key: cord-0703632-m3mhj6op
authors: Pérard-Viret, Joëlle; Quteishat, Laith; Alsalim, Rana; Royer, Jacques; Dumas, Françoise
title: Cephalotaxus Alkaloids
date: 2017-08-16
journal: Alkaloids Chem Biol
DOI: 10.1016/bs.alkal.2017.07.001
sha: 374504f9a761e35a31fadeb37cee82e2d02eca22
doc_id: 703632
cord_uid: m3mhj6op

Cephalotaxus alkaloids represent a family of plant secondary metabolites known for 60 years. Significant activity against leukemia in mice was demonstrated for extracts of Cephalotaxus. Cephalotaxine (CET) (1), the major alkaloid of this series was isolated from Cephalotaxus drupacea species by Paudler in 1963. The subsequent discovery of promising antitumor activity among new Cephalotaxus derivatives reported by Chinese, Japanese, and American teams triggered extensive structure elucidation and biological studies in this family. The structural feature of this cephalotaxane family relies mainly on its tetracyclic alkaloid backbone, which comprises an azaspiranic 1-azaspiro[4.4]nonane unit (rings C and D) and a benzazepine ring system (rings A and B), which is linked by its C3 alcohol function to a chiral oxygenated side chain by a carboxylic function alpha to a tetrasubstituted carbon center. The botanical distribution of these alkaloids is limited to the Cephalotaxus genus (Cephalotaxaceae). The scope of biological activities of the Cephalotaxus alkaloids is mainly centered on the antileukemic activity of homoharringtonine (HHT) (2), which in particular demonstrated marked benefits in the treatment of orphan myeloid leukemia and was approved as soon as 2009 by European Medicine Agency and by US Food and Drug Administration in 2012. Its exact mechanism of action was partly elucidated and it was early recognized that HHT (2) inhibited protein synthesis at the level of the ribosome machinery. Interestingly, after a latency period of two decades, the topic of Cephalotaxus alkaloids reemerged as a prolific source of new natural structures. To date, more than 70 compounds have been identified and characterized. Synthetic studies also regained attention during the past two decades, and numerous methodologies were developed to access the first semisynthetic HHT (2) of high purity suitable for clinical studies, and then high grade enantiomerically pure CET (1), HHT (2), and analogs.

This contribution is an updated and expanded version of the two preceding chapters published in this series in 1984 1 and 1998. 2 Other reviews focused on the structure and synthesis of these compounds, 3e6 and several reviews are available that cover the structureeactivity relationships (SARs), biological, and clinical properties of these alkaloids. 7e19 The literature is reviewed from the last update in this series, which was covering the field till mid-1997, to early 2016. Because they usually exist under the form of unique enantiomer, we numbered the natural products by an arabical digit irrespective to their optical rotation sign, e.g., cephalotaxine (CET) (1) , their antipode by ent-digit, e.g., ent-cephalotaxine ent- (1) , and the corresponding racemic compound usually obtained by chemical synthesis rac-digit, e.g., rac-cephalotaxine rac- (1) . In the asymmetric synthesis section, we have precised the sign of the optical rotation, e.g., CET (À)-(1), ent-CET (þ)- (1) , and rac-CET (AE)- (1) .

For more than 50 years, chemists have studied the extraction, structure determination, biogenesis, and synthesis of CET (1) (Fig. 2) , the major alkaloid of this series, and related alkaloids, because of the reported biological activity of its natural esters 2e5, in view of their therapeutical use in the treatment of leukemia. Homoharringtonine (HHT) (2) and harringtonine (HT) (3), or a mixture of both, have been used in traditional Chinese medicine (TCM) since the 1970s to treat a variety of malignancies. HHT (2) being the easiest to purify and the most abundant in the plant was selected by the National Cancer Institute (NCI) for clinical development. HHT (2) , also known as omacetaxine, a simple ester of CET (1), probably holds the record for the longest time to reach the market 20 after having experienced increasing interest since its first use in TCM, and efforts by pharmaceutical companies for its development as an antileukemic agent. Despite its clinical efficacy, natural HHT (2) was not recorded until now, simply because normal extract formulations have raised several unresolved problems, such as the reproduction of a constant product from defined extraction procedure, the purity profile, the main source of the raw material in China, and biodiversity conservation issues. The Cephalotaxus shrubs with extremely slow growth are mostly protected and endangered species. Moreover, CET (1) and its active esters are present in small quantities in the plant (roots, bark, heartwood, and leaves), which hardly makes possible to exploit it to obtain HHT (2) or its related esters. Semisynthetic HHT (ssHHT) (2) (Myelostat, Ceflatonin, or Omacetaxine mepesuccinate (OM)) is manufactured by a semisynthetic process developed by Robin from naturally occurring alkaloid CET (À)-(1) extracted from the dry leaves of Cephalotaxus harringtonia, and a synthetic precursor side chain. 21 Although HHT (2) was investigated for the treatment of chronic myeloid leukemia (CML) in the 1990s with good results, the advent of Imatinib (IM) mesylate, a breakpoint cluster region and Abelson (BCR-ABL)1 tyrosine kinase inhibitor (TKI) at that time delayed the interest for HHT (2) . However, de novo or acquired resistance to TKIs 22 renewed the interest for this alternative first-in-class simple natural ester of CET (1) . HHT (2) is currently approved in Europe and in the United States, due to its reported activity in patients with CML resistant to currently available TKIs. As a result, on September 2, 2004, orphan designation (EU/3/04/224) was granted by the European Union for OM (International nonproprietary name (INN), trade name Synribo)/HHT (2) for the treatment of CML to Stragen France SAS, France, for the treatment of CML. Exploitation rights were transferred to ChemGenex Europe SAS, France, in January 2009 and subsequently to Teva Pharma GmbH, Germany, in December 2012. Due to their different mechanisms of action, combinations with TKIs represent attractive treatment options. Recently, on October 26, 2012, HHT (2), Synribo (formerly Ceflatonin, Omapro), received an accelerated approval by the US Food and Drug Administration (FDA) for adult patients with chronic-phase (CP) or accelerated-phase (AP) CML with resistance or intolerance to two or more TKIs. The US FDA has granted full approval to HHT (2) (drug substance INN OM, Synribo) for injection, based on the final analysis of two phase II trials evaluating the efficacy and tolerability data of omacetaxine in 2014. This renewed interest for isolation, characterization, biological evaluation, and total synthesis of cephalotaxus alkaloids and analogs to suppress the dependance on natural source supply. cephalotaxane skeleton with either one or two units covalently linked forming dimers having two identical or different units (Fig. 3) . Cephalotaxanes (6) include neutral compounds or salts having this basic skeleton in common and may contain various oxygenated substituents (aliphatic ethers or aromatic units, free or esterified alcohols, substituted or free enols, phenol, and others). Cephalotaxanes include CETs characterized by a free alcohol function on ring D (R ¼ H), and HTs where the 3-OH group is esterified (R s H). Isolation of cephalotaxanes and homoerythrinane alkaloids [skeleton (7) ] from the same plant suggested that they might be formed through similar biogenetic routes. 1,2 Hainanensine (8) is a structurally unique Cephalotaxus alkaloid devoid of optical activity and lacking the spiranic tetrasubstituted carbon center. It was isolated from Cephalotaxus hainanensis and C. fortunei (Fig. 3) and displays a type of single rearranged CET skeleton. 28 Figure 3 Cephalotaxanes (6), the skeleton (7) of minor homoerythrinane alkaloids, and the structurally unique Cephalotaxus alkaloid hainanensine (8) .

