Inhibition of phosphatidylinositol-3-kinase by the furanosesquiterpenoid hibiscone C

Bioorganic & Medicinal Chemistry Letters 27 (2017) 3087–3091

Contents lists available at ScienceDirect

Bioorganic & Medicinal Chemistry Letters journal homepage: www.elsevier.com/locate/bmcl

Inhibition of phosphatidylinositol-3-kinase by the furanosesquiterpenoid hibiscone C Caroline Besley a, Dena P. Rhinehart b, Taylor Ammons a, Brian C. Goess b, Jason S. Rawlings a,⇑ a b

Furman University, Department of Biology, 3300 Poinsett Highway, Greenville, SC 29613, USA Furman University, Department of Chemistry, 3300 Poinsett Highway, Greenville, SC 29613, USA

a r t i c l e

i n f o

Article history: Received 18 April 2017 Revised 8 May 2017 Accepted 13 May 2017 Available online 15 May 2017 Keywords: Hibiscone C Wortmannin PI3K Furanosteroid Furanosesquiterpenoid

a b s t r a c t The phosphatidylinositol-3-kinase (PI3K) pathway regulates cellular metabolism and is upregulated in many cancers, making it an attractive chemotherapeutic target. Wortmannin is a potent inhibitor of PI3K; however, its potential as a chemotherapeutic is limited due to its instability, lack of selectivity, and lengthy chemical synthesis. In contrast, hibiscone C, a structurally simpler and less studied member of the furanosteroid family, has been expediently prepared by total synthesis. We demonstrate that hibiscone C competitively inhibits PI3K activity in intact cells, slows proliferation, and induces cell death. Hibiscone C may therefore serve as a productive scaffold for the development of therapeutically relevant PI3K inhibitors. Ó 2017 Elsevier Ltd. All rights reserved.

The members of the sesquiterpene furanosteroid family of natural products have attracted significant attention as inhibitors of the phosphatidylinositol-3-kinases (PI3K),1 enzymes known to play critical roles in cell growth and differentiation. The key structural similarity shared by each member of this family is a furan ring flanked by two conjugated carbonyl groups. The inherent electrophilicity of these diacylheteroaryl systems (Scheme 1) is enhanced in furanosteroids by ring strain caused by the fusion of many contiguous sp2-hybridized carbons within their polycyclic core structures. As a result, furanosteroids display increased reactivity towards nucleophiles that accounts for their varied and often potent biological activities.2 The most intensely studied member of this family is wortmannin (Scheme 1), which inhibits multiple kinases, including all isoforms of PI3K, and mammalian Polo-like kinase, a cell cycle enzyme that has been proposed as a biomarker for certain types of cancers.3 A crystal structure of wortmannin bound in the ATP binding site of PI3K reveals the key feature of its mechanism of action; namely, covalent attachment of a lysine residue that results in irreversible inhibition (Scheme 1).4 The enhanced electrophilicity of the furanosteroids also contributes to their toxicity and reduces their potential as chemotherapeutic agents. Of particular concern is their lack of enzyme selectivity. For instance, wortmannin is known to inhibit numerous enzymes beyond those that are most important for its antiproliferative activity, which ⇑ Corresponding author. E-mail address: [email protected] (J.S. Rawlings). http://dx.doi.org/10.1016/j.bmcl.2017.05.041 0960-894X/Ó 2017 Elsevier Ltd. All rights reserved.

