Mitochondria as a Novel Target for Cancer Chemoprevention: Emergence of Mitochondrial Targeting Agents

Abstract

Cancer chemoprevention is the most effective approach to control cancer in the population. Despite significant progress, chemoprevention has not been widely adopted because agents that are safe tend to be less effective and those that are highly effective tend to be toxic. Thus, there is an urgent need to develop novel and effective chemopreventive agents, such as mitochondria-targeted agents, that can prevent cancer and prolong survival. Mitochondria, the central site for cellular energy production, have important functions in cell survival and death. Several studies have revealed a significant role for mitochondrial metabolism in promoting cancer development and progression, making mitochondria a promising new target for cancer prevention. Conjugating delocalized lipophilic cations such triphenylphosphonium cation (TPP⁺) to compounds of interest is an effective approach for mitochondrial targeting. The hyperpolarized tumor cell membrane and mitochondrial membrane potential allow for selective accumulation of TPP⁺ conjugates in tumor cell mitochondria versus those in normal cells. This could enhance direct killing of pre-cancerous, dysplastic, and tumor cells while minimizing potential toxicities to normal cells.

1. Introduction

1.1. Role of mitochondria in cell growth and death

Mitochondria efficiently produce ATP via cellular respiration that is essential to fulfill cellular bioenergetic needs. The mitochondrial electron transport chain (ETC) generates a transmembrane proton gradient that is used to generate ATP (1). Some electrons may be prematurely shunted to O₂, mainly by ETC complexes I and III, which results in the generation of superoxide (O₂^•−). Mitochondrial superoxide dismutase can dismutate O₂^•− to hydrogen peroxide (H₂O₂) (1). Therefore, mitochondria are well recognized for their potential to generate reactive oxygen species (ROS). In addition, many mitochondrial TCA cycle metabolites can be used as building blocks to produce nucleotides, amino acids, lipids, heme, and others. Mitochondria also have important roles in redox signaling.

Mitochondria can also initiate the intrinsic pathway that promotes apoptosis and necrosis-like non-apoptotic cell death. Once receiving proapoptotic stimuli, such as from Bax and Bak proteins, Ca²⁺ overload, or other signals, mitochondrial outer membrane permeabilization (MOMP) is induced. Once MOMP is initiated, cytochrome c is irreversibly released from the mitochondrial intermembrane space to the cytosol (2), where it facilitates apoptosome formation, which acts as a platform to activate caspase-9 and the resulting cascades that promote apoptosis (2). Once the mitochondrial permeability transition pore complex (mPTPC) is opened, large amounts of solutes enter the mitochondria matrix, resulting in osmotic swelling which can lead to rupture of the inner and outer membranes, accompanied by loss of mitochondrial membrane potential (ΔΨm), decreased ATP generation and excess ROS production (2).

1.2. Mitochondrial metabolism is required for cancer development

Long ago, Otto Warburg and colleagues observed that cancer cells ferment large amounts of glucose to lactate even in the presence of oxygen (3). In the 1970s, Efraim Racker named this phenomenon ‘aerobic glycolysis’ or the ‘Warburg Effect’ (4). The ‘Warburg Effect’ led to historical viewpoints that cancer cell mitochondria are dysfunctional and that these defects in cancer cell mitochondria render them inconsequential for cancer development. While glycolysis plays an important role in tumorigenesis, mitochondria provide ATP and building blocks (for pyrimidine, amino acid and heme biosynthesis) for cancer cells (5), increasing evidence supports a pivotal role for mitochondrial function in cancer viability. For example, tumor cells lacking mitochondrial DNA (mtDNA) display significantly slower growth and lower tumorigenesis potential (6–8).

Mutations that inactivate tumor suppressor genes and activate oncogenes also alter mitochondrial metabolism. For example, amplification of the Myc oncogene leads to activation of genes that are essential for mitochondrial biogenesis and function (9). Loss of the retinoblastoma tumor suppressor retinoblastoma (RB1) induces mitochondrial protein translation and increases oxidative phosphorylation (OXPHOS) (10). A consequence of increased OXPHOS is the accumulation of low levels of mitochondrial ROS which have been shown to contribute to neoplastic transformation, cell proliferation and survival (11). Some cancer subtypes (e.g. sarcoma, liver cancer, lung cancer and skin cancer) rely predominantly on OXPHOS for ATP production (12). Fatty acid beta-oxidation (FAO), a series of multi-step catabolic reactions that shorten fatty acids, generates NADH and FADH₂ that shuttle into mitochondria to support the ETC. FAO enzymes are dysregulated in cancer; elevated key FAO enzymes and/or high FAO activities are seen in multiple cancer types, including triple negative breast cancer, ovarian cancer, glioma, and mutant KRAS-driven lung cancer (13). FAO has been implicated in contributing to cancer cell survival and proliferation, metastasis, drug resistance (13) and cancer stemness (14). Although a majority of cancer cells have functional mitochondria, some cancers have defects that result from mutations in mtDNA or nuclear DNA that encode mitochondrial proteins (15). Interestingly, malfunctioning mitochondria can ultimately contribute to the formation and progression of cancer via ROS-induced nuclear gene instability, a process called retrograde signaling (15). Collectively, these findings indicate that mitochondrial metabolism plays a critical role in tumorigenesis and cancer development.

1.3. Mitochondrial targeting using ligand-drug conjugates

There are basic two strategies for mitochondrial targeting, non-peptide based and peptide- and amino acid-based. The non-peptide based strategies include delocalized lipophilic cations, and others such as benzothiazepines, modified phospholipids or sulfonylurea-related compounds (16).

1.3.1. Mitochondrial targeting using delocalized lipophilic cations

Compared to normal cells, tumor cells have notable transformations in bioenergetics, biosynthesis, and modulated signal transduction which support non-stop growth (17). For some time, scientists have been working on mitochondrial-targeted agents with high selectivity for cancer cells versus normal cells. The mitochondrial membrane potential in cancer cells (–220 mV) is more hyperpolarized than that in normal cells (–140mV) (18). This discrepancy can be exploited with compounds such as delocalized lipophilic cations (DLCs) that selectively accumulate in cancer cell mitochondria (19). The mitochondrial membrane potential (negative inside) can drive a 100- to 1000-fold uptake of cations. Another advantage of lipophilic cations is that their lipophilicity promotes their ability to cross the plasma membrane and the mitochondrial outer and inner membranes. Moreover, the altered redox homeostasis and elevated ROS production in cancer cells (20) renders them potentially more vulnerable to ROS-generating agents (including DLCs), and/or could further weaken their cellular antioxidant defense systems (Figure 1).

Figure 1. The basis for mitochondrial targeting using mito-compounds for cancer cell killing.

(A) Cancer cells have a more hyperpolarized membrane, which serves to drive the uptake of TPP⁺-conjugated compounds up to 100–1000-fold compared to extracellular compounds. Moreover, elevated ROS production in cancer cells renders them potentially more vulnerable to be killed by ROS-generating agents (like TPP⁺-conjugated compounds). (B) Normal cells have a less hyperpolarized mitochondrial membrane and comparatively lower levels of ROS, which leads to less uptake of mito-compounds and insufficient ROS to initiate cell death.

Mitochondria-targeting DLCs include triphenylphosphonium cation (TPP⁺), Rhodamine 123 and dequalinium (DQA). These lipophilic cations readily accumulate in the mitochondrial matrix (21, 22). Originally, these DLCs were research tools to study mitochondrial structure and metabolism, and to monitor mitochondrial membrane potential (23, 24). Ligation of these DLCs to various parental compounds for de novo drug design led to the development of mitochondria-targeting antioxidants (25). Linking anti-cancer agents with mitochondria-targeting moieties was done to improve anticancer efficacy, minimize multidrug resistance (MDR), and damage cancer cell mitochondria to trigger cell death (26).

Triphenylphosphonium cation (TPP⁺) is the best characterized and most widely used lipophilic cation among the nonpeptide-based strategies for delivering conjugated compounds into mitochondria. TPP⁺ contains a positively charged phosphorus surrounded by three lipophilic phenyl groups. Due to greater uptake of lipophilic cations into cancer cells, phosphonium salts by themselves have some anti-proliferative activities by disrupting mitochondrial membrane integrity and inhibiting respiration in several cancer cell lines in vitro and in human ovarian cancer models in vivo (27, 28). However, the effects of these TPP⁺ cations are limited given that their uptake into mitochondria is reversible (29). More importantly, TPP⁺ salts themselves don’t have anti-tumor effects. However, there are many advantages to using TPP⁺ as a ligand to deliver other compounds into cancer cell mitochondria. TPP⁺-based carriers are relatively simple to synthesize and purify, and they have low reactivity towards their conjugates.

TPP⁺ has been used to deliver various potential anticancer compounds including, but not limited to, antagonists of heat shock proteins, polyphenolic compounds, metabolic modulating agents, triterpernoids, and others (30–33). Parental drugs can be conjugated to mitochondria-targeted ligands via various linkers including esters, amides, carbamates, carbonates, phosphate and others (34). These complexes may be given directly or can be encapsulated in nanoparticle or liposomal vesicles. Recently, their delivery using nanoformulation-based carriers has attracted interest due to potential superior pharmacokinetic properties and improved solubility (35), and relatively higher tumor accumulation (36). Due to the advantage of using TPP⁺ for mitochondrial-targeting and the significant success of this approach for cancer prevention and treatment, we will focus on TPP⁺-based conjugates in this review.

1.3.2. Mitochondrial targeting using peptide-drug conjugates

Tethering peptides or short sequences to bioactive compounds is another type of major strategy for mitochondrial targeting. Common examples include modified gramicidin S (GS)-based pentapeptide sequences(37–41), D-(KLAKLAK)₂ pro-apoptotic peptides (42) and Szeto-Schiller tetrapeptides (43).

