF-7/MDR) [67]. Pgp in cancer cells can pump out chemotherapeutic drugs form cancer cells, resulting in low intracellular drug accumulation. Indeed, Meng et al. have reported higher intracellular doxorubicin content when it was delivered by the nanoparticle, and even more so when co-delivered with siPgp on the nanoparticle [67]. In vivo, systemically delivered nanoparticles carrying both siPgp and doxorubicin worked better than nanoparticles carrying a single cargo. Similarly, Deng et al. have reported on a doxorubicin-loaded liposome, coated layer-by-layer with poly-L-arginine (PLA) and siRNA against multidrug resistance protein 1 (MRP1) [68]. The outermost layer is hyaluronic acid for both prolonging in vivo circulation half-life and targeted delivery via its interaction with CD44 on TNBC cells. Further, Gong et al. have utilized a polypeptide micelle based on a lipoic acid conjugated with cytosol localization and internalization peptide (CL peptide) to co-deliver docetaxel and siRNA against autophagy-related 7 (ATG7) protein [69]. Chemotherapy-induced autophagy is one way cancer cells cope with external stress and could contribute to chemoresistance. In MCF-7 tumor model, siATG7 delivered by the nanoparticle showed no therapeutic benefit by itself, but incorporation of siATG7 on docetaxel-loaded nanoparticle appeared to improve its efficacy. Another example is the regulation of intracellular calcium level, which promotes cancer proliferation and survival during drug treatment. Wang et al. have utilized mesoporous/hollow silica nanocapsules to co-deliver siRNA against T-type Ca²⁺ channels and doxorubicin (DOX) to overcome drug-resistant breast cancer [70]. This combination showed a synergistic therapeutic effect on drug-resistant breast cancer cells MCF-7/ADR, but only an additive effect on the drug-sensitive MCF-7. Intratumoral injection of this siRNA/drug co-loaded nanocapsule better controlled the growth of tumors in mice than nanocapsules carrying a single component.

In addition to its role in metastasis as discussed in the prior section, CXCR4 was also shown to mediate breast cancer’s resistance to cisplatin. Yu et al. have shown that delivering siCXCR4 by chitosan nanoparticle sensitized breast cancer cells to cisplatin in vitro [71]. In another study, metadherin (MTDH) was shown to promote resistance of paclitaxel in breast cancer cells [72]. They also reported that MTDH overexpression in breast cancer patients’ tumors significantly correlated with poorer survival. Yang et al. thus utilized PEI-PLGA nanoparticle to co-deliver siMTDH and paclitaxel on the same particle to achieve a superior anti-tumor (MCF-7) efficacy than NP delivering a single component [72].

Beyond the undruggable targets, we found that siRNA is also superior to antibodies or small molecule inhibitors when targeting the same HER2 protein in drug-resistant HER2+ breast cancer cells [73–75]. This is because siRNA halts protein production, while antibodies or small molecule inhibitors merely bind the protein to prevent its functioning. To more completely inhibit HER2 activity in HER2-positive breast cancer, a combination of two HER2-targeted antibodies (trastuzumab and pertuzumab) that block different subunits of HER2 protein is needed [22]. Even then, resistance is still common. In our prior works, we have shown that using siRNA to knock down HER2 (siHER2) had better efficacy than a HER2-targeted antibody (trastuzumab) and a small molecule inhibitor (lapatinib) in cells [73–75]. In addition, our nanoparticle delivering siHER2 also inhibited growth of tumors that are resistant to HER2-targeted therapies used in clinics [73, 74]. We also reported that the cells treated long-term with siHER2 were less prone to develop therapeutic resistance than those treated long-term with an antibody or a small molecule inhibitor against HER2 protein [75]. This suggests that siRNA can also offer advantages for established druggable targets.

Immune evasion

Immunotherapy has shown impressive clinical success in subsets of patients. However, a larger subset of patients have experienced minimal or no clinical benefit when provided with the same treatment [76]. Understanding the ability of tumors to evade immune response and the development of therapeutics to overcome it are thus of great interest. It is now appreciated that malignant tumors contain not only neoplastic cells, but also non-cancerous host cells that support progression, metastasis, drug resistance, and importantly, immune evasion mechanisms (Fig. 1) [4, 5]. Consequently, components in the tumor microenvironment (TME) are attractive therapeutic targets. However, targeting cancer cells along with the multiple components in the TME are not always feasible because it would require many drugs, potentially increasing toxicity. Moreover, not all promising targets can be drugged by antibodies or small molecule inhibitors. Delivering various siRNAs with nanoparticles to the same tumor to knock down multiple genes at once thus has great promise to meet this demand.