limits its potential as an anticancer agent.3 Additionally, wortmannin is known to undergo hydrolytic opening of the furan ring in neutral buffer,3 decomposes to inactive products via lactone hydrolysis under mild conditions, and reacts with a variety of nucleophiles under physiological conditions.5 Numerous studies aimed at preparing less toxic wortmannin analogs have identified molecules that exhibit similar potencies as wortmannin, and efforts targeting more stable and more potent analogs are ongoing (Fig. 1).6 Given their structural similarities, it is reasonable to suggest that all members of this family of natural products may share a similar mechanism of action, and a number of them in addition to wortmannin7 have succumbed to total synthesis (Fig. 2).8 Despite intense recent synthetic interest in this family of natural products, surprisingly little is known about the biological activity of one of its structurally simplest members, the furanosesquiterpoenoid hibiscone C.9 We were attracted to this synthesis target because hibiscone C lacks the lactone functionality that contributes to the instability of wortmannin. Aided by our racemic hibiscone C total synthesis effort (Scheme 2),8c herein we describe results of biological assays that suggest the hibiscone C scaffold may be a productive starting point for the discovery of new furanosteroids with desirable biological activity. To evaluate the biological activity of racemic hibiscone C, we first tested its ability to inhibit PI3K signaling in T lymphocytes. T lymphocytes (T cells) are white blood cells of the immune system that have multiple functions including facilitating activation of other immune cells, maintaining peripheral tolerance, and killing

3088

C. Besley et al. / Bioorganic & Medicinal Chemistry Letters 27 (2017) 3087–3091 CH3 O

H3 CO Ac O CH3

O

CH3

O

H

O

O

H

O

OH

O

H

O

O

O OH

HN

HN

wo rtm a n n in

CH3 O

H3 CO Ac O

CH3

P I3 K-Lys -NH2

O

CH3 O

H3 CO Ac O

Lys -P I3 K

Lys -P I3 K

Scheme 1. Mechanism of action of wortmannin.

H3 CO Ac O

CH3 O

CH3 O

(a)

CH3 O

(b)

H

O

CH3 O

H3 CO Ac O

O

H

O

O

O OH

N

H3 C

Fig. 1. Previously prepared analogs of wortmannin: (a) showed diminished activity,6a whereas (b) proceeded to clinical trials under the name PX-866.6b

of infected cells. Metabolically, activated T cells resemble cancer cells in that they engage in aerobic glycolysis.10 Furthermore, the cellular metabolism of T cells, like cancer cells, is dependent on PI3K pathway activity.11 It has been shown that activated T cells can respond to stimulation by the growth factor interleukin-2 (IL-2) by activating PI3K. Activated PI3K phosphorylates PIP2 to generate PIP3, which subsequently phosphorylates the serine/threonine kinase Akt.12 Therefore, the phosphorylation of Akt serves as an excellent readout of PI3K activity. To determine if hibiscone C can inhibit PI3K pathway activity, activated T cells growing in culture were starved of IL-2, which resets the PI3K pathway, and were then treated with 100 lM wortmannin, racemic hibiscone C, or vehicle. The extent of Akt phosphorylation was measured in response to IL-2 restimulation by Western blot (Fig. 3). Hibiscone C prevented the phosphorylation of Akt, similar to results obtained with the known PI3K inhibitor wortmannin. As the critical diacylfuran functional group responsible for the irreversibility of the PI3K inhibitory activity of wortmannin is also present in hibiscone C, it is reasonable to hypothesize that hibis-

Fig. 3. Hibsicone C inhibits PI3K pathway activity. Activated T cells were grown in culture (growing), then deprived of IL-2 overnight (starved). Cells were then treated with 100 lM wortmannin (Wort.), 100 lM racemic hibiscone C (HibC), or vehicle (DMSO) for 30 min and then stimulated with 1000 U/mL recombinant human (rh) IL-2 for the times indicated. Phosphorylation of Akt was detected by Western blot. Detection of actin serves as a loading control. Data are representative of three independent experiments.

cone C, like wortmannin, may irreversibly bind PI3K and competitively inhibit its enzymatic activity. To test this possibility directly, growing T cells (those not starved of IL-2) were treated with 100 lM wortmannin, hibiscone C, or vehicle, and the extent of Akt phosphorylation was again measured (Fig. 4). We observed no detectable Akt phosphorylation following the administration of hibiscone C, suggesting that hibiscone C competitively inhibits PI3K. Similar results were observed for wortmannin. Wortmannin is known to be a potent inhibitor of PI3K activity. Since the more structurally simple hibiscone C lacks many of the

O O H3 C

CH3 H

CH3

O

H3 CO

O

OH CH3

O

CH3

O

O

O

O

hibiscone C

viridin

O

O O

halenaquinone

Fig. 2. Some previously synthesized furanosteroids.