Modified forms of antibacterial peptide GS have been successfully used as a mitochondrial-targeting moiety; a β-turn within the peptide motif inserts into the mitochondrial membrane and this insertion does not rely on the mitochondrial membrane potential. Examples of this approach include GS-based nitroxide ROS scavenger conjugates that can protect cells from oxidative damage and apoptosis and can prolong survival in a rat model of hemorrhagic shock (37, 38). One of these conjugates (XJB-5–131) suppresses DNA damage and motor decline in a mouse model of Huntington’s disease (41). A GS-based PARP inhibitor (XJB-veliparib) inhibits mitochondrial PARP which prevents hypoxia/ischemia-induced death in neurons (39). A GS-conjugated ROS generator, XJB-Lapachone, can stimulate mitochondrial ROS and promote autophagic cell death in cancer cells (40).

1.4. TPP⁺-based compounds for cancer chemoprevention

Chemoprevention is the use of agents to prevent cancer in healthy persons (Primary prevention), to prevent pre-cancerous lesions from becoming cancer (Secondary prevention), and to prevent cancer recurrence or metastasis (Tertiary prevention). It would be ideal to use them for primary or secondary prevention. However, their use in tertiary prevention could also be highly valuable as conventional cancer therapeutics are often unable to stop recurrence or metastasis. Chemopreventive efficacy is generally estimated by preclinical in vitro cell culture and in vivo animal studies followed by phase I clinical trials. Toxicity is determined in animals and clinical trials. An ideal chemoprevention agent should be highly effective, cause minimal side effects, and have known mechanisms of action. The selective accumulation of TPP⁺ conjugates in tumor cell mitochondria make them promising agents for chemoprevention with enhanced efficacy and minimal toxicity for normal tissues.

This review will summarize and recent published approaches in using mitochondrial-targeted TPP⁺ based-compounds (Figure 2 and Table 1) as well as modifications or improvements to classical mito-compounds (Table 2) for cancer chemoprevention and treatment. We will also discuss their mechanisms of action and propose future perspectives.

Figure 2.

Chemical structures of representative mitochondria-targeted compounds.

Table 1.

Features of direct conjugation of mitochondria-targeted compounds

Parental drug	Bond	Conjugates name	In vitro model and IC50 or other values	In vivo model and antitumor efficacy	Mode of actions	Reference
Pyriplatin	C-C	OPT	IC50 (A549): 8.7±1.6 μM;	A549 xenograft; Dosage: 5mg/kg BW; i.v.; TGII: > 50%	nDNA Damage, OXPHOS and glycolysis inhibition, apoptosis	(112)
Terpyridine platinum (II)	C-C	TTP	IC50 (CaoV): 9.0 ± 0.2 μM; IC50 (A549): 9.5 ±1.2 μM IC50 (A549/DDP): 24.1 ±1.4 μM	-	Mitochondrial TrxR inhibition, OXPHOS and glycolysis inhibition, ΔΨm decrease	(113)
Pt (IV) complexes	C-C	PDT	IC50 (A549): 12.20 ± 2.29 μM; IC50 (MCF7): 7.54 ±2.27 μM IC50 (HeLa): 6.42 ±1.03 μM	-	Apoptosis, ΔΨm decrease, OXPHOS and glycolysis inhibition	(115)
Chlorambucil	Amide	Mito-Chlor	IC50 (MCF7): 7.0 ± 1.2 μM IC50 (MIA PaCa-2): 1.6 ±0.9 μM IC50 (BXPC3): 2.5 ± 2.8 μM	MIA PaCa-2 xenograft; Dosage: 10mg/kg BW; TGII: 61.5%	mtDNA and nDNA Damage, cell cycle arrest, ROS production	(118)
Betulinic acid (BA)	Triethylene glycol	Mito-BA compound 9	IC50 (TET21N): 0.74 ± 0.14 μM IC50 (MCF-7): 0.70 ±0.11 μM	-	Apoptosis, OXPHOS inhibition	(93)
Glycyrrhetinic acid (GA)	Alkyl	Mito-GA compound 2f	IC50 (HeG-2): 7.58 ± 0.89 μM IC50 (A549): 5.25± 0.84 μM IC50 (MCF07): 7.23 ± 1.01 μM IC50 (HT-29): 8.71 ± 1.06 μM Inhibit A549 cell migration @ 5 μM	-	G2 cell cycle arrest, apoptosis, cell migration inhibition, ΔΨm decrease	(96)
Desi-iodo analog of PU-H71	C-C	PU-H71-TPP-2	Reduced the viability of AML cell lines (HL-60, THP-1 and MV-411)	-	OXPHOS inhibition, apoptosis, ATP decrease	(30)
Apoptozole	Ester	Az-TPP-O3	A panel of cancer cell lines with IC50 ranges from 0.5 to 1.5 μM	-	Mortalin ATPase activities inhibition, interaction between mortalin and p53 disruption, MOMP formation, apoptosis	(152)
Curcumin	Ester	Mito-curcumin	IC80 (A549): 5 μM	-	Mitochondrial TrxR2 decease, ROS overproduction, ΔΨm decrease, mtDNA damage, apoptosis	(69)
Honokiol	C-C	Mito-HNK	IC50 (H2030): 0.26 μM	H2030-BrM3 and DMS-273 orthotropic model; Dosage: 3.75 μmol/kg BW; oral gavage; TGII: 70% and 40%, respectively	ETC complex I inhibition, ROS overproduction, mitoSTAT3 phosphorylation suppression	(31)
Phenol	C-C	Phenol TPP-derivative 2	IC50 (A549): 0.55 ± 0.07 μM IC50 (MDA-MB-231): 0.34 ± 0.11 μM IC50 (U-02OS): 0.63 ± 0.10 μM	-	Mitochondrial mass decrease, nuclear PGC-1α levels decrease, electron transport rate decrease, ROS overproduction, ΔΨm decrease, apoptosis	(79)
Caffeic acid	Ester	Mito- caffeic acid	IC50 (HepG2): 9.19 ± 0.25 μM	-	mPTP open, ΔΨm decrease, apoptosis	(73)
Gentisic acid (GA)	C-C	GA-TPP+C10	IC50 (MCF-7): 2.42 ± 0.43 μM IC50 (MCF-7-rho0): 14.72 ± 0.37 μM IC50 (MCF-TAMR): 11.47± 0.30 μM	-	Mitochondrial ABC transporter increase, ETC complex I inhibition, αKGDHC activity inhibition, cell cycle arrest in G1 phase	(75)
Lonidamine	Aliphatic chain	Mito-LND	IC50 (H2030BrM3): 0.74 μM IC50 (A549): 0.69 μM	H2030-BrM3 orthotropic model; Dosage: 7.5 μmol/kg BW; oral gavage; TGIIprimary tumor: >40% TGIILN metastasis: >50% H2030-BrM3 brain metastasis model; Dosage: 7.5 μmol/kg BW; oral gavage	ETC complex I and II inhibition, ROS overproduction, AKT/mTOR/p70S6K axis disruption, autophagic cell death	(32)
Pyruvate dehydrogenase kinase 1 (PDK1) inhibitors	Alkyl	TPP-PDK1–1f	IC50 (MCF-7): 0.50 ± 0.11 μM IC50 (A375): 0.51 ± 0.09 μM IC50 (NCI-H1650): 0.21 ± 0.04 μM	-	PDK activity inhibition, pyruvate glycolysis decrease, ΔΨm decrease	(85)
Dichloroacetate	Ester bond	Mito-DCA	IC50 (PC-3): 17 ± 1 μM IC50 (DU145): 42 ± 2 μM IC50 (LNCaP): 30 ± 6 μM	-	Glycolysis decrease, cellular respiration and ATP decrease, anti-tumor immunity increase	(87)
Tamoxifen	C-C	Mito-Tam	IC50 (MCF-7): 1.25 μM IC50 (MDA-MB-436): 3.4 μM IC50 (SK-BR-3): 3.5 μM	NeuTL syngeneic model; Dosage:0.54 μmol/mouse/dose; oral gavage; TGII: 80% MCFHer2high model; Dosage:0.25 μmol/mouse/dose; oral gavage; Tumor didn’t grow	ETC complex I inhibition, ROS overproduction, tumor-derived senescent cells elimination	(99, 101)
Peptide nucleic acids (PNA)	Amide	PNAs-mito	IC50 (A549): 15.6 μM	A549 xenograft model; intratumoral injection; TGII: > 50%	Mitochondrial fusion disruption, ΔΨm, ATP decrease	(109)
Triethoxypropylsilane	Amide	Triethoxypropylsilane–TPP (1)	IC50 (SCC7): 3 μM IC50 (PC3): 10 μM IC50 (MCF-7/ADR): 21 μM	Xenograft model; Dose: 50 mg kg−1 in PBS; Tumor growth was inhibited	Mitochondrial biomineralization and morphology disruption, ΔΨm decrease, apoptosis	(156)
Furocoumarin Derivatives	C-C	PAPTP	Induce more than half of B16F10 cells death at 1 μM; Induce up to 85% of B-CLL cells death at 10 μM	Orthotopic mouse B16F10 melanoma model; Dosage:5nmol/g BW; i.p.; TGII: 90% when combine used with cisplatin Orthotopic xenograft human pancreatic ductal adenocarcinoma; Dosage: 5nmol/g BW; i.p; Tumor sizes were significantly reduced.	Kv1.3 channel inhibition, mitochondrial damage, ROS overproduction, apoptosis	(155)
Imidazolium salt	C-C	TPP1	For 1-hour of exposure: IC50 (RT112): 200 μM IC50 (SW780): 200 μM IC50 (UMUC3): 80 μM	-	ΔΨm decrease, apoptosis	(185)

Abbreviation: TGII: Tumor growth inhibition index; αKGDHC: α-ketoglutarate dehydrogenase complex; BW: body weight

Note: Targeting moieties are all TPP cations

Table 2.