Until 1997, 36 compounds related to the CET series (Tables 1 and 2) , were described, of which 23 are natural esters and 1 is dimeric.

CET (1) and HHT (2) were also isolated conventionally from fruits, leaves, and branches of Cephalotaxus sinensis. 29 Among the alkaloids derived from CET (1), compounds possess different D rings such as cephalotaxinone (10) , demethylcephalotaxine (11) , and isocephalotaxinone (13) and ring C such as cephalotaxinamide (18) . Oxygenated compounds with modifications on their B ring as 11-hydroxycephalotaxine (14) and 4-hydroxycephalotaxine (15) or compounds having an oxygen bridge between the rings B and D such as drupacine (16) were also described (Fig. 4) . 30e32 11-Hydroxycephalotaxine (14) was converted to drupacine (16) in acidic medium. 31 Among the minor alkaloids isolated from the bark of C. hainanensis Li (in addition to 11 known alkaloids), CET analogs were identified as cephalotaxinone (10) , cephalotaxinamide (18) , and demethylneodrupacine (17) , which possesses an ether bridge between C11 and C3, while drupracine (16) exhibits an ether bridge between C11 and C2. The structure of desmethylcephalotaxinone (12) , isolated from Cephalotaxus harringtonia Table 1 Cephalotaxus alkaloids that have been described from natural sources till mid-1997 1, 2 Cephalotaxine analogs Cephalotaxine (1), C 18 H 21 NO 4 , mp 132e133 C, [a] D À183 (c 0.22, EtOH) Hainanensine (8) , C 17 H 17 NO 4 , mp 240e244 C, [a] D 0 (CH 3 OH) Epicephalotaxine (9) , C 18 H 21 NO 4 , mp 136e137 C, [a] D À150 (c 0.8, CHCI 3 ) Cephalotaxinone (10), C 18 H 19 NO 4 , mp 198e200 C, [a] D À146 (c 0.63, CHCI 3 ) Desmethylcephalotaxine (11) , C 17 H 19 NO 4 , mp 109e111 C, [a] D À110 (c 0.28, CH 3 OH) Demethylcephalotaxinone (12) , C 17 H 17 NO 4 , mp 102e107 C, [a] D þ2. 3 (c 0.52, CH 3 OH) Isocephalotaxinone (13) , C 18 H 19 NO 4 , mp 130.5e132 C, ½a 21 D À47 (c 0.5, CHCI 3 