OH H3 C

O

H3 C

O CH3

H3 C

OH H O CH3

H3 C

H3 C

O

OH

H3 C

O H O CH3

H3 C

O H O CH3

(±)-h ib is c o n e C Scheme 2. Summary of the synthesis of racemic hibiscone C.8c Starting from a commercially available dione, the natural product can be prepared in seven steps through a series of annulations and oxidation state adjustments.

C. Besley et al. / Bioorganic & Medicinal Chemistry Letters 27 (2017) 3087–3091

Fig. 4. Hibsicone C competitively inhibits PI3K pathway activity. Growing activated T cells (growing) were treated with 100 lM wortmannin, 100 lM racemic hibiscone C, or vehicle (DMSO) for the times indicated. PI3K pathway activation was determined by detection of phosphorylated Akt. Detection of actin serves as a loading control. Data are representative of three independent experiments.

Fig. 5. Dose response analysis of hibiscone C. Growing activated T cells (growing) were treated with the indicated doses of wortmannin, racemic hibiscone C, or vehicle (DMSO) for 1 h. PI3K pathway activation was determined by detection of phosphorylated Akt. Detection of actin serves as a loading control. Data are representative of two independent experiments.

functional groups present in wortmannin, including the oxygenated functional groups of wortmannin that are known to interact with the ATP binding pocket of PI3K,13 we suspected that the absence of these groups in hibiscone C may have a deleterious effect on its potency relative to that of wortmannin. To assess this we performed a dose-response analysis on growing T cells treated with varying concentrations of wortmannin or hibiscone C ranging from 100 lM to 0.01 lM (Fig. 5). While wortmannin showed inhibition of Akt phosphorylation at all doses tested, we were not able to detect inhibition of PI3K activity at doses below 10 lM. These results suggest that the additional functional groups present on wortmannin do, in fact, meaningfully contribute to its potency relative to hibiscone C. Ultimately, the utility of wortmannin as a chemotherapeutic agent rests on its ability to kill cells. To determine if hibiscone C is also cytotoxic, growing T cells were treated with wortmannin, hibiscone C, or vehicle, and viability was measured over a twenty-four hour period (Fig. 6A and B). While we observed reduction in viability with doses as low as 1 lM for wortmannin, only the maximum administered dose of hibiscone C (100 lM) resulted in a reduction in viability. These findings are consistent with our dose-response observations related to Akt phosphorylation. Although hibiscone C is not cytotoxic at lower concentrations, it is possible that hibiscone C could slow cell proliferation at such concentrations. To text this possibility, proliferation of growing T cells was monitored in the presence of wortmannin, hibiscone C,

