Design, synthesis and biological evaluation of a macrocyclic discodermolide/dictyostatin hybrid

Abstract

A 22-membered macrocyclic discodermolide/dictyostatin hybrid has been designed and synthesised; biological evaluation against a range of human cancer cell lines revealed significant levels of growth inhibition.

---

Discodermolide (1, Fig. 1), originally isolated from the deep-sea sponge Discodermia dissoluta, displays potent antiproliferative activity against a wide range of human cancer cell lines and inhibits the growth of drug-resistant solid tumours.¹ It shares a similar microtubule-stabilising mechanism to that of Taxol, while having a greater tubulin binding affinity,² and has progressed into clinical development as a novel anticancer agent.³ Furthermore, the synergistic combination of Taxol and discodermolide induces tumour regressions and suppresses angiogenesis in animal models of ovarian carcinoma, supporting their potential use together in cancer therapeutics.⁴

Fig. 1 Structures of discodermolide, dictyostatin and designed hybrid 3. Overlay of the lowest energy conformation of hybrid 3 (purple) and the X-ray structure of discodermolide (green).

Similarly, dictyostatin (2) displays further elevated levels of growth inhibition across the same wide range of human cancer cell lines, and has emerged as a new microtubule-stabilising agent with promising anticancer properties.^5,6 Strong structural similarities exist between discodermolide and dictyostatin, particularly with regard to stereochemical homology, as determined by our recent configurational assignment of the latter structure using detailed NMR analysis,^5c suggesting that they interact in a similar fashion with the Taxol binding site on β-tubulin.^5d,6 Thus, building on initial encouraging findings reported by the Curran group,^7a constraining the conformation of the more flexible open-chain structure of discodermolide into the macrocyclic ring motif of dictyostatin might provide active hybrids of these marine-sponge-derived polyketides . Herein, we report the synthesis of the designed 22-membered macrolide 3, incorporating the full C2–C24 linear sequence of discodermolide and the (Z)-enoate of dictyostatin, and that it shows significant growth inhibitory activity against human cancer cell lines.

At the outset, we considered it essential to achieve a suitable overlay of the energetically preferred conformation of our designed hybrid with that of discodermolide. A 10 000 step Monte Carlo conformational search was performed using Macromodel (Version 8.0) with the MM2* force field and a Born/surface area (GB/SA) water solvent model. The calculated global minimum for analogue 3 (see the ESI† ) correlates well with the X-ray crystal structure of discodermolide,^1a as shown in the overlay in Fig. 1. The match for the C9–C26 region (corresponding to C7–C24 of discodermolide) region is particularly striking. Consequently, if the tubulin-bound conformation of discodermolide resembles its X-ray structure, the constrained macrocyclic analogue 3 may possess a similar binding affinity and cytotoxicity to that of the natural product.

Our synthetic strategy leading to analogue 3 was adapted from previous work on the total synthesis of discodermolide.⁸ A complex aldol coupling of enone 4 and aldehyde 5, derived from previously prepared advanced intermediates, would thus be used to assemble a suitable linear precursor for macrolactonisation (Scheme 1).

Scheme 1 Retrosynthetic analysis of hybrid analogue 3.

Synthesis of the enone 4 started from bis-PMB ether 6,^8a which incorporates the stereochemistry required for the C11–C26 region (Scheme 2). Selective removal of the primary PMB ether was achieved with BCl₃·DMS,⁹ followed by oxidation of the resulting alcohol with TEMPO/PhI(OAc)₂ and Still–Gennari olefination¹⁰ to give enone 4 (73%). Synthesis of the aldehyde partner 5 began with the selective primary oxidation of diol 7,^8b,c again making use of TEMPO/PhI(OAc)₂, followed by a Still–Gennari olefination to provide (Z)-enoate 8 (60%). Removal of the PMB ether,¹¹ bis-silylation of the corresponding diol with TBSOTf/2,6-lutidine and DIBALH reduction of the methyl ester generated allylic alcohol 9 (48%). Finally, PMB ether formation with PMBTCA/Sc(OTf)₃, selective primary TBS cleavage and Dess–Martin oxidation provided aldehyde 5 (68%) in readiness for the pivotal aldol coupling step.

Scheme 2 Synthesis of enone 4 and aldehyde 5.

Enolisation of enone 4 with c-Hex₂BCl/Et₃N in Et₂O, and subsequent addition of aldehyde 5, gave adduct 10 with >95 : 5 dr in favour of the desired (7S)-adduct (48%, Scheme 3). This anti-Felkin–Anh outcome is consistent with our earlier work on 1,6-induction in similar boron aldol reactions, and can be rationalised by invoking the favoured transition state model TS 1.^8b,c

Scheme 3 Key boron aldol coupling.