Oxyoenated cephalotaxines , C 18 H 21 NO 5 , mp 235e242 C, [a] D À139 (c 0.56, CHCI 3 ) 4-Hydroxycephalotaxine (15) , C 18 H 21 NO 5 , m. 135e137 C, ½a 27 D þ120 (c 0.025, CH 3 OH) Drupacine (16) , C 18 H 21 NO 5 , mp 70e72 C, [a] D À137 C (c 0.79, CHCI 3 ) Demethylneodrupacine (17) , C 17 H 19 NO 5, mp 197e200 C Cephalotaxinamide (18) , C 18 H 19 NO 5 , mp 230 C NO 9 , mp 73e75 C, [a] D À104.6 (c 1.0, CHCI 3 ) Isoharringtonine (4), C 28 H 37 NO 9 , mp 69e72.5 C, [a] D À99.6 (c 1.06, CHCI 3 ) Deoxyharringtonine (5), C 28 Esters of oxygenated cephalotaxine analogs Drupangtonine (33), C 28 H 37 NO 9 , oil, [a] D À24 (c 0.06, MeOH) 11a-Hydroxyhomodeoxyharringtonine (34) , C 29 H 39 NO 9 , oil, ½a 25 D À115 (c 0.07, MeOH) 11b-Hydroxyhomodeoxyharringtonine (35) , C 29 H 39 NO 9 , oil, ½a 25 D À153 (c 0.10, MeOH) 11-Hydroxydeoxyharringtonine (36), C 28 H 37 NO 9 , oil, ½a 25 D À77 (c 0.0085, MeOH) (Forbes) K. Koch var. harringtonia cv. fastigiata, was determined by semisynthesis from CET (1) via oxidation to cephalotaxinone (10) and cleavage of the enol ether thus confirming desmethyl CET (12) as a natural metabolite and biosynthetic intermediate. 33 Repeated chromatography of an alkaloidal fraction of the ethanolic extract of Cephalotaxus fortunei Hook f. 34 afforded as minor component (þ)-4-hydroxycephalotaxine (þ)- (15) recently also identified in Cephalotaxus koreana Nakai. 35 Epicephalotaxine (9) is another minor metabolite of C. fortunei Hook f. accompanied by CET (1) (50%e54% of the total alkaloids), cephalotaxinone (10) , (þ)-acetylCET (þ)- (26) , demethylCET (11) , HT (3), and HHT (2) . 36 Molecules coupled with side chains bearing an ester function as in nordeoxyharringtonine (nordeoxyHT) (21) , homodeoxyharringtonine (homodeoxyHT or deoxyHHT) (22) , and bishomodeoxyharringtonine (bishomodeoxyHT) (23) are found among HTs (Fig. 5) . 37 Including esters 2e5, these compounds are esters of CET (1) having its polycyclic core coupled to different side chains deriving from (R)-citramalic acid. Although occurrence of enantiomeric pairs in the same family of plants is unusual, 38 acetylcephalotaxine (acetylCET) (26) was reported under both enantiomeric forms. (À)-AcetylCET (À)- (26) was obtained by semisynthesis through acetylation of natural CET (À)-(1) 24 and was isolated from C. fortunei 30 and C. wilsoniana. 39 , 40 The opposite enantiomer (þ)-acetylCET (þ)- (26) was isolated in C. fortunei 41 and C. hainanensis. 42 Although an error cannot be excluded in literature data, these findings need to be confirmed since this natural Cephalotaxus Alkaloids product (þ)-26 is the only one portraiting an opposite absolute configuration (AC) of alkaloid core.

Anhydroharringtonine (20) bearing an additional heterocycle in the side chain is the only natural ester of this type. 43 It induced 98% growth inhibition of P388 leukemia cell lines at 1 mg/mL, a level comparable to that of deoxyHT (5) . There are also molecules coupled with side chains carrying a free acid group (Fig. 5) : 5 0 -des-O-methylHT (30) (harringtonic acid), (2 0 R,3 0 S)-3 0 -hydroxy-5 0 -des-O-methylHT (28) (cephalozemic acid C), 5'-des-O-methylHHT (31) (homoharringtonic acid, also identified as a metabolite of HHT (2) in rats and rabbits) and 5 0 -des-O-methylisoHT (27) (isoharringtonic acid). 1, 2 Others possess an ester function and an aromatic ring, as in homoneoHT (24) and (2 0 R,3 0 S)-hydroxyneoHT (25) . 44 Some of these esters deliver biological activities comparable to those of the first four isolated esters 2e5, but they are available in much lower quantities in the plant. The first dimeric Cephalotaxus alkaloid, cephalotaxidine (37) , was isolated from C. harringtonia var. drupacea (C. drupacea) in 1996. 45 It is a dimer of HHT (2) and homoharringtonamide (32) with an IC 50 of 1.8 mg/mL against P-388 murine leukemic cell lines. Drupangtonine (33) 46 was the most active (IC 50 7 nM against P-388 murine leukemic cell lines) of the four esters of oxygenated CET analogs, 11a-hydroxy-homodeoxyHT (34) , 11b-hydroxyhomodeoxyHT (35) , and 11b-hydroxydeoxyHT (36) (IC 50 0.38, 0.33, and 0.17 mg/mL, respectively, against P-388 murine leukemic cell lines) identified from C. harringtonia var. drupacea (Fig. 6) . 47 Since the last reviews in this series, 1,2 more than 30 Cephalotaxus alkaloids have been isolated from various Cephalotaxus sp. These new alkaloids include the following compounds. 

N-oxides of CET, CET a N-oxide (38) , CET b N-oxide (39) , and 11bhydroxycephalotaxine b N-oxide (40) , together with isocephalotaxine (41) were identified and characterized in the ethyl acetate extract of C. fortunei fruits by Jossang et al. in 2001 (Fig. 7) . 48 Although it was described as a synthetic compound earlier, 49 the relative configuration of isocephalotaxine (41) isolated during this work remains unknown due to the lack of Nuclear Overhauser effect spectroscopy (NOESY) correlations. These alkaloids exhibited low cytotoxicity against nasopharynx KB cells with IC 50 values of 30, 14, 31, and 15 mg/mL, respectively. ACs of cephalezomines G (42) and H (43) described by Kobayashi and Morita 50 from the leaves of Cephalotaxus harringtonia var. harringtonia were determined by circular dichroism (CD) as 2S,3R for cephalezomine G (42) and 2R,3R for cephalezomine H (43) . However, the structure of cephalezomine H (43) , initially assigned as a trans diol was revised to a syn-b,b-diol after total synthesis (Fig. 7) . 51 Some uncertainty remained, however, due to the divergent optical rotations of the synthetic and the natural compounds.