Features of mitochondria-targeted compounds assembled in nano-formulations

Nanocarrier	Parental drug	In vitro model and IC50 value	In vivo model and antitumor efficacy	Mode of actions	Reference
PLGA-b-PEG NPs	Platin-Cbl	IC50 (A2780/CP70): 1.0 ± 0.5 μM IC50 (PC3): 0.22 ± 0.04 μM	-	ΔΨm decrease, mitochondrial respiration disruption, citrate synthase activity inhibition	(119)
PLGA/CPT/ PEI-DMMA NPs	Doxorubicin	IC50 (MCF-7/ADR): 19.82 μM at pH 6.5	MCF-7/ADR xenograft; Dosage: 5mg/kg BW; i.v.; TGII: 84.9%	ΔΨm decrease, OXPHOS inhibition, mtDNA damage, apoptosis	(125)
Folate-cholesteryl albumin NPs	Docetaxel	IC50 (B16F10): around 0.1 μM IC50 (MCF-7): around 1 μM	MCF xenograft; Dosage: 5mg/kg BW; i.v.; TGII: 69.5%	ΔΨm decrease, apoptosis	(133)
NCs with stabilizing agent (TPP+-Brij98)	Paclitaxel	A mortality rate of 73.5% ± 3.7% (MCF-7/ADR): 50% B-TPP/PTX NCs @ 5 μg/ml; The diameters for MCF-7 ADR spheroid in control and 50% B-TPP/PTX NCs on days 5: 633.1 ± 17.8 μm vs. 416.0 ± 30.1 μm	-	ΔΨm decrease, apoptosis	(139)
DSPE-PEG liposomes	Resveratrol	IC50 (B16F10): around 10 μg/ml	-	ΔΨm decrease	(54)
PEG-NPs with pH-sensitive de-PEGylation	Quercetin	IC50 (HeLa and HepG2): 50 μM ~ 100 μM	H22 xenograft; TGII: 65.7 %	ΔΨm decrease, apoptosis	(59)

Abbreviation: NPs: Nanoparticles; NCs: Nanocrystals; PLGA: Poly (lactic-co-glycolic acid); PEG: Polyethylene glycol; CPT: Camptothecin; PEI: Polyethylenimine; DMMA: 2,3-Dimethylmaleic anhydride; DSPE: Distearoylphosphatidylethanolamine; BW: body weight

Note: Targeting moieties are all TPP cations, except for special indication

2. Mitochondria-Targeted Chemopreventive Agents

Known chemopreventive drugs and natural polyphenolic compounds have been intensively studied for their preventive and therapeutic effects on cancer. Their anticancer efficacies are largely attributed to their roles in cell cycle regulation, anti-proliferation, pro-apoptosis and modulating redox state. To enhance their potential chemopreventive effects, several compounds have been modified to enhance mitochondrial targeting. While several examples have direct evidence for cancer prevention, other examples represent in vitro and in vivo studies showing enhanced efficacy and decreased toxicity and strong potential for further testing.

2.1. TPP⁺-based natural phenolic compounds

2.1.1. Mito-Honokiol

Honokiol (HNK), a phenolic compound from magnolia bark, has chemopreventive/chemotherapeutic effects in many cancer types by suppressing tumor initiation, retarding malignant progression, reducing tumor load, preventing therapy-resistance and primary tumor metastasis (44). Its chemopreventive effects have been attributed to its potent anti-proliferative, anti-inflammatory, anti-angiogenic, anti-oxidative, cell cycle arrest, and anti-migration properties (44, 45). Recent studies have pointed to potential mitochondrial effects, with high-dose HNK decreasing cancer mitochondrial bioenergetics and inhibiting the development of squamous cell carcinoma in a mouse model (46). These findings led to the hypothesis that creating mitochondrial-targeted HNK could further increase its anticancer and chemopreventive efficacy and potency.

Mito-HNK was synthesized by linking TPP⁺ to HNK via a long alkyl chain (31). Mito-HNK is 100-fold more potent than HNK in suppressing mitochondrial bioenergetics and retarding cell proliferation/invasion of human lung cancer cells derived from brain metastases and a small cell lung cancer line with high metastatic potential. Mito-HNK causes pronounced inhibition of respiratory chain complex I which stimulates ROS generation (31). Further evidence of its mitochondrial effects were its promotion of mitochondrial Prx3 oxidation, its suppression of the phosphorylation of mitochondrial STAT3, and its inability to significantly inhibit cancer cells that lack mitochondria (31). In mice, Mito-HNK suppresses lung cancer development and brain metastases, whereas the same dose of HNK failed to have an effect (31). Since Mito-HNK was administered starting 1 day after engrafting metastatic lung cancer cells into the arterial circulation, its ability to suppress brain metastases indicates chemopreventive activity. Mito-HNK crosses the blood-brain barrier. RNA-seq analysis of brain metastatic lesions showed that Mito-HNK strongly induced mediators of cell death in the malignant cells but not in the non-malignant tumor stroma (31), and downregulated pathways involved in tumor invasion and proliferation. In the tumor stroma, Mito-HNK downregulated inflammatory and angiogenic pathways which typically support metastasis (31). Mito-HNK is effective against models with mutant EGFR and mutant KRas, suggesting efficacy regardless of driver mutation status. Similar to the good safety profile of HNK (47), Mito-HNK showed no toxicity in mice even when tested at ~20-fold higher than the effective dose (31). These findings make Mito-HNK a highly attractive agent for the prevention and treatment of lung cancer and its brain metastases, and warrant its further development.

2.1.2. Mito-resveratrol

Resveratrol, a phytoalexin in fruit skins, has diverse activities including anti-tumorigenic, anti-viral, cell-cycle inhibitory and cardiovascular and neurological protective effects, which make it ideal for chemoprevention (48). Although it has anti-cancer potential (49), its poor bioavailability and rapid metabolism are major impediments (50). Resveratrol was one of the first natural polyphenols to be modified for mitochondrial targeting (51, 52) which greatly improved its water solubility. These resveratrol show selectivity for fast-growing and tumoral cells (51, 52) and their effects on mitochondrial complexes I and III could contribute to ROS generation (53). Resveratrol liposomes modified with TPP-polyethylene glycol ((TLS (Res)) have increased blood circulation time, enhanced mitochondrial uptake (>5-fold), enhanced ROS generation and dissipation of the mitochondrial membrane potential, and enhanced anti-tumor effects (54). Further investigation of its in vivo activity is warranted.

2.1.3. Mito-quercetin

Quercetin, a natural polyphenolic flavonoid, has well-documented chemopreventive/chemotherapeutic properties in many tumor models (55, 56). Its mechanisms include redox-modulation, cell-cycle arrest, anti-metastatic and direct pro-apoptotic effects on many cancer cells via mitochondrial- and caspase-dependent pathways (55, 56). Low dose quercetin has selectivity for cancer cells versus normal cells (57). However, its poor water solubility is a major obstacle.

Mattarei et al. (58) synthesized a TPP derivative of quercetin (compound 7) which accumulated in mitochondria while maintaining its ability to inhibit mitochondrial ATPase. Compound 7 was markedly more effective on colon cancer CT26 cells, but did not affect slow-growing normal fibroblasts (58). To diminish its rapid clearance by the reticuloendothelial system, Xing et al. (59) developed novel self-assembled quercetin PEGylated nanoparticles (TPP-Que-NP) which enhances mitochondrial localization and disruption of mitochondrial membrane potential, increases caspase activation, and greatly improves anti-proliferative effects on four cancer lines (MCF-7, A459, HepG2, and HeLa) (59). In vivo, these quercetin nanoparticles showed greater tumor suppressive activity than doxorubicin in H22 tumor bearing mice, with no signs of toxicity (59).

2.1.4. Mito-curcuminoids

Curcumin, a major polyphenol in turmeric, has long been recognized for its chemoprevention potential for various cancer lines in vitro (60). However, its low bioavailability and limited intracellular accumulation limit its in vivo potential. Selective DLC-based mitochondria targeting (61) and nanoparticle-mediated delivery (62) may improve its potential. Linking curcumin with mitochondriotropic molecules greatly enhanced its potency for cancer cells including cell cycle arrest and apoptosis.

Recent attention has examined how mito-curcumin analogs alter redox homeostasis. Mammalian thioredoxin reductases (TrxRs) have central roles in maintaining redox homeostasis (63), and in promoting cell growth, preventing cell apoptosis and supporting angiogenesis (64). The TrxR/Trx system can greatly influence cancer development, progression and chemoresistance (65). Therefore, TrxRs are considered potential chemopreventive/therapeutic targets (65).

Curcumin irreversibly inhibits TrxR by alkylation of its active site (66). Qiu et al. (67) found that curcumin analogues also inhibit TrxR. One such analog (compound 2a) has 100-fold more TrxR inhibitory activity than curcumin, although its anti-tumor effect was not significantly improved. In order to enhance anti-tumor effects, compound 2a was tethered to TPP⁺ via a decyl linker to create TPP2a (68). TPP2a is a potent inhibitor of mitochondrial TrxR (IC₅₀ =1.44 μM) and promotes mitochondrial-dependent apoptosis (68) in multiple human cancer cells (A549 lung adenocarcinoma, HeLa cervical cancer, PC3 prostate cancer, MDA-MB-231 triple-negative breast cancer, and SW480 colon cancer) (IC₅₀ range 0.91–2.77 μM). Jayakumar et al. (69) found that mito-curcumin also greatly decreased the frequency of cancer stem cells (69); it increased ROS generation, decreased mitochondrial TrxR2 activity by >50%, decreased glutathione, and induced the loss of membrane potential. Overexpression of TrxR2 in cancer cells accentuated its anti-tumor effects (69). Mito-curcumin analogues are highly promising agents for primary and tertiary cancer prevention.