Among immune cells in the TME, macrophages have received the most attention since M2 Tumor Associated Macrophages (TAMs) have been reported to occupy up to 50% of the breast tumor mass [77]. The number of TAMs in tumors also correlated with the prognosis of multiple cancers, including breast cancer. Song et al. have co-delivered siRNAs against vascular VEGF and placental growth factor (PIGF) to both M2-TAMs and breast cancer cells in the tumor by using a polymeric nanoparticle based on PEG- and mannose-doubly modified trimethyl chitosan and citraconic anhydride grafted poly (allylamine hydrochloride) [78]. The material homed to tumors upon systemic delivery by the EPR effect. In tumors, the low pH condition triggers the cleavage of PEG exposing mannose and cationic charge on the particle for enhanced delivery to macrophage and cancer cells in the tumor, respectively. Both siVEGF and siPIGF successfully and synergistically suppressed breast cancer cell proliferation and at the same time repolarized M2 macrophages to M1 macrophages. Together, the treatment effectively inhibited orthotopically implanted 4T1 breast tumors along with their lung metastasis.

On the other hand, Shen et al. took a different approach to modifying the TME [79]. Instead of targeting macrophages in the TME, Shen et al. utilized cationic nanoparticles to deliver siRNA against C-C chemokine receptor type 2 (CCR2) to inflammatory monocytes (precursors of macrophage) throughout the body. CCR2 is expressed on inflammatory monocytes and mediates monocyte recruitment to the TME via C-C chemokine ligand 2 (CCL2) expressed by cancer cells. PEG-PLA nanoparticles with varying charges were evaluated, and Shen et al. reported that higher number of cationic nanoparticles got taken up by CD11b⁺Ly6C^hi monocytes than neutral nanoparticle counterparts. To disrupt the CCR2-CCL2 axis of monocyte recruitment to tumors, siCCR2 delivered by cationic polymer nanoparticle substantially decreased CCR2 expression in monocytes isolated from blood, spleen, and bone marrow. This treatment also resulted in reduced macrophage levels in the TME and higher recruitment of cytotoxic CD8+ T cells, compared to the scrambled siRNA counterpart or siCCR2 delivered by neutral polymeric nanoparticles. No changes in body weight throughout the treatment regimen suggest a preliminary good systemic safety profile.

Besides monocytes and macrophages, attempts were also made to manipulate other cellular components in the TME. For example, Su et al. have reported that the TAM-produced CCL18 chemokine recruits circulating naive CD4+ T cells to the tumor by binding to their PITPNM3 receptors [80]. They also showed that intratumoral regulatory T cells (Treg) are mainly developed in the tumor from the recruited CD4+ T cells, as opposed to the recruitment of Treg from the peripheral blood. Therefore, they utilized CD4 aptamer-conjugated siRNA against PITPNM3 to systemically knock down PITPNM3 in CD4+ T cells throughout the body. In humanized mice bearing TNBC, daily i.p. injection with CD4 aptamer-siPITPNM3 substantially reduce the number of intratumoral Tregs, primary tumor growth, and lung metastases. In another example, Wu et al. have reported lipid-coated calcium phosphate nanoparticles to deliver siPD-1 to T cells and siPD-L1 to breast cancer cells (MCF7) and have shown in an ex vivo experiment that knocking down both molecules promoted the cytotoxic effect of T cells when co-incubated with cancer cells [81]. The effect is also better than knocking down either PD-1 or PD-L1 alone.

Soluble factors in the TME also significantly contribute to cancer’s immune escape. Adenosine is one of the major immunosuppressive factors, produced by CD39 and CD73 enzymes [82]. These enzymes are expressed by both cancer and immune cells in the TME. Jadidi-Niaragh et al. employed chitosan lactate nanoparticles to deliver siRNA against CD73 in the tumor in mice [83]. Successful CD73 knockdown in cancer cells, primary breast tumor growth inhibiton, and decreased number of lung metastases were reported. Importantly, siCD73-NP was found to synergize with the anti-tumor activity of tumor lysate-pulsed DC vaccine when given in combination.