Reduction of β-hydroxyketone 10 proved problematic, affording mixtures of epimeric alcohols at C9 when subjected to standard Evans–Saksena conditions (Scheme 4).¹² An improvement was found by following the same protocol as previously employed for the reduction of similarly troublesome β-hydroxyketones^8b using stoichiometric quantities of (R)-CBS and BH₃·THF,¹³ which provided the desired isomer 11 in a 75 : 25 dr.¹⁴ The 1,3-anti diol 11 was protected as its acetonide ; subsequent oxidative PMB ether cleavage (with DDQ) and sequential oxidation of the resultant alcohol 12 with TEMPO/PhI(OAc)₂ then NaClO₂/NaH₂PO₄ provided seco-acid 13 (58%).

Scheme 4 Completion of the synthesis of hybrid analogue 3.

Finally, a Yamaguchi macrolactonisation of 13 (65%) and global deprotection using 3M HCl in MeOH (53%) afforded the 22-membered macrolide 3 after HPLC purification. Full proton and carbon assignments were carried out using COSY and HMQC NMR experiments. The close correlation between the calculated (by molecular modelling ) and observed ³J_H–H coupling constants of 3 (see the ESI† ) suggests that the conformation adopted is similar to our modelled prediction.

The cell growth inhibitory activity of macrocyclic discodermolide/dictyostatin hybrid 3 was evaluated in vitro against three cancer cell lines: MDA-MB-231 (breast), A549 (non-small cell lung) and HT29 (colon) (Table 1). Notably, analogue 3 displayed significant antiproliferative activity against these human carcinoma cells, with a cytotoxicity around one-tenth that of discodermolide. This preliminary data is consistent with the conformation adopted by 3 closely resembling the X-ray structure of discodermolide, which itself has now been reported to correlate strongly with the NMR-derived conformation of discodermolide bound to tubulin.¹⁵

Table 1 Human cancer cell growth inhibitory properties of macrocyclic discodermolide analogue 3 relative to discodermolide (1)

	GI50/µM^a
	MDA-MB-231 (breast)	A549 (non-small cell lung)	HT29 (colon)
a 50% growth inhibitory concentration after 72 h of continuous incubation.
1	0.029	0.020	0.015
3	0.208	0.399	0.170

---

In conclusion, we have designed and synthesised the most active macrocyclic discodermolide/dictyostatin hybrid reported to date.^7,16 The encouraging antiproliferative activity of analogue 3 can be attributed to its constrained (dictyostatin-like) macrocyclic structure, which bears a strong resemblance to the bioactive conformation of discodermolide.¹⁵ In ongoing work, this rational design approach is being extended to the synthesis of further novel hybrid analogues of discodermolide and dictyostatin.

Financial support was provided by the EPSRC and AstraZeneca. We thank Dr. Carmen Cuevas (PharmaMar, Madrid) for providing the biological data, Dr. Stuart Mickel (Novartis) for the gift of chemicals, Rob Paton and Dr. Jonathan Goodman (Cambridge) for assistance with the modelling studies, and Dr. John Leonard (AstraZeneca) and Dr. Isabelle Lyothier (Cambridge) for helpful discussions.