2.1.5. Mito-hydroxycinnamic acid derivatives (Mito-HCAs)

Hydroxycinnamic acids are plant phenolic acids with excellent chemopreventive potential in preclinical models (70). When administered at high dose, they may trigger mitochondrial malfunction and mitochondria-mediated cell death in various cancer lines (71, 72). Li et al. (73) synthesized a series of TPP-conjugated hydroxycinnamic acid derivatives (mito-HCAs), including p- coumaric (p- CoA), caffeic (CaA), ferulic (FA) and cinnamic acid (CA). At relatively high concentrations, these mito-HCAs selectively inhibit the proliferation of human hepatoma HepG2 cells with minimal effects on non-cancerous cells (73). Mito-caffeic acid dose-dependently caused fragmented mitochondria morphology, collapsed ΔΨm, and caused apoptotic cell death in cancer cells (73). Cyclosporin A, which inhibits the opening of the mitochondria permeability transition pore (mPTP), partially rescued cancer cells from mito-caffeic acid, which implied that mPTP could contribute to its anti-tumor activity. Unlike most of the mitochondria-targeted compounds discussed above, for which their pro-oxidant properties induce cell killing, these Mito-HCAs increased antioxidant enzymes (catalase, glutathione peroxidase) and decreased mitochondrial lipid peroxidation (73).

2.1.6. TPP⁺-based polyhydroxybenzoates

Breast cancer is a heterogenous disease with different subtypes. Sandoval-Acuña et al. (74) studied the anticancer effects of mitochondria-targeted polyhydroxybenzoates on breast cancer lines representing four cancer subtypes. They linked TPP⁺ with various decyl-polyhydroxybenzoates (SA-TPP⁺ C₁₀, GA-TPP⁺ C₁₀, PIA-TPP⁺ C₁₀, and PCA-TPP⁺ C₁₀). All of these compounds were effective on all breast cancer lines tested, which implied that estrogen receptor or HER2/neu status was inconsequential for efficacy (74). MCF-7 cells were more prone to apoptosis induction by these agents than MDA-MB-231 cells (74). The different basal metabolic states in these cells may account for this difference. Their efficacy corresponded with an initial loss of mitochondrial membrane potential, and a modest decline in ATP. These TPP⁺-polydroxybenzoates also inhibited the migratory ability of MDA-MB-231 cells, which are highly metastatic (74). Subsequent mechanistic studies (75) on GA-TPP⁺ C₁₀ demonstrated its time-dependent effects on mitochondrial bioenergetics, including a significant drop in Complex I dependent respiration, and late-stage inhibition of the α-Ketoglutarate Dehydrogenase Complex. Cells responded to this treatment with several adaptations that could promote survival, including AMPK activation, metabolic remodeling towards glycolysis, increased expression of for mitochondrial biogenesis and electron transport, and decreased expression of uncoupling proteins. Combined treatment with doxorubicin and GA-TPP⁺ C₁₀ showed synergistic effects against breast cancer cells (75). While additional in vivo studies are needed to to fully evaluate GA-TPP⁺ C₁₀’s chemopreventive potential, its parental compound gentisic acid has anti-neoplastic effects (76, 77) and may protect against the cardiac myofibrillary and endothelial damage induced by doxorubicin (78).

2.1.7. Phenol TPP-derivatives

Gazzano et al. (79) developed two phenol TPP derivatives of 4-hydroxybenzaldehyde (compounds 1 and 2). Both compounds were effective against various human cancer cell lines and had much higher IC₅₀ values for their normal cell counterparts. The differential effect was the greatest for human lung cancer A549 cells versus immortalized human bronchial epithelial cells (BEAS-2B) (79). Both compounds accumulated in lung cancer cell mitochondria, but not in BEAS-2B cell mitochondria. These TPP-phenols decreased mitochondrial mass, decreased levels of nuclear PGC-1α (the transcriptional regulator for mitobiogenesis), decreased electron transport, increased ROS generation, and triggered mitochondrial depolarization and caspase 3/9 dependent apoptotic cell death of tumor cells (79). These characteristics warrant further examination of their chemoprevention potential.

2.2. TPP⁺-based metabolic regulation agents

2.2.1. Mito-lonidamine

The potential of lonidamine (LND; 1-(2,4-dichlorobenzyl)-1H-indazole-3-carboxylic acid), in combination with other agents, to prevent and treat multiple cancers has been explored (80). Although it has the potential to reverse resistance and enhance sensitivity to standard-of-care therapies, LND by itself has limited anti-cancer efficacy. Historically, LND’s anti-neoplastic properties were proposed to result from effects on energy metabolism including glycolysis (81) and complex II inhibition (82). To enhance its mitochondrial effects, our group developed Mito-LND by linking LND with TPP⁺ via an aliphatic chain (32). In vitro, Mito-LND is 180- and 300-fold more potent than LND in blocking the growth of H2030BrM3 cells (lung cancer cells isolated from brain metastases) and A549 adenocarcinoma cells, respectively (32). In vitro, Mito-LND is ~100-fold more potent in suppressing invasion of highly metastatic lung cancer lines (32). Compared with LND, Mito-LND is 370- and 162-fold more potent at inhibiting mitochondrial complexes I and II, respectively. Mito-LND enhances ROS generation and the oxidation of mitochondrial Prx3, but not cytosolic Prx1, implying that the ROS is primarily mitochondrial. Mito-LND cause decreased phosphorylation of energy sensing proteins (AKT and P70S6K), and induces autophagy in lung cancer cells (32). In orthotopic models of lung adenocarcinoma, a low dose Mito-LND significantly reduced primary tumor growth and metastasis to lymph nodes, and suppressed the development of brain metastasis (32). Given the timing of Mito-LND administration, its ability to suppress brain metastases is consistent with chemopreventive activity. To the best of our knowledge, Mito-LND is the one of the least toxic mitochondria-targeted compounds, causing no toxic effects in mice that were treated for 8 weeks with doses 50 times above the effective anti-cancer dose. Mito-LND induced autophagy in vivo in mouse lung tumors and their brain metastases, but not in normal mouse lung and brain (32). The low toxicity and pronounced anti-cancer and anti-invasive effects of Mito-LND make it a highly promising agent for the tertiary prevention and treatment of lung cancer or its metastases to lymph nodes and the brain.

2.2.2. Mitochondria-targeted pyruvate dehydrogenase kinase inhibitors

Mitochondrial pyruvate dehydrogenase kinase (PDK) is activated in several cancers (but not in normal tissues) where it inhibits (through phosphorylation) the pyruvate dehydrogenase enzyme complex (PDC) which converts pyruvate into acetyl-CoA to support the TCA cycle and OXPHOS. PDK therefore favors glycolysis over OXPHOS, and contributes to the Warburg effect. Inhibition of PDK, and especially PDK1 (the most common isoform in many cancer types) (83), could provide a promising anti-cancer strategy to reverse the Warburg effect. Several PDK inhibitors can retard cancer cell proliferation (84) and overcome chemotherapeutic drug resistance (83). However, these inhibitors have some undesirable off-target effects.

To improve the selectivity for PDK1, Xu et al. (85) designed a series of mitochondria-targeted and tumor-specific PDK1 inhibitors (compounds 1a-f). Unlike the parental drugs, which were equally effective on cancer and normal cells, all these TPP⁺-conjugated PDK1 inhibitors were selective for cancer cells. Compounds 1c-f showed both improved antiproliferative effects and anti-PDK ability. IC₅₀ values of their antiproliferative effects against NCI-H1650 cells ranged from 0.21 μM to 0.88 μM. As expected, the most potent compound, 1f, converted the glucose metabolic profile in NCI-H1650 cancer cells from lactate production to increased oxygen consumption (85). While TPP-PDK1 inhibitors did not sabotage OXPHOS, they exhibited profound anti-cancer properties including disruption of mitochondrial membrane potential and promoting apoptotic cell death (85).

Dichloroacetate (DCA) is so far the only PDK inhibitor tested in clinical trials; it was well-tolerated in patients with recurrent malignant brain tumors and brain metastases (86). However, its anionic nature limits its cellular and mitochondrial uptake. To circumvent this problem, Pathak et al. (87) conjugated TPP⁺ to DCA via labile ester bonds. Mito-DCA was ~1000-fold more active than DCA against multiple prostate cancer cells, with greater effects on highly glycolytic cells (PC3) than less glycolytic cells (LNCaP). In addition to suppressing glycolysis, mito-DCA reduced basal cellular respiration and ATP synthesis (87). Since lactic acid can be immunosuppressive (88), mito-DCA’s ability to decrease lactate production can improve anti-cancer immunity. Mito-DCA did not affect metabolism in non-malignant cells (87).

2.3. Mitochondria-targeted triterpenoids

Pentacyclic triterpenes are plant metabolites that include oleanolic acids, ursolic acids, betulinic acids, betulonic acids, glycyrrhetinic acids, and others. These compounds possess many activities including anti-proliferative, pro-apoptotic, anti-angiogenic anti-oxidative and anti-cancer effects (89). Their anti-tumor effects are mainly mediated through activation of mitochondrial apoptosis (90, 91). However, their poor bioavailability and limited intracellular accumulation restrain their use in vivo (92).