3089

or vehicle over a 24 h period (Fig. 6C and D). Doses of hibiscone C with concentrations as low as 0.01 lM resulted in a reduction in cell proliferation in the first 18 h of culture, similar to what was observed with wortmannin. We have demonstrated that hibiscone C inhibits PI3K activity in intact T cells (Fig. 3). Furthermore, the ability of hibiscone C to competitively inhibit PI3K activity suggests it binds irreversibly to PI3K (Fig. 4). These observations, which are consistent with those found for wortmannin, are not surprising because hibiscone C contains the same electrophilic diacylfuran that has been shown in wortmannin to bond covalently in the ATP binding pocket of PI3K. However, we observed that hibiscone C was less potent than wortmannin (Fig. 3). This finding is consistent with a number of observations regarding other members of the furanosteroid family. First, a wortmannin homolog with an altered diacylfuran subunit (Fig. 1a), is able to inhibit PI3K activity, albeit with reduced activity.6a Second, a diallylamino product mimic (Fig. 1b) lacks this functional group entirely yet retains biological activity.6b These observations suggest that other functional groups within these furanosteroids contribute to their ability to inhibit PI3K activity. Indeed, non-covalent binding points between PI3K and wortmannin have been mapped, which include S806, Y867, V882, and D964.13 Residues Y867, V882, and D964 interact with the diacylfuran of wortmannin, whereas S806 reacts with a different region of the molecule. Hibiscone C contains the functional groups that bind to all of these residues, except S806. Thus, our data suggest that the interaction between PI3K and the diacylfuran ring of hibiscone C is sufficient for inhibition and that the S806 interaction is necessary to achieve heightened inhibition of PI3K activity. As such, furanosesquiterpenoids like hibiscone C that are less structurally complex than wortmannin should retain some, but not all, of the activity of wortmannin. Importantly, it is known that in addition to PI3K wortmannin can inhibit the enzymatic activity of the Polo-like kinases,14 which have been shown to play critical roles in facilitating cell cycle progression. While we did not detect biological activity of hibiscone C below doses of 10 lM as assessed by Akt phosphorylation (Fig. 5) or cytotoxicity (Fig. 6B), we did observe that hibiscone C affects cell proliferation at lower doses (Fig. 6D). These observations suggest that hibiscone C, like wortmannin, may also inhibit Polo-like kinases. Evaluating this possibility will be an important component of gaining a more thorough understating of its mechanism of action. Moreover, resolving racemic hibiscone C into enantiopure materials will permit more precise quantification of the bioactivity of the natural enantiomer.8d There are several limitations on the use of wortmannin and wortmannin analogs as chemotherapeutics, which helps explain why only one derivative of wortmannin has been advanced to clinical trials (Fig. 1b).15 In addition to its toxicity, wortmannin is virtually insoluble in neutral buffers and slowly hydrolyzes to an inactive form at room temperature in aqueous solution. Furthermore, the low quantities of wortmannin available from natural sources coupled with the limited efficiency of existing total synthesis efforts renders wortmannin a challenging core structure for medicinal chemistry efforts. In contrast, hibiscone C lacks the hydrolytically unstable ester of wortmannin and retains PI3K inhibition activity. Furthermore, as the structurally simplest known member of the furanosteroid family of natural products, the furanosesquiterpenoid hibiscone C can be synthesized with relative ease,8c which will facilitate the preparation of hibiscone C analogs and derivatives. We therefore suggest that hibiscone C may serve as a valuable structural core for the discovery and investigation of new furanosteroid derivatives with enhanced stability and biological activity.

3090

C. Besley et al. / Bioorganic & Medicinal Chemistry Letters 27 (2017) 3087–3091

Fig. 6. Hibiscone C slows cell proliferation and is cytotoxic. Duplicate samples of growing T cells were treated with DMSO (control) or the indicated doses of wortmannin or racemic hibiscone C. At the indicated times, cells were counted (C and D) and viability (A and B) was measured by trypan blue exclusion (see Supporting Information).

Acknowledgements Funding from the National Science Foundation – United States (RUI: 1213569), The Camille and Henry Dreyfus Foundation – United States (Henry Dreyfus Teacher-Scholar Award to BG), and Furman University is gratefully acknowledged. Valuable preliminary contributions from Hudson Tryon are also acknowledged.

6.

A. Supplementary data

7.

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.bmcl.2017.05. 041.

8.

References 1. (a) Garcia-Echeverria C, Sellers WR. Oncogene. 2008;27:5511–5526; (b) Wu P, Hu Y. Med Chem Commun. 2012;3:1337–1355; (c) Falasca M, Hamilton JR, Selvadurai M, Sundaram K, Adamska A, Thompson PE. J Med Chem. 2017;60:47–65. 2. Hanson JR. Nat Prod Rep. 1995;12:381–384. 3. Wipf P, Halter R. Org Biomol Chem. 2005;3:2053–2061. 4. Wymann MP, Bulgarelli-Leva G, Zvelebil MJ, et al. Mol Cell Biol. 1996;16:1722–1733. 5. (a) MacMillan J, Vanstone AE, Yeboah SK. Chem Commun. 1968;613–614; (b) MacMillan J, Simpson TJ, Yeboah SK. J Chem Soc Chem Commun. 1972;1063; (c) MacMillan J, Simpson TJ, Vanstone AE, Yeboah SK. J Chem Soc Perkin Trans 1. 1972;2892–2898;

9.

