Mitochondrial-targeted Hsp90 C-terminal Inhibitors that Manifest Anticancer Activities

Abstract

The development of C-terminal heat shock protein (Hsp90) inhibitors has emerged as a potential treatment for cancer. Small molecules that target the mitochondria have proven to be efficacious towards cancer, as the reprogramming of mitochondrial function is often associated with oncogenesis. Herein, we developed triphenylphosphonium (TPP)-conjugated Hsp90 C-terminal inhibitors and evaluated their anticancer activity and their accumulation in the mitochondria. TPP-conjugated Hsp90 C-terminal inhibitors manifested increased activity against various cancer cell lines as compared to the parent compounds.

Graphical Abstract

As a member of the molecular chaperone family, heat shock protein 90 (Hsp90) is essential for the conformational maturation of nascent polypeptides and the rematuration of denatured proteins.¹ Specifically, Hsp90 regulates the folding, stability, and function of more than 300 client protein subsstrates, many of which are critical for cancer cell survival and malignant transformation.^2–7 In fact, Hsp90-dependent client proteins (e.g., Her2, CDK6, Akt) are directly associated with all ten hallmarks of cancer. Consequently, Hsp90 inhibitors may be beneficial for the treatment of cancer as they target multiple oncogenic pathways simultaneously.

Numerous investigations of the Hsp90 N-terminal ATP-binding site have been pursued, which ultimately produced 17 small molecules derived from natural products (eg., Figure 1) that underwent clinical evaluation.^8–9 Unfortunately, the pro-survival heat shock response (HSR) that is induced upon the administration of N-terminal inhibitors is detrimental, as increased levels of Hsp90 are produced.¹⁰ In contrast, Hsp90 C-terminal inhibitors can segregate the HSR from the degradation of client proteins, which provides an opportunity to develop small molecules that overcome this undesired effect.^11–19

Figure 1.

(a) Natural product-derived Hsp90 N-terminal inhibitors

(b) Natural product-derived Hsp90 C-terminal inhibitors

Novobiocin was the first Hsp90 C-terminal inhibitor identified and shown to degrade client protein substrates without induction of the HSR.^20–21 Compounds based on the coumarin core of novobiocin and structurally related analogs, have been synthesized to improve their anticancer activity (eg., ~700 μM against SKBr3), which ultimately, led to molecules that manifest ~1000-fold greater activity than novobiocin.^22–26 Unfortunately, no co-crystal structure of an inhibitor bound to the Hsp90 C-terminus exists, which makes the development of improved analogs challenging.

Since mitochondrial bioenergetics are required for cancer cell survival, anticancer agents that accumulate in the mitochondria may overcome some of the undesired effects resulting from inhibition of the cytosolic Hsp90 isoforms (HSP90α and HSP90β). The mitochondrial Hsp90 homolog, TRAP1, can induce cell death in cancer cells if inhibited.²⁷ Therefore, the development of mitochondrial-targeted Hsp90 inhibitors represents a promising strategy for the development of new anti-cancer agents. Several N-terminal Hsp90 inhibitors have been modified to contain the triphenylphosphonium (TPP) salt, which is commonly used to direct compounds towards the mitochondria.^28–29 The TPP containing N-terminal inhibitors demonstrated enhanced anti-proliferative activity as compared to their non-TPP conjugated analogs. Giorgio et al. also reported a TPP-conjugated molecule that manifested ~10 fold greater activity.³⁰ Therefore, the TPP moiety was introduced onto novobiocin-based Hsp90 C-terminal inhibitors and they were evaluated for anticancer activity to determine whether such modifications enhanced Hsp90 C-terminal inhibitory activity.

Three Hsp90 C-terminal inhibitors (1, 2, and 3) were chosen for modification, based on prior evaluations and confirmation of Hsp90 inhibitory activity. Compound 1 contains the coumarin core as found in novobiocin, whereas compounds 2 and 3 were identified as more potent Hsp90 C-terminal inhibitory scaffolds in recent studies (Figure 2). ^{22, 31–32}

Figure 2.

Hsp90 C-terminal inhibitors

Since the benzamide side chain of these molecules is required for anti-cancer activity, the TPP moiety was not introduced onto the benzamide. Instead, studies have shown that the incorporation of an ionizable N, N-dimethylaminopropyl moiety in lieu of the sugar on novobiocin can enhance activity against various cancer cell lines. Furthermore, extensive structure-activity relationship (SAR) studies have shown that modifications to the amine can produce compounds that manifest similar activity. When combined, these data supported modification to the sugar region of novobiocin is solvent-accesible and represents a location at which replacement of the amine can occur along with TPP to give a mitochondrial-targeted Hsp90 C-terminal inhibitor (Scheme 1).

Scheme 1.