To improve cellular uptake and mitochondria targeting, Nedopekina et al. (93) synthesized a series of TPP⁺ conjugates of ursolic and betulinic acids. In vitro, the conjugates were more effective on three human cancer lines (MCF-7, HCT-116 and TET21N) than with the parental compounds. Betulinic acid conjugate 9 exhibited the highest efficacy. Its superior cell killing was mediated through mitochondria-mediated caspase-dependent apoptosis (93). Tsepaeva et al. (94) developed a series of TPP-betulin derivatives, of which compound 8–13 had the highest efficacy (IC₅₀ as low as 0.045 μM) against various cancer cell lines, with ~10-fold specificity for cancer cells. The length of the alkyl linker and the location of the TPP conjugate affected anti-tumor activity. Compounds 12 and 13, bearing two phosphonium groups, reduced mitochondrial membrane potential more than compounds with only one phosphonium. When tested against PC-3 prostate adenocarcinoma cells, no changes in the cell cycle phase were observed (94). Recently, the same group developed a series of C-28-TPP-conjugated triterpenoids with variable length alkyl linkers and C-3 oxygen-containing groups (33). Elongation of the linker generally lowered their anti-tumor effects, whereas variation in C-3 oxygen groups did not considerably influence efficacy. TPP⁺-betulinic conjugates 6a-d lowered ΔΨm ~3-fold (33). Compounds with short propylene and butylene linkers caused greater ROS production. In summary, the C-28-TPP-triterpenoid conjugates with shorter alkyl linkers have the most promising anticancer activity within this class (33).

Derivatives of Glycyrrhetinic acid (GA), the pharmacologically active metabolite of licorice, have extensively substantial chemopreventive potential (95). Jin et al. (96) recently designed a series of GA-TPP conjugates, of which compound 2f had excellent activity against human cancer cells and improved selectivity for cancer cells, inhibiting proliferation and migration, and triggering mitochondrial-dependent apoptosis via collapse of the mitochondrial membrane potential (96). The potential of GA-TPP conjugates for tertiary chemoprevention warrants further evaluation.

2.4. Other types of TPP⁺-based compounds

2.4.1. Mitochondria-targeted tamoxifen

Tamoxifen is FDA-approved for primary breast chemoprevention in premenopausal and postmenopausal women at high risk, and for the adjuvant treatment of early stage estrogen receptor (ER)-positive breast cancer and ER-positive metastatic breast cancer (to suppress recurrence or metastasis). However, tamoxifen is insufficient for HER2-positive breast cancer. Doses of tamoxifen well above typical therapeutic levels can compromise mitochondrial complex I in breast cancer cells (97, 98). Mitochondria-targeted tamoxifen (Mito-Tam) was designed to enhance its mitochondrial effects (99). Mito-Tam kills not only ER-positive, but also HER2-positive and triple-negative breast cancer, including lines that are tamoxifen-resistant. Mito-Tam does not cause systemic toxicity. Mito-Tam is particularly effective against HER2^high tumors (99). Molecular modeling predicts that Mito-Tam may block the interaction between the ubiquinone-binding pocket and Complex I. It has also been found that a fraction of HER2 protein is localized at the mitochondrial inner membrane, and HER2^high cells have elevated OXPHOS. These findings provide new potential for using Mito-Tam for the tertiary prevention of HER2^high breast cancer.

Many chemotherapeutics often eventually induce cell senescence, which can later contribute to tumor relapse (100). However, Mito-Tam suppresses tumor progression without inducing senescence (101). Furthermore, Mito-Tam eliminates both tumor-derived senescent cells and tumor-unrelated senescent cells. The effects of Mito-Tam on senescent cells were distinct from those of other agents, including tamoxifen, and inhibitors of Complex I (rotenone and pieridicin A), and Complex II (Mito-VES and atpenin 5). The mechanisms by which Mito-Tam kills senescent cells may involve suppression of adenine nucleotide translocase 2 (ANT2), which transfers ATP into mitochondria that are faced with ATP crisis (101). The forced overexpression of ANT2 in senescent cells renders them less sensitive to Mito-Tam.

In summary, Mito-Tam is a promising agent for tertiary prevention of multiple types of breast cancer and tumor recurrence. To the best of our knowledge, Mito-Tam’s potential for primary or secondary chemoprevention remains to be determined. Due to its excellent efficacy and safety profile, it is entering Phase I clinical trials in the Czech Republic.

2.4.2. TPP⁺-based plastoquinone (SkQ1)

Recent studies have shown that anti-cancer effects of mitochondria-targeted antioxidants are mostly mediated though their pro-oxidant properties when used at higher concentrations (102). Titova et al. (103) investigated how low nanomolar concentrations of the mitochondria-targeted antioxidant SKQ1 regulates the cell cycle of fibrosarcoma and rhabdomyosarcoma cells. SkQ1 suppressed the growth of fibrosarcoma and rhabdomyosarcoma cells and caused cell cycle arrest during transition from mitosis to cytokinesis (103). This effect could partly be attributed to decreased levels of Aurora family and Rb proteins, both of which play important roles in regulating mitosis progression and the cell cycle (104, 105). Since tumorigenesis is highly regulated by cell cycle disorders, the cytostatic properties of mitochondria-targeted SkQ1 are considered to be protective against tumor development. Appropriate animal models are required to test its preventive potential.

2.4.3. TPP⁺-based peptide nucleic acids (PNA-mito)

Mitochondrial architecture changes such as mitochondrial fusion and fission have roles in tumor development, stemness of tumor cells (106), and chemotherapy drug response (107). Peptide nucleic acids (PNAs) are synthetic DNA analogues in which the deoxyribose phosphate backbone is replaced with a neutral peptide-like N-(2-aminoethyl) glycine skeleton. PNAs hybridize with complementary DNA or RNA with high specificity, and can disrupt DNA replication and transcription (108). In order to interfere with the displacement loop regulatory region of mitochondrial DNA, Chen et al. (109) developed a mitochondria-targeting PNA by conjugating TPP to a PNA that targets the D-loop (PNA-mito). In A549 lung cancer cells, PNA-mito blocked mitochondrial gene expression, disrupted mitochondrial fusion and the mitochondrial membrane potential, and decreased ATP generation. PNA-mito inhibited lung tumor growth in mice by >50%. Low dose (both 1 μM and 5 μM) PNA-mito significantly enhanced the sensitivity of drug-resistant A549 tumor spheres to cisplatin (109). These findings suggest tertiary lung cancer prevention potential for PNA-mito including tumors that are cisplatin-resistant.

3. Mitochondria-Targeted Chemotherapeutic Drugs

TPP⁺ has been linked with many classical chemotherapy drugs to attempt to enhance mitochondrial targeting and anti-tumor effects. Lipophilic TPP⁺-based chemotherapeutic drugs are expected to have enhanced efficacy but significantly decreased toxicity. Some of these conjugates may therefore have potential for cancer chemoprevention, especially for tertiary prevention.

3.1. TPP+ based standard chemotherapeutics

3.1.1. Mitochondria-targeted Pt(II) and Pt(IV) based drugs and pro-drugs

Platinum (Pt)-based drugs (cisplatin, carboplatin, and oxaliplatin) have long been a mainstay for many cancer types (110). However, they have well-known serious side effects, low bioavailability and poor stability, and can fail clinically due to acquired resistance (111).

Developing non-conventional monofunctional Pt(II) complexes is one potential way to circumvent these drawbacks. Recently, a series of mitochondria-targeted unconventional monofunctional Pt(II) complexes were developed (112). One of the derivatives, Pt(ortho-PPh₃CH₂Py) (NH₃)₂Cl] (NO₃)₂ (OPT), showed higher efficacy than cisplatin against lung cancer cells in vitro and tumor growth in vivo. In addition to its effects on nuclear and mitochondrial DNA, OPT disrupted mitochondrial structure and respiration, and induced cytochrome c release and apoptosis (112). Another TPP⁺-based terpyridine platinum(II) complex (113) was developed to inhibit thioredoxin reductase (TrxR). It inhibited cancer cell mitochondrial TrxR activity by 70%, and caused severe damage to mitochondrial morphology and function, which resulted in dual inhibition of mitochondrial and glycolytic bioenergetics. The resulting hypometabolic state was effective in human ovarian cancer cells (113).

Another way to overcome the shortcomings of conventional Pt(II)-based drugs is to develop Pt(IV) pro-drugs which can be conjugated with ligands at the axial positions to render “dual actions” (114). Jin et al. (115) developed and synthesized two mitochondria-targeted TPP⁺-based Pt(IV) complexes, namely, c,c, t-[Pt(NH₃)₂Cl₂(OCOCH₂CH₂CH₂CH₂PPh₃)(OH)]Br (PMT) and c,c,t-[Pt(NH₃)₂Cl₂(OCOCH₂CH₂CH₂CH₂PPh₃)₂]Br₂ (PDT). The TPP⁺ moiety on the axial position of Pt(IV) markedly changed the reactivity and anti-tumor effects. Although TPP⁺ increased their lipophilicity, the generation of Pt(II) species from these Pt(IV) prodrugs was decreased and thus their reactivity with DNA. Although these Pt(IV) compounds caused more mitochondrial damage (disrupted morphology, dissipation of ΔΨm, and decreased ATP production) than cisplatin, PMT and PDT were less effective on cancer cells in vitro (115). These modified Pt(IV) pro-drugs may therefore not be useful candidates.