While the aforementioned studies performed immune analysis in tumors or organs, it is not always apparent to what degree the effect in reducing tumor burden is from the activated immune system as opposed to direct cytotoxic effect of therapeutics on cancer cells. Thus, using antibodies to deplete immune cells (e.g., CD8, CD4), as conventionally performed in immunology research may shed light on this. Although indirectly, one may also utilize a bilateral tumor mouse model, wherein only one of the two tumors is intratumorally injected, and systemic immune effects are observed in the contralateral untreated tumor. For example, Zhang et al. have utilized a glucan-based cationic nanoparticle to deliver siRNA against macrophage migration inhibitory factor (MIF) in mice inoculated with 4T1 breast tumors on left and right mammary fat pads [84]. The siMIF-NP treatment was injected directly into only the left tumor, and the effect in both left and right (uninjected) tumors were monitored. They have reported decreases (though not statistically significant) in the burden of the uninjected tumor, suggesting that locally manipulating immunosuppression in the tumor caused some degree of systemic antitumor immune effect. In a single tumor model, they have reported that knocking down MIF in the tumor reduced CD206 level (M2 macrophage marker) and increased MHCII level (important for antigen presentation to CD8+ T cells) of the TAM. As a result of reduced immunosuppression, they have reported a higher number of infiltrating CD8+ T cells in the tumor treated with siMIF-NP.

Conclusion

Targeted therapy by monoclonal antibodies and small molecule inhibitors has substantially improved the prognosis of cancer patients in the past few decades. However, there is significant room for improvement. Monoclonal antibodies can target only accessible proteins (e.g., surface or secreted proteins). Small molecule inhibitors can target a wider range of proteins, but the off-target toxicity is often unavoidable. Treatment toxicity in patients is exacerbated by the combination of drugs typically required for effective control of cancer, especially in advanced stages. Further, there are a vast number of targets that cannot be feasibly drugged by antibodies or small molecule inhibitors. siRNA can overcome these shortcomings owing to its established mechanism of action and its capability to knock down the expression of any protein precisely, opening doors to many novel therapeutic possibilities.

Herein, we present the recent efforts in targeting other cancer hallmarks beyond the conventional cancer cell proliferative and apoptotic targets. They include metastasis, drug resistance, and immune evasion, regulated by proteins that emerge as novel therapeutic targets with advancement in cancer research and systems biology. Research has also shown that cancer hallmarks often intersect and synergize to a cancer’s benefit. For example, cancer cells undergoing the EMT process in metastasis are also drug-resistant. This type of convergence is also observed with common components in the tumor microenvironment, such as macrophages. Macrophages have been reported to mediate cancer’s metastasis, drug resistance, and immune evasion [85]. Certain chemotherapies were reported to enrich cancer cells with immune evasive phenotypes (e.g., CD47⁺ CD73⁺ PDL1⁺), explaining one of the drug resistance mechanisms [86].

Recently, the role of the immune response in cancer has been well recognized, leading to many successful immunotherapeutics, such as immune checkpoint inhibitors. However, a large fraction of cancer patients do not respond to these immunotherapeutics, which is partly due to the immunosuppressive tumor microenvironment. Nanoparticle-mediated delivery of siRNAs can be designed to target both cancer and TME components concurrently without significant added off-target toxicity found in certain combination of small molecule drugs.

Like other therapeutics, siRNA is often evaluated in immunodeficient hosts, necessitated by the use of human tumor xenografts. However, evaluating siRNA therapeutics in mice with intact immune systems may provide important information that better predicts the safety and efficacy in patients. In particular, off-target toxicity may be overlooked since siRNA against human protein would have no effect on mouse protein in the non-targeted host mouse cells. Likewise, the efficacy of siRNA therapeutics may be inaccurately predicted in the immunodeficient mouse model. For example, siRNA may kill cancer cells, resulting in tumor antigen release that triggers systemic adaptive immune response against tumors, further promoting the effect of siRNA therapeutics. Alternatively, siRNA may have other actions on immune cells systemically or in the TME that either promote or weaken the anti-tumor immunity. While use of humanized mice (with intact immune response) implanted with human tumors is attractive, the cost is too prohibitive by many for routine evaluation. Thus, use of a syngeneic tumor model in immunocompetent mice in addition to xenografted immunocompromised mice should be considered for evaluating and translating novel siRNA therapeutics.

While many recent focus areas of siRNA therapeutic development have been on undruggable targets, siRNA also has advantages in the druggable target arena. First, when targeting the same protein/pathway (e.g., HER2), siRNA was found to overcome cancer resistance to the antibody and small molecule inhibitor against HER2. Further, a cocktail of siRNAs targeting several pathways can be delivered simultaneously with a fewer limitations than combinations of small molecule inhibitors, which often pose toxicity issues due to their cumulative off-target effects. Lastly, although this review focuses on breast cancer as a case study, the concept extends to other types of cancer. siRNA technology holds great promise in cancer therapy because it can rapidly translate the discovery of any novel cancer pathways to drugs for cancer patients. Effective targeted delivery of siRNA thus has great potential to revolutionize cancer treatment.