A new stereoisomer (44) of desmethylcephalotaxinone (12) has been isolated from the leaves and heartwood of Formosan C. wilsoniana (Fig. 8) . 52 Although the UV, 1H NMR, and ½a 23 D of compound 44 were similar to those of natural desmethylcephalotaxinone (12) , C1 to C7 and C10 signals in the 13 Cephalotaxus Alkaloids synthetic sample of 12. 100 Evidence that compound 44 was a stereoisomer of desmethylcephalotaxinone (12) was also based on H1a and H6a cross-peak in the NOESY correlation chart of 44. While this correlation seems doubtful, since only one stereogenic carbon center is present in structures 12 and 44, these findings imply that the second stereogenic center is the nitrogen atom that lost its fast inversion properties, as observed in the X-ray studies showing the rigid structure of CET (1) . 53 Compound 44 showed no cytotoxic activity against the Hep G2, MCF-7, Hep 3B, and HT-29 human cancer cell lines.

A phytochemical investigation of the branches and leaves of Cephalotaxus lanceolata, a Cephalotaxus species native to Yunnan Province, China, resulted in the isolation of three new Cephalotaxus alkaloids, cephalancetines A (45) and B (46) , and unsymmetrical dimer cephalancetine D (47) together with known cephalancetine C (48) isolated for the first time from this plant, along with nine other known alkaloids (Fig. 8) . 54 The structure of cephalancetine C (48) was unambiguously confirmed by single-crystal X-ray diffraction as consisting of CET (1) linked at C8 to C7 (C26) of cephalotaxinamide (18) thus establishing its AC as (3S,4S,5R,8S,22S,23S,24R,26S). Cephalancetine D (47) is a dimer consisting of drupacine (16) linked at C8 to C7 (C26) of cephalotaxinamide (18) . Cephalancetine C (48) was identical to the saponification product of bis-cephalezomines (see Fig. 12 ) with LiOH. 55 All these isolated alkaloids were tested for their cytotoxicities against four human tumor cell lines, A549, HCT116, SK-BR-3, and HepG2 that evidenced that HT (3) and HHT (2) can strongly inhibit the proliferation of these human tumor cell lines as well, however with limited potency against SK-BR-3 cells (Table 3) . 54 HT (3) showed remarkable cytotoxicity against A549, HCT116, and HepG2 cell lines with IC 50 values of 0.15, 0.054, and 0.38 mg/mL, respectively. HHT (2) also exhibited potent cytotoxicity against A549, HCT116, and HepG2 cell lines with IC 50 values of 0.085, 0.001 and 0.087 mg/mL, respectively, 85 times stronger than that of doxorubicin (DOX) against HCT116 cell line. Comparing the relationship between the structures with their cytotoxicities, the ester side chains at C3 of the CET skeleton seem to play the most prominent role in the cytotoxic activity of these alkaloids.

Four new cephalotaxus alkaloids, cephalotines AeD (49)e(52) were isolated from the leaves and twigs of C. lanceolata and Cephalotaxus fortunei var. alpina (Yunnan province, P.R. China) along with 24 known alkaloids ( Fig. 9 ). None of these compounds showed significant activity against HeLa, SGC-7901 gastric cancer, and A-549 lung cancer cell lines (IC 50 > 20 mM). 56 

In 2000, Kobayashi et al. isolated from the leaves of Cephalotaxus harringtonia var. nana new cytotoxic alkaloids, cephalezomines A (53) and B (54) , which are esters of drupacine (16) , as well as cephalezomines CeF (55)e (58) , which are esters of CET (1) (Fig. 10 ). 57 These new compounds were found together with seven known alkaloid esters, HT (3), isoHT (4), deoxyHT (5), homodeoxyHT (22) , CET (1), and related CET alkaloids, demethylCET (11), 30 11b-hydroxyCET (14) , and drupacine (16) . Their structure and stereochemistry were elucidated by spectroscopic data, CD, and X-ray diffraction analysis.

Among the new Cephalotaxus alkaloids possessing a CET type backbone with various side chains at C3, cephalezomine D (56) is the first one with (2 0 R,3 0 R) AC. In this study 57 the team of Kobayashi also described the cytotoxicity of these alkaloids against L1210 and KB cell lines (Table 4) . Two years later, Kobayashi and Morita reported the isolation of five new alkaloids, cephalezomines G, H, and JeL from the leaves of C. harringtonia var. nana (Sapporo, Japan). Among them, cephalezomines K (59) and L (60) are esters of CET (1) epimers at C4" (Fig. 10) . 58 In 2010 bis-cephalezomine A (65), homodeoxyHT (22) , and two HHT N-oxides, HHT a N-oxide (62) and HHT b N-oxide (63), previously prepared by Takano 60, 61 for SAR studies (Fig. 11 ). HHT (2) and HT (3) showed pronounced antiplasmodial activity against Plasmodium falciparum 3D7 but to a lesser extent against Leishmania major (Table 5) , while deoxyHT (5) and homodeoxyHT (22) were active in the two tests. Figure 11 Oxygenated Cephalotaxus esters and Cephalozemine J.

HHT a N-oxide (62) and HHT b N-oxide (63) reported for the first time as natural products, did not display any antileishmanial or cytotoxic activity against A549 cell line, despite of an esterified 3-OH function in their structures (Table 5) . SAR directed toward antiplasmodial activity of isolated Cephalotaxus alkaloids was also discussed. Cephalezomine J (64) from the leaves of C. harringtonia var. nana (Fig. 11) is the unique CET conjugate with a sugar type side chain (D-glucose). 50 Although the data collected in this study are cytotoxicity values (Table 5) , a selectivity index was calculated to rate this effect on human versus parasitic cell lines. In all the cases, the cytotoxicity was greater against parasites, rather on tumor cells (SI > 1).