3.1.2. Mito-Chlor and T-Platin-Cbl-NPs

The DNA alkylating agent chlorambucil is commonly used to treat chronic lymphocytic leukemia/small lymphocytic lymphoma (116). However, it has several severe adverse effects, low stability, non-selective reactivity towards all peripheral cells, and the development of resistance (117). Mitochondria-targeted chlorambucil (Mito-Chlor) developed by Millard et al. (118) preferentially localizes in cancer cell mitochondria where it damages mitochondrial DNA and causes cell cycle perturbation and apoptosis. Mito-Chlor is 80-fold more effective than chlorambucil on breast and pancreatic cancer cell lines that are insensitive to the parent drug. Mito-Chlor significantly retarded the growth of human pancreatic cancer in mice (118). Platin-Cbl (119), a combination of cisplatin and chlorambucil (Cbl), demonstrated efficacy for several tumor cell lines in the NCI60 cell panel, but was unable to completely block energy generation from fatty acid metabolism in cisplatin-resistant cells. To improve its mitochondrial effects, they encapsulated Platin-Cbl inside polymeric nanoparticles and further linked Platin-Cbl with TPP⁺. These mitochondria-targeted TPP-Platin-Cbl nanoparticles arrested all metabolic pathways in cisplatin-resistant cells, and further enhanced anti-tumor efficacy 5-fold (119).

3.1.3. Mitochondria-targeted doxorubicin

Doxorubicin (DOX), an anthracycline antitumor antibiotic, stabilizes topoisomerase 2 which prevents the resealing of DNA strands (120). Previous attempts to improve its anticancer efficacy and potency involved either directly linking TPP⁺ to DOX or encapsulating DOX in various liposomal or nanoparticle carriers (121). The efficacy of TPP⁺-DOX for wild-type cancer cell lines varied between different carrier formulations (121).

Efflux proteins are a major reason for DOX resistance (122). Adding a membrane-permeating peptide (with both hydrophobic and cationic properties) to DOX followed by attachment to copolymers can successfully bypass efflux pumps (123). This strategy enhances mitochondrial accumulation by >11-fold and markedly improves the suppression of DOX-resistant tumor spheroids. Another approach is dual-targeting DOX-loaded carrier systems, such as TPP-functionalized PEG-modified polydopamine (PDA-PEG-TPP-DOX) (124). Compared to DOX nanoparticles without TPP, PDA-PEG-TPP-DOX exhibited much higher mitochondrial penetration and a greater lowering of mitochondrial membrane potential. In MDA-MD-231 human breast cancer cells, PDA-PEG-TPP-DOX overcame drug resistance to a greater extent than PDA-PEG-DOX (124). Yu et al. (125) used an acidity triggered cleavable polyanion PEI-DMMA (PD) to coat lipid-polymer hybrid nanoparticles (DOX-PLGA/CPT) to form DOX-PLGA/CPT/PD. Although the net charge of this compound was negative, it could still target tumor sites, taking advantage of the acidic tumor microenvironment. After it was endocytosed by tumor cells, TPP⁺ guided DOX-PLGA/CPT to mitochondria, where it disrupted mtDNA and induced apoptosis (125). In vitro, both DOX-PLGA/CPT and DOX-PLGA/CPT/PD overcame DOX resistance of MCF-7/ADR cells (125). However, these two formulations showed completely different anticancer efficacy in mice. DOX-PLGA/CPT/PD caused almost four-fold greater tumor inhibition (84.9%) compared with DOX-PLGA/CPT (19.0%). Such drastic differences between in vitro and in vivo efficacy were ascribed to the overall negative charge of DOX-PLGA/CPT/PD which decreased its rapid clearance from the blood and enhanced its tumor tissue accumulation. Liu et al. recently embedded DOX-TPP into a series of nanoparticle formulations (126, 127). By conjugating DOX-TPP to water-soluble hyaluronic acid (HA) using an acid-cleavable bond to achieve pH-responsive selective targeting of cancer cells (128), HA-hydro-DOX-TPP mainly accumulated inside mitochondria (and not the nucleus) after its separation from the HA-nano carrier. The cellular content of HA-hydro-DOX-TPP in MCF-7/ADR cells was ~7-fold higher than that achieved with free DOX, and HA-hydro-DOX-TPP was much more effective on MCF-7/ADR cells (126). Its antitumor effects were further confirmed in vivo, and higher drug levels were consistently observed in tumors of the HA-hydro-DOX-TPP group (126). This nanoparticle formulation possessed a good safety profile (126).

HA-ionic-TPP-DOX, another hyaluronic acid modified DOX-TPP NC derivative with a different (ionic) linkage(127) showed comparatively inferior in vitro activities (127). However, in vivo HA-ionic-TPP-DOX exhibited much higher efficacy than DOX-TPP. Its higher antitumor efficacy in vivo could result from a higher tumor accumulation of HA-ionic-TPP-DOX (127). These results suggest its potential for future clinical trials.

Kim et al. (129) developed a self-assembled mitochondria-targeted TPP⁺-based nanoparticle probe (N1) with a cyanostilbene fluorescent tag to visualize the therapeutic effects. N1 was later modified with DOX to form the model drug N1-DOX. N1-DOX inhibited tumor growth by ~60%, exhibited higher cancer cell uptake and predominately mitochondrial accumulation, and caused mitochondrial dysfunction (massive ROS generation, MMP collapse) and eventually apoptosis.

3.1.4. Mitochondria-targeted taxanes

The taxanes, including paclitaxel and docetaxel, block at the G2/M phase of the cell cycle by enhancing the assembly and stability of microtubules (130). While the taxanes have been used clinically for many types of malignancies (131), they cause severe adverse reactions including neurologic toxicity and bone marrow suppression (132). Mitochondrial targeting may reduce these side effects on normal tissues.

Battogtokh et al. (133) synthesized TPP-Docetaxel (TPP-DTX) conjugates and an albumin-nanoparticle formulation TPP-DTX@FA-chol-BSA to provide for better drug stability and longer retention in the circulation. They proposed that drug-loaded nanoparticles would release about 48–60% of TPP-DTX in a low pH environment (such as tumor cell lysosomes). TPP-DTX conjugate and its nanoparticles showed significantly higher accumulation in mitochondria vs. the cytosol in MCF-7 cells, induced 2-fold more ROS, and triggered mitochondrial depolarization (133). However, the selectivity for cancer versus normal cells was not reported. In vivo, TPP-DTX nanoparticles redcued tumor size 4.8-fold. Zhang et al. (134) generated a different liposome system to deliver TPP-DTX. EphA10 is a receptor that is predominately expressed on some breast cancer cells (135). They added an EphA10 antibody onto the surface of the liposome carrier, and also synthesized a pH-sensitive PEG-Schiff base-cholesterol (PSC) to modify the liposome (PSLP) to ensure sustained release of the drug. This final drug-loaded product (EPSLP/TD) exhibited the highest cellular uptake, mitochondrial targeting, and the highest levels of cleaved poly (ADP-ribose) polymerase (PARP) and cytochrome c release to the cytosol (134).

Clinical responses to paclitaxel (PTX) are variable, and resistance mechanisms include multidrug resistance (MDR) transporters (136). Studies have tested the hypothesis that mitochondria-targeted drugs can overcome MDR by suppressing ATP synthesis. TPP⁺-conjugated PTX-loaded liposomes can improve efficacy against MDR-positive PTX-resistant ovarian cells and cisplatin-resistant lung cancer cells (137, 138). Han et al. (139) developed PTX nanocrystals stabilized with TPP⁺-modified Brij 98 (a P-glycoprotein inhibitor) to overcome MDR. All of the Brij-TPP/PTX nanocrystals (with different percentages of Brij98 in the formulations) showed much higher anti-proliferative effects on human breast cancer MCF-7 cells and its MDR line MCF-7/ADR (139). The anti-proliferative effects were correlated with loss of mitochondrial membrane potential, and the Brij-TPP/PTX nanocrystals were effective against 3D spheroids (139).

3.2. TPP⁺-based Hsp inhibitors

3.2.1. TPP⁺-based Hsp90

Heat shock protein (Hsp90) is a major chaperone protein that assists with proper folding and function of client proteins. Many Hsp90 clients, such as signaling kinases (e.g. MAPK, PI3K, AKT), p53 and telomerase, play crucial roles in cancer development (140). By binding to inactive clients, Hsp90 facilitates their activation. Hence, inhibiting Hsp90 could have anti-cancer effects by indirectly inactivating other oncoproteins (140). Though initial Hsp90-based anti-cancer strategies had modest effects (141). Kang et al. (142) developed the first in class mitochondria-targeted Hsp90 inhibitor, which contains the Hsp90 ATPase antagonist 17-(allylamino)-17-demethoxy-geldanamycin (17-AAG) linked to TPP⁺ (GA-TPP-OH). GA-TPP-OH induced loss of mitochondrial membrane potential, triggered cytochrome c release, and induced mitochondrial apoptosis (142). Hsp90 protein physically interacts with the PTP component protein CypD to inhibit pore opening to prevent initiation of cell death (142). GA-TPP-OH reduced the physical interaction between Hsp90 chaperones and CypD to potentiate the onset of cell death. GA-TPP-OH was tumor cell selective (142). GA-TPP-OH inhibited lung tumor growth in vivo by ~70%, a much larger effect than the the parental drug 17-AAG. GA-TPP-OH did not cause apparent toxicity to mice (142).

Gamitrinib specifically acts on the mitochondrial Hsp90 protein TRAP1. Mitochondria-targeted Gamitrinib was developed to study its regulation of the protein-folding response inside the mitochondria of tumor cells (143). At high dose, Gamitrinib-TPP causes rapid mitochondria-mediated apoptosis of tumor cells, whereas lower doses trigger an unfolded protein response (UPR), autophagy, and complete inhibition of NF-κB-dependent cell survival. Gamitrinib-TPP’s ability to blunt pro-survival mediators enhanced the sensitivity of tumor cells to death receptor agonists. Gamitrinib-TPP worked synergistically with the apoptosis-inducing agent TRAIL to arrest intracranial glioblastoma growth in mice (143). A recent study (144) focused on how Gamitrinib-TPP alters glutamine metabolism in non-small cell lung cancer (NSCLC) which relies heavily on glutamine metabolism (145). Glutamine-dependent NSCLCs are especially sensitive to Gamitrinib-TPP. TRAP1 inhibition by Gamitrinib-TPP increased the enzymatic activity of glutamine synthetase (GS) which synthesizes glutamine from glutamate. Increased phosphorylation of AMPK in Gamitrinib-TPP-treated cells implied that these cells were deprived of ATP (144). Thus, Gamitrinib-TPP could be a therapeutic strategy for glutamine-dependent cancers.