(d) MacMillan J, Vanstone AE, Yeboah SK. J Chem Soc Perkin Trans 1. 1972;2898–2903; (e) Yuan H, Barnes KR, Weissleder R, Cantley L, Josephson L. Chem Biol. 2007;14:321–328. (a) Norman BH, Paschal J, Vlahos CJ. Bioorg Med Chem Lett. 1995;5:1183–1186; (b) Hong DS, Bowles DW, Falchook GS, et al. Clin Cancer Res. 2012;18:4173–4182; (c) Wipf P, Minion DJ, Halter RJ, et al. Org Biomol Chem. 2004;2:1911–1920; (d) Zask A, Kaplan J, Toral-Barza L, et al. J Med Chem. 2008;51:1319–1323; (e) Sun D, Prasad BAB, Schuber Jr PT, et al. Bioorg Med Chem. 2013;21:5182–5187. (a) Mizutani T, Honzawa S, Tosaki S, Shibasaki M. Angew Chem Int Ed. 2002;41:4680–4682; (b) Jacobi PA, Konekamp T, Mascall KC, O’Connor RT, Onyango EO, Sessions EH. Adv Heterocycl Chem. 2013;100:119–143. Hibiscone C: (a) Koft ER, Smith III AB. J Am Chem Soc. 1982;104:5568–5570; (b) Kraus GA, Wan Z. Synlett. 2000;3:363–364; (c) Ungureanu A, Meadows M, Smith J, Duff B, Burgess JM, Goess BC. Tetrahedron Lett. 2011;13:226–228; (d) Lu Y, Yuan H, Zhou S, Luo T. Org Lett. 2017;19:620–623; Viridin: (e) Anderson EA, Alexanian EJ, Sorensen EJ. Angew Chem Int Ed. 2004;43:1998–2001; Halenaquinone: (f) Harada N, Sugioka T, Ando Y, Uda H, Kuriki T. J Am Chem Soc. 1998;8483–8487; (g) Kojima A, Takemoto T, Sodeoka M, Shibasaki M. J Org Chem. 1996;61:4876–4877; (h) Sutherland HS, Souza FES, Rodrigo RGA. J Org Chem. 2001;66:3639–3641; (i) Kienzler MA, Suseno S, Trauner D. J Am Chem Soc. 2008;130:8604–8605. (a) Joshi KC, Singh P, Pardasani RT, Pelter A, Ward RS, Reinhardt R. Tetrahedron Lett. 1978;19:4719–4722; (b) Ferreira MA, King TJ, Ali S, Thomson RH. J Chem Soc Perkin Trans 1. 1980;249–256; (c) Zhang X, Zhang S, Xuan L. Heterocycles. 2008;75:661–668.

C. Besley et al. / Bioorganic & Medicinal Chemistry Letters 27 (2017) 3087–3091 10. 11. 12. 13. 14.

Gerriets VA, Rathmell JC. Trends Immunol. 2012;33:168–173. Frauwirth KA, Thompson CB. J Immunol. 2004;172:4661–4665. Cantrell D. Semin Immunol. 2002;14:19–26. Walker EH, Pacold ME, Perisic O, et al. Mol Cell. 2000;6:909–919. Liu Y, Shreder KR, Gai W, Corral S, Ferris DK, Rosenblum JS. Chem Biol. 2005;12:99–107.

3091

15. (a) Workman P, Clarke PA, Raynaud FI, van Montfort RLM. Cancer Res. 2010;70:2146–2157; (b) Karve S, Werner ME, Sukumar R, et al. Proc Natl Acad Sci USA. 2012;109:8230–8235.