In 2009, Kobayashi et al. isolated five new dimeric alkaloids from C. harringtonia var. nana, especially heterodimers, named bis-cephalezomines AeE (65)e(69) (Fig. 12) . 55 The alkaloids that constitute the dimers (HHT (2), HT (3), deoxyHT (5) , and cephalezomine A (53)) showed potent cytotoxic 

The new alkaloid cephalocyclidin A (72) with an unprecedented polycyclic rearranged skeleton derived from CET (1) and six contiguous asymmetric centers was isolated from fruits of C. harringtonia var. nana. 62 ACs at C2 and C3 were assigned by the exciton chirality method as 2R and 3S, respectively, indicating 1R,2R,3S,4S,5R,11S configuration, compatible with the absolute stereochemistry of the hydrochloride of CET (1) deduced through the Flack parameter of 0.01 (5) , in the X-ray analysis. It could be derived biogenetically by formation of the of cephalocyclidin A (72) against the L1210 mouse lymphoma and human epidermoid carcinoma cells KB is 0.85 and 0.80 mg/mL respectively.

The proportion of the four major natural esters 2e5 with antitumor activity is between 0.7% in the needles of C. fortunei and 36% in the roots of C. harringtonia var. harringtonia of the total amount of alkaloids. This plant is the best natural source for these four compounds of pharmacological interest. At present, the plants of the genus Cephalotaxus are endangered species and face the threat of extinction. Only the renewable parts of the plant are used for the extraction of alkaloids. For example, the use of modern methods of extraction provides up to 2 kg of pure CET (1) from a ton of dry leaves, that is 0.2% of dry matter. Robin et al. studied a method of preparation of HHT (2) by semisynthesis from natural CET (1) by attachment of the side chain (see Section 3). Purification provided a product, the purity of which was 99.8% against 98.5% for NCI products and 95%e97% for products of Chinese origin (Table 7) . 63 For example, 550 g of HHT (2) was injected into a preparative high-performance liquid chromatography (HPLC) system (diameter 45 cm, height 1 m containing 48 kg of ODS stationary phase). The elution was performed using a gradient of buffered solution of pH 3 methanol-water as mobile phase (flow rate 540 L/h, 1200 psi). The collected fractions (20e50 L) were separated by UV detection and analyzed by HPLC. After eliminating fractions with about 0.5% of the total content of HHT (2), fractions suitable for drug specification (purity > 99.8%) were combined. HPLC analysis of the organic phase after HHT recovery (basified to pH 8.5 and extracted with dichloromethane) showed a level of impurities less than 0.5%. HHT (2) (210 g) was then recrystallized by dissolution in 240 mL of methanol at 30 C and filtered at 0.25 microns for sterilization, and then deionized water (2.4 L) was added and methanol was distilled off. The aqueous solution of HHT (2) DS (quality "drug substance") or "drug substance" quality (DS) was held in a decontaminated rotary evaporator until the appearance of white crystals of pure HHT (2) . After filtration and drying under vacuum at 60 C for 2 days, 88% overall yield (OY) of HHT (2) was obtained when compared to its potential content in the semisynthetic gross HHT (2) . This corresponds to a clinical batch allowing for the preparation of 40,000 therapeutic dosage units containing 5 mg containing less than 0.03% of impurities. 63 Numerous methods were described for the extraction and purification of CET (1), its natural esters, and related analogs. Extraction of fruits, leaves, or branches of C. sinensis with 90% ethanol under refluxing conditions followed by acid base treatment and purification on alumina allowed for the isolation of CET (1) and HHT (2) . 64 High speed countercurrent chromatography, a form of liquideliquid partition chromatography in which the stationary liquid phase is retained in the apparatus without the use of a solid support, was performed with a pH gradient elution to shorten the duration of the separation and improve resolution, allowing a very efficient separation of crude alkaloid extract (800 mg) from C. fortunei to yield drupacine (16) (9.3 mg), wilsonine (15.9 mg), CET (1) (130.4 mg), epi-wilsonine (64.8 mg), fortuneine (12.8 mg) , and acetylCET (26) 65 The recovery of each alkaloid was greater than 90%. The extraction step plays an important role in the overall purification process of HHT (2) . Traditional methods of isolation and purification of the chemical constituents of plant tissues have some disadvantages. These approaches require very long extraction times and usually a large amount of solvent with sometimes low efficiency. Furthermore, many natural products are thermally unstable and may degrade during hot extraction. To overcome these drawbacks, a new method of extracting HHT (2) from Cephalotaxus koreana using microwave extraction (MWE) was developed by Kim et al. 66 The MWE process of HHT (2) recovery (biomass/MeOH 1/6 m/v, 250 rpm) gave a performance of the first biomass extraction 25% higher than the conventional solvent extraction (CSE) method for which at least four extractions were required. It was possible to recover more than 99% of HHT (2) extracting once the biomass by MWE at 40 C for 15 min or 50 C for 5 min. Because the content of tars and waxy compounds were also increased, an adsorbent was added to the biomass (1/1.5 m/m) during the MWE extraction. The use of the MWE process for the extraction of HHT (2) when compared to a CSE one has many advantages such as shorter extraction time, reduced amount of solvent, and higher yield resulting in a less costly production system for a superior product. Ultrasonic-MWE of C. fortunei produced CET (1) in a simple, highly efficient, and low cost process, suitable for its industrial production in high purity after two recrystallizations from anhydrous ethanol. 67 Using a prepurification method with adsorption prior chromatography, HHT (2) with over 52% purity can be obtained simply from C. koreana biomass in good yield, minimizing the use of solvents, scale, and complexity of operations for the final HPLC purification of HHT (2) ( Table 6) . 68 This process allows increasing the purity of the raw HHT (2) from 0.5% to over 52% with three simple steps and inexpensive pretreatment of the MeOH extract.