Bryant et al. (30) developed two variants of mitochondria-targeted Hsp90 inhibitors, PU-H7-TPP1 and PU-H7-TPP2, by linking the purine-based chemical scaffold PU-H71 with TPP⁺. The concentration of PU-H7-TPP in mitochondria was 17-fold higher than that in the cytosol. Both PU-H7-TPP1 and PU-H7-TPP2 reduced the viability of multiple acute myeloid leukemia (AML) cell lines, with PU-H7-TPP2 having greater efficacy (30). However, both treatments were ineffective for the chronic myeloid leukemia cell line K562. Unmodified PU-H71 was ineffective. PU-H7-TPP’s mechanistic effects include MMP dissipation, OXPHOS inhibition and ATP depletion (30). Mitochondrial metabolism is important for AML progression (146). Since mitochondrial Hsp90 (TRAP1) is upregulated in AML patient samples, and increased TRAP1 correlates with worse outcomes, PU-H7-TPP2 induced apoptosis and disrupted mitochondrial function in patient-derived AML samples (30). Thus, mitochondria-targeted Hsp90 inhibitors have therapeutic potential for AML patients.

All clinically proven Hsp90 inhibitors target the N-terminal part of Hsp90 to disconnect the interaction between Hsp90 and ATP. However, this inhibition ultimately leads to heat shock response, with more Hsp70 which can suppress apoptosis (147). To avoid this pitfall, inhibitors that target the C-terminus of Hsp90 have been developed that do not generate heat shock response. Zhang et al. (148) tcreated mitochondria-targeted Hsp90 C-terminal inhibitors by replacing an amine with TPP⁺. These modified compounds accumulated in the mitochondria and had a two-fold greater anti-proliferative effect on human breast cancer MCF-7 cells. The TPP⁺-modified compound only triggered the degradation of Hsp90 client proteins Raf-1 and CDK-4, but not Her2 or AKT (148).

3.2.2. TPP⁺-based Hsp70

Hsp70 expression is highly elevated in various cancers, and correlates strongly with tumor initiation and progression (149). Various small inhibitors have been developed to target various members of Hsp70 that localize to different subcellular compartments. Ko et al. (150) developed a lysosomal Hsp70 antagonist, apoptozole, which binds to Hsp70’s ATPase domain to disable its ATPase activity. The Hsp70 in mitochondria, known as mortalin, promotes cell proliferation and downregulates apoptosis (151). Park et al. (152) modified the lysosomal antagonist apoptozole with TPP⁺ to create Az-TPP-O3 which enabled the targeting of mitochondrial Hsp70 (mortalin). Its anti-tumor effects for a panel of 20 human cancer cell lines was improved by ~6-fold (152). Mitochondrial-localized Az-TPP-O3 disrupted the interaction of mortalin with p53; the dissociated p53 helped initiate Bak-mediated mitochondrial outer membrane permeabilization, leading to caspase-dependent apoptosis in cancer cells. Apoptozole-TPP-O3 did not affect lysosome-mediated apoptosis or autophagy (152).

3.3. Other types of TPP⁺-based compounds

3.3.1. TPP⁺-based mitochondrial K⁺ channel inhibitors

Blockade of mitochondrial ion channels can result in loss of ΔΨm, ROS production, mitochondrial swelling, cytochrome c release, and cancer cell death induction that bypasses upstream players in intrinsic apoptosis (e.g. Bcl-2 family proteins, p53 status, and others) (153). The voltage-gated potassium channel mtKv1.3 is overexpressed in many cancer cells compared to normal cells, and plays a pivotal role in cancer development (154). The plant-derived furocoumarins are specific Kv1.3 channel inhibitors, but they have only modest tumor-killing effects at pharmacologically relevant concentrations.

Leanza et al. (155) developed two derivatives by directly linking TPP⁺ with PAP-1’s carbon chain (PAPTP), or via indirect conjugation using a carbamic ester bond (PCARBTP). PAP-1 and its mitochondriotropic analogs only induced apoptotic cell death in cancer cells that highly expressed mtKv1.3, but not in cancer cells with lower expression or healthy counterparts. Both derivatives could induce the death of more than half of the B cells derived from patients with chronic lymphocytic leukemia. Upon accumulating in mitochondria, they induced an initial increase in membrane potential and ROS release, which facilitated opening of the PTP, followed by mitochondrial swelling and structure fragmentation, membrane depolarization and release of cytochrome c (155). The modified compounds showed stunning effects in decreasing tumor volume in both orthotopic B16F10 melanoma and a mouse pancreatic ductal adenocarcinoma model. Their tumor-killing effects were likely mediated by ROS production, as pre-treatment with an anti-oxidant eliminated their efficacy (155). These psoralen derivatives did not affect fast-proliferating normal cells including immune cells or major organs (155). In summary, mitochondria-targeted furocoumarin derivatives are ideal anti-cancer agents for cancers that highly express mtKv1.3 channels and have altered redox states.

3.3.2. TPP⁺-based trialkoxysilane

Biomineralization is a natural process of living organisms to form mineral-based materials from available elements within the biological matrix. Kim et al. (156) incorporated intracellular mitochondrial-targeting strategies into the biomineralization system to improve selectivity for tumor cells. Trialkoxysilane can cause mitochondrial dysfunction by forming silica particles in response to the alkaline pH of 8 inside the mitochondrial matrix. They linked TPP⁺ to trialkoxysilane, creating compound 1 which accumulated and initiated biomineralization mostly within tumor cell mitochondria (156). High concentrations of compound 1 within mitochondria formed large silica particles which disrupted mitochondrial morphology and depolarized the mitochondrial membrane (156). These events triggered apoptosis in various cancer cells, including drug-resistant cell lines. In contrast, compound 1 had low toxicity for normal cells. Systemic administration of compound1 inhibited tumor growth, but did not cause toxicity (156).

4. Mitochondria-targeted agents to modulate the tumor microenvironment

Mitochondria have important roles in establishing immune cell phenotypes, modulating their metabolic and physiologic states, and their functionality (157). In fact, almost every mitochondrial component, including catabolic and anabolic metabolism, redox systems, mtDNA and mitochondrial dynamics/turnover has a great impact on the immune system (157). Mitochondrial pathways (TCA, OXPHOS and amino acid metabolism) regulate immune cell activation status and and functional responses. For example, resting T cells (naïve and memory type) require mostly OXPHOS respiration and lipid oxidation, whereas proliferating effector T cells (Teff) switch to aerobic glycolysis and glutamine metabolism (158). Immune suppressive regulatory T cells have distinct metabolism to maintain their suppressive function, including increased complex-III dependent OXPHOS and decreased glycolysis (159). Mitochondria-generated ROS act as second messengers for both inflammasome activation (160) and antigen-specific T cell activation (161). And, mtDNA released from dysfunctional mitochondria can bind to and activate the NLRP3 inflammasome, which plays a critical role in innate immunity by triggering secretion of pro-inflammatory cytokines (e.g. interleukin-1β and IL-18) (162). Mitochondrial dynamics also control T cell activation and differentiation. Activated T cells tend to have more fragmented mitochondrial membrane structure and more expanded cristae to reduce ETC efficiency and enhance glycolytic metabolism, whereas memory T cells have more fused mitochondrial structure that favors OXPHOS and fatty acid oxidation (163).

The immune system is a major determinant of the tumor microenvironment. The dual forces of how the tumor affects immunity and host-protection have established the foundation for the “cancer immunoediting” theory (164). One way tumors restrain host immunity is by rendering immune cells metabolically anergic. Non-stop tumor growth modifies the surrounding environment to be nutrient-deficient/altered, acidic, hypoxic, and with accumulated waste. As a result, immune cells are not properly activated or have lost their battle for resources. Mitochondria-targeted compounds could potentially influence immunosurveillance in a cancer cell-intrinsic or cancer cell-extrinsic fashion. Through selectively inducing energy crisis and death in tumor cells, mitochondria-targeted compounds could potentially make more nutrients available to the microenvironment. Moreover, the release of mtDNA or mtROS from dying tumor cells could help to initiate innate immunity and adaptive immunity. For example, mitochondria-targeted IR-780 was recently developed as an immunogenic cell death inducer through its ability to trigger anticancer immunity (165). After accumulating inside tumor cell mitochondria and killing cancer cells, large amounts of tumor-associated antigens were released and triggered dendritic cell maturation, and more T cells infiltrated into the tumor microenvironment which finally led to suppression of tumor development and the prevention of metastasis (165). Another study developed the TPP⁺-based Ce6-loaded nanocarrier for mitochondria-targeted anticancer photodynamic therapy (166). This Ce6-loaded nanocarrier exhibited superior in vivo anticancer efficacy, specifically lowering tumor load in two rodent models by ~90%. The immunological mechanisms were increased activation of DC cells and T cells, T cell infiltration and upregulated cytokines (IFN-γ and TNF-α) in tumors (166). Tumor metabolism waste could also negatively affect anticancer immunity, with lactate being the most prominent. Lactate can inhibit activation and maturation of DCs (167) and proliferation of T_eff cells (168) in the tumor milieu. After shutting down tumor metabolism (OXPHOS or/and glycolysis) by mito-compounds, metabolism waste would decrease and exert less inhibitory effects on immunosurveillance. As previously discussed, mito-DCA helps dendritic cells to excrete more anticancer-related cytokines (87). Alternatively, mitochondria-targeted compounds could regulate anticancer immunological activities via directly modulating mitochondrial metabolism in immune cells. However, there is not yet direct evidence with TPP⁺-based mito-compounds to support this viewpoint. However, biguanides (phenformin and metformin), which can inhibit mitochondrial respiratory complex I, can suppress the inhibitory function of granulocytic MDSC ex vivo and their accumulation in tumor microenvironment in vivo through enhanced AMPK phosphorylation (169, 170).