Design of TPP conjugated Hsp90 C-terminal inhibitors

Syntheses of the proposed analogs began via their corresponding phenols, 7, 9 and 11, which have been previously reported.^{22, 31–32} In general, 1-bromoethanol was attached to the phenol via a Mitsunobu coupling reaction. The resulting alkyl bromides were then heated at reflux in the presence of triphenyl phosphine in acetonitrile until the corresponding TPP salts precipitated from solution to give 4, 5 and 6, respectively. (Scheme 2).

Scheme 2. Synthesis of TPP conjugated Hsp90 C-terminal inhibitors 4, 5, 6.

(a) 1-bromoethanol, DIAD, PPh₃, THF, 0 °C to r.t., overnight, 30-50% (b) PPh₃, acetonitrile, reflux, overnight 25-45%

An MTT assay was conducted to evaluate the anticancer activities of 4, 5 and 6, and all of the compounds were found to exhibit efficacious activity against various cancer cell lines (SKBr3, MCF-7, PC3 and HCT-116) as shown in Table 1. A two-fold increase in anti-proliferative activity was observed for 4, 5 and 6, against the MCF-7 cell line when compared to the amino-counterparts 1, 2 and 3, respectively. Compound 5 manifested ~ 10-fold increased activity against the SKBr3 cell line when compared to amine 2 (Table 1).

Table 1.

Cytotoxicity of 4, 5 and 6 in comparison of their parent compounds

	MCF-7 (μM)	SKBr3 (μM)	PC3 (μM)	HCT-116 (μM)
4	0.80 ±0.07	0.31 ± 0.19	1.94 ± 0.84	0.64 ±0.21
1	1.80 ± 1.48	0.85 ± 0.10	1.01 ± 0.06	0.83 ±0.20
5	0.93 ±0.25	0.21 ±0.05	2.94 ± 1.32	0.45 ± 0.15
2	2.14 ± 0.75	2.03 ± 0.23	3.53 ± 0.91	1.49 ± 0.47
6	0.17	0.23 ±0.04	1.56 ±0.19	0.38 ±0.06
3	0.34 ±0.02	0.25 ±0.05	N.D.	N.D.

Compounds 4, 5 and 6 were also evaluated via western blot analyses of SKBr3 cell lysates following incubation for 24h in an effort to confirm that TPP conjugated compounds localized in the mitochondria. Since GDA is a known Hsp90 inhibitor and resides in the cytoplasm, client protein degradation (such as Her2 and AKT) was observed. In contrast, 4, 5 and 6 did not degrade related proteins at concentrations that mirrored the cellular IC₅₀ value. Previous results showed that parent compounds 1, 2, and 3 function as Hsp90 C-terminal inhibitors and induce degradation of these client proteins.^{22, 31–32} Consequently, we proposed that installation of TPP group to these parent compounds will redirect them into the mitochondria. As a consequence, cytoplasmic client proteins will not be degraded by the mitochondrial-targeted Hps90 inhibitors. The TPP conjugated compounds continued to produce steady levels of Hsp90 and Hsp70, which is a hallmark of Hsp90 C-terminal inhibition.

A mitochondria-targeted accumulation assay was also performed to quantify the relative concentrations of 4, 5 and 6. These compounds all showed significant accumulation in the mitochondria as compared to the cytoplasm. Surprising, the parent compounds also showed similar levels of accumulation in the mitochondria. It appears that the N,N-dimethyl amino moiety present in the parent compound is ionized under physiological conditions and appears to facilitate their transport into the mitochondria, much like the TPP moiety. Nonetheless, compounds 4 and 6 showed increased concentrations of mitochondria accumulation as compared to their corresponding parent compounds. The TPP moiety not only can drive these compounds into the mitochondria, but is capable of having similar effects as other ionizable functional groups.

In this study, we developed three TPP conjugated Hsp90 C-terminal inhibitors and evaluated their antiproliferative activity, as well as their mitochondria accumulation. Experimental results clearly showed these modifications produced compounds that manifested increased anti-proliferative activity. A mitochondria accumulation assay confirmed that TPP conjugation with Hsp90 C-terminal inhibitors produced compounds that target the mitochondria. Future research will be conducted to generate additional compounds that contain TPP in an effort to enhance activity and to study the consequences mechanism of mitochondrial inhibition.

Figure 3. Western bolt for TPP conjugated Hsp90 C-terminal inhibitors 4, 5, 6.

D: DMSO, G: geldanamycin (GDA), 4, 5, 6 showed as low & high concentrations, L: ½ of IC₅₀, H: 5 × IC₅₀.

Figure 4.

Mitochondrial targeting of compounds. LC-MS quantification of compound 1 (black), compound 2 (red), compound 3 (cyan), compound 4 (blue), compound 5 (magenta), compound 6 (green) in isolated fractions of 24 h treated PC3 cells. Cyto, cytosol; mito, mitochondria and AU is the ratio of the compound and the internal control.