A recent study by Robin related a method for isolation and purification of CET (1) from the crude alkaloid extract from 10 kg dry leaves (not from the bark) of C. fortunei var alpina (Sichuan, China). 69 CET (1) purification was performed for the first time by reverse phase HPLC. The crude base (24.5 g) containing 71% of CET (1) was dissolved in a mobile phase consisting of ortho-phosphoric acid and deionized water containing 1.55% triethylamine (pH 3). The acid additive constrained the alkalization of the obtained aqueous phase and its extraction with dichloromethane. CET (1) (18 g) was fully recovered from the crude alkaloid extract as a white solid with an HPLC purity greater than 95%. The limited supply of promising secondary metabolites from their natural sources is a major hurdle to their pharmaceutical development; however, strategies to overcome this problem were developed. For instance, to increase the yield of natural anticancer drug alkaloids from natural sources, methods of extraction are developed. Recovering CET (1) from C. harringtonia, var. fastigiata with vegetable oil in the presence of an aqueous carbonate aqueous phase yielded 5% CET (1) with a purity of 26%. 70 The extraction performed with a mixture of scCO 2 (supercritical CO 2 ) with MeOH-10% Et 2 NH yielded 0.0029% CET (1) of undescribed purity. 71 HHT (2) was also obtained from C. koreana. 72 He et al. have developed an HPLC analysis method for HHT (2) determination, 73 allowing to reveal three congeners present as impurities in the HHT (2) samples. Two impurities are new: a-isoHHT (73) is an isomer of HHT (2) in which the hydroxyl group is alpha to the isopropyl group and ethyl-HHT (74) , the ethyl ester of HHT (2) , is an isolation artifact of HHT (2) with ethyl acetate, the third being HT (3) Table 6 The prepurification summary (Steps 2e4) of homoharringtonine (HHT) biomass (16 g) 64 Steps

HHT (2) (Table 7) . a-IsoHHT (73) was found concentrated in the crystallization mother liquors of HHT (2) (10.7% against 1%e3% in crude extracts). The HPLC analysis according to He's protocol therefore allows the determination of HHT (2) in crude samples with a linear response from 4 to 12 mg and a limit of quantitation (LOQ) of 27 mg. It was used to assess the purity of several HHT (2) commercial batches. 63 A tandem HPLC-electron-spray mass spectrometry (MSeMS) method was developed for the analysis of alkaloids contained in the leaves of C. harringtonia. Nine alkaloids bearing an ester group were detected with a good sensitivity. 74 Content in HT (3) was higher than in HHT (2) in the needles of Cephalotaxus griffithii alkaloid fraction (CGAF) (122.14 and 16.79 mg/g of CGAF, respectively) by HPLC analysis. 75 Another major concern is the enantiomeric purity of CET (1), which has a direct effect on the diastereoisomeric purity of its natural esters. A process for purification of HHT (2) was early developed in view of the clinical studies, 63 although recently, a new process for the purification of CET (1) allowed the resolution of racemic or partially racemized samples through 69 This study also pointed out that the cristalline structure of HHT (2) deposited at the Cambridge Crystallographic Data Centre (CCDC) exhibits the reverse AC to the commonly accepted one. Eventually, selection of a Cephalotaxus cultivar grown in Europe allowed high recovering of CET (1) (up to 113e139 g of highly pure cristalline CET (1) from 25 kg of dry leaves). 69 Similarly, ammonium salts of HT (3) and HHT (2) were characterized and, as expected, protonation occurred at N9, 76 a finding in accordance with biological studies of N-oxides by Takano 61 and mechanistic studies by Steitz (see Section 6) .

Apart from the extraction of the natural source, plant cell cultures represent alternative, efficient, and sustainable chemical factories for the production of biologically important secondary metabolites, especially when the natural source become endangered. The production is limited by the low yields associated with harvest and the high costs are associated with complex chemical synthesis. Cephalotaxus evergreen trees are somewhat difficult to propagate and they are relatively rare. Some of them have been listed in the red list of IUCN (International Union for Conservation of Nature) as national plant of second-grade protection in China and are "near threatened" (Cephalotaxus latifolia), "vulnerable" (Cephalotaxus oliveri, Cephalotaxus mannii) or rare and "endangered" (C. lanceolata or Cephalotaxus fortunei Hook. var. lanceolata (K.M.Feng) Silba, C. hainanensis) species. 77 A variety of strategies are being developed to overcome the limitation of low product yields. A renewed interest in the study of Cephalotaxus crops 78 and the development of methods for analyzing the composition of different plants have been stimulated by the need for raw materials and the search for natural sources rich in CET (1). An HPLC protocol was used to determine the content of CET (1), HT (3), and HHT (2) in crop roots and calluses of C. harringtonia. As expected, CET (1) is the major alkaloid (10 mg/kg dry matter) while HT (3) and HHT (2) are more present in the roots (6.6 and 7.5 mg/kg, respectively). 79 In plant cell and tissue culture, cell growth and product formation are not necessarily correlated. Cell culture studies have sought to identify factors that tend to yield low yields and slow metabolic rates. Zhang et al. proposed the technique of periodic oscillation in temperature between 10 and 25 C every 12 h for 45 days to overcome this key problem. 80 The total production of HT (3), HHT (2) , and isoHT (4) (1.22 mg/L) in solid cultures (SoC) of C. fortunei was enhanced by 1.8-and 1.3-fold compared to the controls at constant temperatures of 10 or 25 C. When subjected to such an oscillation every 24 h for 30 days, suspension cultures (SuC) gave a total alkaloid production of 0.18 mg/L, a 2.0-and 1.05-fold improvement compared to the SuC controls at 10 or 25 C, respectively. HT (3) represents the main alkaloid (72%e92% for SoC in 45 days, and 50%e67% for SuC for 30 days) while the HHT (2) content is only 4%e21% in SoC and 23%e40% in SuC, the remaining being isoHT (4).