Notes and references

1. (a) S. P. Gunasekera, M. Gunasekera, R. E. Longley and G. K. Schulte, J. Org. Chem., 1990, 55, 4912 . CCDC reference VINTAN for the X-ray crystal structure of 1; (b) E. ter Haar, R. J. Kowalski, E. Hamel, C. M. Lin, R. E. Longley, S. P. Gunasekera, H. S. Rosenkranz and B. W. Day, Biochemistry, 1996, 35, 243.
2. R. J. Kowalski, P. Giannakakou, S. P. Gunasekera, R. E. Longley, B. W. Day and E. Hamel, Mol. Pharmacol., 1997, 52, 613.
3. A. Mita, A. C. Lockhart, T.-L. Chen, K. Bochinski, J. Curtright, W. Cooper, L. Hammond, M. Rothenberg, E. Rowinsky and S. Sharma, J. Clin. Oncol., 2004, 22, 2025.
4. G. S. Huang, L. L. Barcons, B. S. Freeze, A. B. Smith, III, G. L. Goldberg, S. B. Horwitz and H. M. McDaid, Clin. Cancer Res., 2006, 12, 298.
5. (a) G. R. Pettit, Z. A. Cichacz, F. Gao, M. R. Boyd and J. M. Schmidt, J. Chem. Soc., Chem. Commun., 1994, 1111; (b) R. A. Isbrucker, J. Cummins, S. A. Pomponi, R. E. Longley and A. E. Wright, Biochem. Pharamacol., 2003, 66, 75; (c) I. Paterson, R. Britton, O. Delgado and A. E. Wright, Chem. Commun., 2004, 632; (d) I. Paterson, R. Britton, O. Delgado, A. Meyer and K. G. Poullennec, Angew. Chem., Int. Ed., 2004, 43, 4629.
6. C. Madiraju, M. C. Edler, E. Hamel, B. S. Raccor, R. Balachandran, G. Zhu, K. A. Giuliano, A. Vogt, Y. Shin, J. Fournier, Y. Fukui, A. M. Brückner, D. P. Curran and B. W. Day, Biochemistry, 2005, 44, 15053; R. M. Buey, I. Barasoain, E. Jackson, A. Meyer, P. Giannakakou, I. Paterson, S. Mooberry, J. M. Andreu and J. F. Diaz, Chem. Biol., 2005, 12, 1269.
7. (a) Y. Shin, N. Choy, R. Balachandran, C. Madiraju, B. W. Day and D. P. Curran, Org. Lett., 2002, 4, 4443; (b) For an excellent review on natural product hybrids, see: L. F. Tietze, H. P. Bell and S. Chandrasekhar, Angew. Chem., Int. Ed., 2003, 42, 3996.
8. (a) I. Paterson, G. J. Florence, K. Gerlach, J. P. Scott and N. Sereinig, J. Am. Chem. Soc., 2001, 123, 9535; (b) I. Paterson, O. Delgado, G. J. Florence, I. Lyothier, M. O'Brien, J. P. Scott and N. Sereinig, J. Org. Chem., 2005, 70, 150; (c) I. Paterson, O. Delgado, G. J. Florence, I. Lyothier, J. P. Scott and N. Sereinig, Org. Lett., 2003, 5, 35.
9. M. S. Congreve, E. C. Davison, M. A. M. Fuhry, A. B. Holmes, A. N. Payne, R. A. Robinson and S. E. Ward, Synlett, 1993, 663.
10. W. C. Still and C. Gennari, Tetrahedron Lett., 1983, 24, 4405.
11. Attempts to remove the PMB ether using DDQ resulted in both PMP acetal formation and over-oxidation to the benzoate.
12. D. A. Evans, K. T. Chapman and E. M. Carreira, J. Am. Chem. Soc., 1988, 110, 3560 . NaBH(OAc)₃ gave 67 : 33 dr in favour of the 1,3-syn-diol while Me₄NBH(OAc)₃ gave little or no diastereoselectivity.
13. E. J. Corey, R. K. Bakshi and S. Shibata, J. Am. Chem. Soc., 1987, 109, 5551.
14. After chromatographic separation, the syn-diol was subsequently used to synthesise the C9 epimer of 3.
15. V. M. Sanchez-Pedregal, K. Kubicek, J. Meiler, I. Lyothier, I. Paterson and T. Carlomagno, Angew. Chem., Int. Ed., 2006, 45, 7388.
16. For recent reports of non-macrocyclic discodermolide analogues, see inter alia: I. Paterson and O. Delgado, Tetrahedron Lett., 2003, 44, 8877; A. B. Smith, III and M. Xian, Org. Lett., 2005, 7, 5229; A. B. Smith, III, B. S. Freeze, M. J. LaMarche, T. Hirose, I. Brouard, M. Xian, K. F. Sundermann, S. J. Shaw, M. A. Burlingame, S. B. Horwitz and D. Myles, Org. Lett., 2005, 7, 315; S. J. Shaw, K. F. Sundermann, M. A. Burlingame, D. C. Myles, B. S. Freeze, M. Xian, I. Brouard and A. B. Smith, III, J. Am. Chem. Soc., 2005, 127, 6532; J. M. Minguez, S. Y. Kim, K. A. Giuliano, R. Balachandran, C. Madiraju, B. W. Day and D. P. Curran, Bioorg. Med. Chem., 2003, 11, 3335; S. P. Gunasekera, R. E. Longley and R. A. Isbrucker, J. Nat. Prod., 2002, 65, 1830.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental and modelling details, and ¹H/¹³C NMR data for compounds 3 and 10–15. See DOI: 10.1039/b615122a

---