5. Mitochondria-targeted agents in clinical trials

Numerous mitochondria-targeted agents have shown improved anticancer efficacy and potency in preclinical settings. However, most of these mito-compounds have not yet advanced to clinical trials for cancer chemoprevention. To our knowledge, Mito-TAM is the only such compound that has advanced to a Phase I clinical cancer trial.

Regarding human use of mitochondria-targeted TPP⁺-based agents for diseases other than cancer, the antioxidant mitoquinone (mitoQ) has been the most extensively investigated. It consists of the parental ubiquinone CoQ10 covalently linked with TPP via a ten-carbon chain (171). In many preclinical studies, mitoQ is protective against mitochondrial oxidative stress in wide variety of diseases and conditions (172). Eighteen studies on the diverse actions of mitoQ can be found on http://www.clinicaltrials.gov/. Of these, four have been completed, one was terminated, and 13 are in the recruiting or active stage. Although considered a neuroprotective compound, mitoQ did not slow down the progression of Parkinson’s Disease in a Phase II trial (173). In another phase II trial, mitoQ suppressed the levels of liver enzymes that are indicative of liver damage in hepatitis C patients (174). Recent human trials showed that chronic supplementation with mitoQ could safely ameliorate age-related vascular function deterioration (175). Although conflicting results have been seen in diseases having oxidative damage as a pathological cause, mitoQ may still hold promise as an effective antioxidant in the ongoing trials.

Mitochondria-targeted compounds that are promising candidates for clinical translation for cancer prevention should ideally have good bioavailability, prolonged retention in circulation, selective targeting of tumor cells, and high accumulation in the mitochondria of tumor cells. Oral administration of these agents would be preferred for ease of use. Because of TPP’s lipophilicity, TPP⁺-conjugates are generally expected to traverse phospholipid bilayers and enter the systemic circulation from the gut. Consistent with this premise, TPP conjugation has been shown to generally enhance bioavailability and cellular uptake (52, 61, 93). However, once in the circulation, the positive charge on TPP⁺ renders them susceptible to rapid clearance by the reticuloendothelial system. To overcome this barrier, it is ideal to have a negative surface charge when in the circulation, and to later reveal the positive charge once inside tumor cells to target mitochondria; the DOX-PLGA/CPT nanoparticles discussed above are an example of this approach (125). This type of charge conversion can be achieved by coating nanoparticles with an acidity-triggered-cleavable polyanion material; once triggered by the acidic pH in the tumor microenvironment, the negatively charged coating will be stripped off to expose the TPP cations. Another way to prolong the retention time of TPP⁺ compounds in the circulation and to deliver a sufficient amount is to encapsulate TPP⁺ compounds inside nanosized carriers. There has been an increasing trend to use of nano-formulations for Enhanced Permeability and Retention (EPR) effect and tumor site targetability. Mitochondrial accumulation is usually enhanced many-fold by the TPP⁺ moiety as noted for the many examples discussed in this review. However, variations in the alkyl carbon chain that links TPP⁺ to its cargo group can also affect the rate of mitochondrial uptake, the specific site of action (mitochondrial membrane or matrix), and the corresponding effects on cellular respiration. In general, longer linkers (usually carbon n = 3–10) result in more hydrophobic conjugates, faster mitochondrial localization, and greater ability to inhibit OXPHOS (176, 177).

6. Future Perspectives

Relative to the parental compounds, TPP⁺ conjugants have improved in vitro efficacy against cancer cell lines on the order of ≥5-fold. The greatest improvement of efficacy has been observed with mitochondria-targeted metformin that has a 10-carbon linker (IC₅₀ of 1.1 μM vs. 1.3 mM for metformin) (178). Although numerous in vitro anti-tumor effects have been shown, only some studies examined their anticancer efficacy in vivo. In fact, not all in vitro data can be translated in vivo. Hence, it will be useful for future studies to include in vivo tumor models to better evaluate chemopreventive efficacy as well as potential toxicity. To date, most in vivo tests have been done using mouse models which omit the potential effect that mitochondria-targeted compounds could have on modulating anticancer immunity. Future studies will need more syngeneic tumor models to fully understand how various mitochondria-targeted species perform in the context of a functional immune system. Although TPP⁺ has low toxicity, non-specific toxicities have been reported for some TPP-derived compounds that were given systemically (134). To reduce unwanted toxicities, attempts have included modification of the TPP cation and liposome incorporation of the modified compounds to improve selectivity for tumor cells and their mitochondria (134, 179). Regional administration of the compounds, such as intratumoral injection, could also be an alternative method to systemic administration. Advantages of intratumoral injection include the ability to greatly enhance drug concentration around the target injection site, and to lower the exposure of other organs to the drug (180, 181). Moreover, in situ vaccination using mitochondria-targeted compounds could also help to prevent distant/secondary tumor challenge (unpublished data).

Overcoming resistance to standard chemotherapeutics and preventing distant metastasis have been a recent focus for broadening the utility of mitochondria-targeted compounds for cancer prevention and treatment. Besides conjugating parental drugs to TPP⁺ as a potential way to evade efflux pumps, a recent trend has been to encase the conjugates in nanosized carriers for better drug stability, or to incorporate them with inhibitors of P-glycoprotein pumps (124, 125, 139). Multiple animal studies have demonstrated that some TPP-derived compounds can cross the blood-brain barrier and be taken up into brain tissue, including drugs that have the potential to protect from mitochondrial damage in neurodegenerative diseases (29, 182–184). Therefore, some mitochondria-targeted compounds may be useful to treat or prevent brain metastasis. As lead examples, both Mito-HNK and Mito-LND have shown excellent ability to retard brain metastases (31, 32). It is therefore likely that other agents could be developed to treat brain metastases or neurological tumors.

7. Conclusion

Only in the last decade has mitochondrial metabolism emerged as a promising target for cancer prevention/treatment, thanks to our better understanding of the vital role of mitochondria in cancer development, progression, and regulating the tumor immune microenvironment. A diverse array of mitochondria-targeting drugs has been developed that show good anticancer activities in preclinical settings. In general, many of these mito-targeting compounds have better pharmacological activities (bioavailability, cellular and mitochondrial uptake) and markedly improved anticancer efficacy compared with the parental compounds without harming normal cells. These properties make this class of drugs attractive for development as preventive and therapeutic agents. However, translating these promising preclinical agents to clinical use will require considerably more effort. In this review, we summarized our vision for the next frontier of this field including anti-cancer strategies aimed at mitochondrial metabolic and bioenergetic modifications that are characteristic of many cancers (Figure 3 and Table 3) as well as newly developed drug carrier delivery systems. In addition to suppressing primary tumors, some mito-compounds (Mito-HNK, Mito-LND, and Mito-TAM) can efficiently prevent or suppress tumor metastasis. We also pointed out the potential role of mito-compounds in reshaping the tumor immune microenvironment in a manner that would promote cancer elimination. Although this new direction is in its infancy and needs more in-depth study, we predict that targeting immuno-metabolism will lead to new breakthroughs for prevention and treatment. As this field continues to advance, and with many promising candidates under development, there is considerable potential for several of these candidates to advance to clinical trials and eventual preventive use.

Figure 3. Potential anticancer activities of mitochondria-targeted compounds (MTC).

Inhibition on ETC by MTC (A) leads to mitochondrial uncoupling, with overproduction of mitochondrial ROS (B) (31, 32), and decreases in ATP generation (C) (30, 109). The initial production of mitochondrial ROS can lead to membrane potential dissipation (D) (69, 96, 115), and opening of the mitochondrial permeability transition pore (mPTPC) (E) (73). The opening of the mPTPC could intensify mitochondrial ROS production (F) (191), allowing large amounts of solutes to enter the mitochondrial matrix and cause necrosis-like cell death (G) (191), and release of cytochrome c (Cyt C) into cytosol to trigger apoptosis (H) (73, 115, 156). Induction of mitochondrial outer membrane permeabilization (MOMP, I) (152), inhibition of Kv1.3 channels (J) (155) and TrxR activity (K) (69) by MTC could also promote more ROS production. Excessive mitochondrial ROS can damage mitochondrial DNA (L) (118) and telomeric nuclear DNA (nDNA) (M) (192). These mitochondrial stressors can potentially lead to autophagy-induced cell death (N) (32).

Table 3.

Regulation on mitochondrial ETC complexes and glycolysis

Mitochondrial- targeted compounds	ETC complexes	Glycolysis	Reference
Mito-honokiol	Complex I inhibition	Inhibition	(31)
Mito-lonidamine	Complex I inhibition	Inhibition	(32)
Mito-metformin	Complex I inhibition	Inhibition	(178)
Mito-resveratrol	Complex I and III inhibition	-	(53)
GA-TPP+C10	Complex I inhibition	Increase	(75)
Mito-tamoxifen	Complex I inhibition	-	(99)
Mito-DCA	-	Inhibition	(87)
TPP- pyruvate dehydrogenase kinase inhibitors	-	Inhibition	(85)
MitoVit2 E	Complex I inhibition	-	(186)
Mito vitamin E succinate	Complex II inhibition	-	(187)
MitoVE11S	Complex II and III inhibition	-	(188)
Mito-Coenzyme Q10	Complex I inhibition	-	(189)
Mito-3-bromopyruvate	-	Inhibition	(190)