In 2014, Li showed that the combination of 10 mg/L sodium fluoride (NaF) as glycometabolic regulator and 100 mmol/L methyl jasmonate (MJ) as the elicitor both promoted the biosynthesis of CET (1) in C. mannii suspension cells. 81 NaF was more effective to increase cell membrane permeability and product secretion compared to MJ. The yield of HT (3) in the cells was 4.8 greater under NaF þ MJ conditions than in the control, 1.7-fold greater under NaF and 1.6-fold under MJ conditions, respectively (Table 8) . No HHT (2) was found in NaF þ MJetreated cells. The product release rates were 0%, 78%, 24%, and 62% in control, NaF, MJ, and NaF þ MJ treatment, respectively. The combined NaF and MJ treatment on cells provided an efficient strategy for producing CET (1) .

Another study of the effects and action mechanisms of NaF on the growth and CET (1) production of C. mannii suspension cells indicated that NaF acted as an inhibitor of the EmbdeneMeyerhof pathway (EMP), not as an elicitor to promote CET (1) production. 82 

The Cephalotaxus alkaloids have drawn the attention of many chemists due to their unique structure and potent antileukemic activities of certain ester derivatives they contain. Excellent review articles on this subject have been published in this series by Huang and Xue, 1 Miah et al. 2 , and more recently among others, by Nay et al. 6 Since 1997, 41 syntheses of CET (1) have been reported, 18 of which being enantioselective.

CET (1) can be described as the superposition of two structural units relevant for synthetic strategies, namely the pyrrolobenzazepine core ABC 75 and the synthetically challenging azaspiro unit CD 76 (Scheme 2). The three stereogenic centers (C3, C4, and C5) are contiguous on the D ring. Weinreb showed that the stereochemistry of the B/D ring fusion (C4/ C5) is thermodynamically favored. In addition, carbonyl reduction of cephalotaxinone (10) delivers CET (1) with the natural configuration at C3. 83 33 (23 since 1997 are described in this review) syntheses of CET (1) concerned its racemic form ( Table 9 ). The number of asymmetric syntheses has grown considerably from only 3 before 1997 to 18 until 2015, inflecting the development of asymmetric synthesis, and the efforts to provide a fully synthetic access to the drug ( Table 10 ). The present section describes the recent synthetic strategies organized according to the main carbonecarbon bond disconnections of the central azepine ring B (Fig. 13 ). Asymmetric syntheses are also classified based on the type of asymmetric center control.

The key steps are highlighted for each strategy. Most of these syntheses are formal; nevertheless, the OY of each synthesis up to CET (1) was evaluated. Eight racemic and five asymmetric syntheses of CET (1) are total syntheses, with only one racemic and three asymmetric routes described since mid-1997, with experimental OY and enantiomeric excess (ee). The OY and Cephalotaxus Alkaloids the number of steps (NS) are given from commercially available reagents for comparison purpose. In addition, for convergent syntheses, total number of steps (TNS), which represents the total efforts to be accomplished up to the target molecule, and the corresponding average yield (AY) are given. Synopses of the syntheses of CET (1) reported so far are presented in Table 9 (racemic) and Table 10 (asymmetric). Besides formal syntheses targeting Weinreb's enamine grouped together in Table 9 , racemic syntheses are classified according to the type of disconnection of ring B (aee, Fig. 13 ), the type of intermediate in the formal syntheses and then in a chronological order. The asymmetric syntheses are presented in chronological order in Table 10 . It is apparent from Table 9 that most of the syntheses of rac-CET (AE)-(1) are linear (only 10 being convergent) and proceed predominantly (19 of 33) by the C4eC13 disconnection a. The shortest access to CET (1) was described in the first synthesis of CET (1) by Weinreb et al. 103 in 11 steps including the preparation of homopiperonylic acid (89) in two steps from piperonal (82) . 134 The most effective total syntheses were described by Kuehne et al. 94 in only 11 steps and 11.2% OY (average 82% yield per step) or 12 steps and 17.1% OY, by Ishibashi and Ikeda et al. 138 115 The favorite choices as ring A precursors are piperonal (82) and homopiperonylic acid (89) with C11 and the methylenedioxy group already at the right position. Table 10 shows that proline (140) is a widely used chiral source (5 over 21 syntheses), either in its natural L form, or in its unnatural D one. Herein, 15 syntheses are convergent and 6 are linear. The most efficient total syntheses of CET (À)-(1) (OY > 5%) were reported by Weinreb et al. 103 OY and 4.5% AY in 16e18 steps, 99% ee), Ikeda et al. 137 (5. 1% OY and AY in 19 steps, 88% ee), Dumas and d'Angelo 135 (5. 4% OY in 20 steps, >95% ee; 4.1% AY in 22 TNS), El Bialy et al. 140 (12. 2% OY in 14 steps, 89% ee; 4.3% AY in 20 TNS) and Trost et al. 129 (13. 9% OY in 19 steps, 91% ee; 8.4% AY in 21 TNS) were also efficient.

The ring closure C4eC13 a (Fig. 14) to form the azepine cycle was frequently adopted after Semmelhack's early extensive developments in the second synthesis of rac-CET (AE)-(1). 83 3.2.1. 1 Tietze's Formal Synthesis (1997) In 1997, Tietze et al. 84 Cephalotaxus Alkaloids and 5.2% OY from 1,3-cyclopentanedione (77) (2.5% AY in 15 TNS). It should be mentioned that Mori's protocols differ for the synthesis of racemic or optically active CET (1). Indeed, the transformation of desmethyl CET (12) into cephalotaxinone (10) using 2,2-dimethoxypropane and p-TsOH led to racemization of the product (76% yield). On the contrary, when p-TsOH and methyl orthoformate were used for this transformation at room temperature (rt), limited racemization occurred (ee 88%), but the yield went down to 47%.

Suga and Yoshida developed an approach to the azaspiro-compound 101 via ring-closing metathesis (RCM) of a functionalized pyrrolidine 100 generated by electrochemical oxidation of a,a-bis(trimethylsilyl)pyrrolidine 95 (Scheme 4). 86, 87 First, homopiperonylic acid (89) was used to prepare the A unit 91 bearing a nosylate and an aromatic iodide suitable for ring B annulation, relying on Semmelhack's procedure. 83 , 88 The two silyl groups were introduced by Beak's methodology 89 (deprotonation with sec-BuLi and quenching with TMSCl) furnishing compound 94, which was protected into N-methoxycarbonyl compound 95. Its anodic oxidation at low temperature yielded N-acyliminium ion 96, reacting in situ with homoallylmagnesium bromide (97) furnishing pyrolidine 98. This latter compound was next submitted to a second anodic oxidation followed by direct trapping with the organozinc reagent 99 to provide the diolefin 100. Diene 100 was converted to azaspiro CD unit 101 via ring closing metathesis promoted by Grubbs I catalyst. After deprotection using TMSI, the spiranic amine intermediate reacted with nosylate 91 obtained from homopiperonylic acid (89) following Semmelhack's procedure. 88 The so-formed aryl iodide (AE)-102 was submitted to an intramolecular Heck-type ring closing reaction according to Tietze 84 Regioselective bromination of homopiperonylic acid (89) yielded bromocarboxylic acid 103 converted to Weinreb's amide, 2,83 and then reduced by DIBAL to furnish aldehyde 104. Different routes were proposed to prepare the ACD ringecontaining aryl bromide 86. The best one starts from b-ketoester 105 transformed by reduction to allylic alcohol 106, 91 which was treated with triethyl orthoacetate under Johnson orthoestere Claisen conditions to give ethyl ester 107. Afterward, ester 107 was converted in three steps into sulfonamide 108 by successive reduction by LiAlH 4 , tosylation, and nucleophilic displacement by tosylamine. The The g-butyrolactam 113 was prepared from 2-arylethanol 90 in 65% OY by Mitsunobu's substitution of the primary alcohol, followed by partial reduction of the crude succinimide 112 with NaBH 4 and reductive dehydroxylation of the resulting hemiaminal. Huang's methodology, namely activation of compound 113 by triflic anhydride (2,6-di-tertbutyl-4-methylpyridine (DTBMP)) was the best base for this reaction, followed by successive addition of buten-1-yl Grignard (97) 84 and employed by Suga and Yoshida. 86 Once again, four steps (Mori's protocol) 85 have to be performed to obtain rac-CET (AE)-(1), to give a 14-steps pathway in 5.4% estimated OY from homopiperonylic acid (89). Reaction of 1,2-epoxycyclopentane 116 with in situ generated dimethylsulfonium methylide gave allylic alcohol 106, which was directly used for the subsequent JohnsoneClaisen rearrangement. The resulting ester 107 was subjected to amidation reaction conditions with hydroxylamine to afford hydroxamic acid 117. The nitroso-ene cyclization was accomplished on oxidation with tetrapropylammonium periodate via a highly reactive acylnitroso intermediate 118 delivering N-hydroxy-1-azaspiro [4.4] non-2-one 119, which was trapped as an acetate 120 yielding in moderate yield spiro-lactam (AE)-121 90 on SmI 2 reduction. Alkylation of the sodium salt of (AE)-121 with iodoaryl tosylate 115 gave quantitatively (AE)-122. The subsequent Heck reaction promoted by HerrmanneBeller catalyst 87 proceeded smoothly to deliver the ABCD Kuehne's intermediate 94 (AE)-123, and the corresponding regioisomeric product (AE)-124 in 82% combined yield, based on recovered starting material (brsm) as an inseparable 54:46 mixture. rac-CET (AE)-(1) could be obtained in 12 steps from cyclopentene oxide (116) and 6.2% OY provided that separation of the regioisomers is effective and four supplementary steps applied to transform the ABCD intermediate (AE)-123 (TNS 15 steps, AY 2.5%, 4.3% brsm). Although Huang and Wang described a rapid access to the CD spirocyclic amide (AE)-121, Heck ring closure of ring B led to a mixture of two regioisomers and decreasing dramatically the yield. as key transformations. 99 We describe here the best route that led to Li's intermediate 97 (AE)-148 via the spirolactone 154 (Scheme 10). Spirolactone 154 was obtained from methyl prolinate (AE)-150 by N-allylation with allyl bromide, leading to allylamine 151, which underwent a [2,3]-Stevens rearrangement of the resulting quaternary ammonium salt by reaction with benzyl bromide providing N-benzyl pyrrolidine 152. After lactonization of 152 under acidic conditions followed by debenzylation through catalytic hydrogenation, spirolactone 154 was acylated with 3,4-dimethoxyphenacetyl chloride (145) to produce the amide spirolactone 155 containing the AC ring units, which was selectively reduced with LiBH 4 to diol 156 in 63.9% OY. Swern oxidation of diol 156 furnished the amido keto-aldehyde 157, which was treated with potassium carbonate to give the amido spirocyclopentenone (AE)-148 with the ACD ring units. This formal synthesis is especially suitable for the production of (AE)-148 at a scale up to 10e50 g. Li 97 and Hanaoka's 98 procedures (6 steps, 15% OY) would allow to complete the synthesis of rac-CET (AE)-(1) in 15 steps and 7.1% OY from rac-proline (AE)- (140) . 

In 2013, Zhang et al. proposed two pathways for the formal synthesis of rac-CET (AE)-(1) targeting Li's enone (AE)-174 100 and Hanaoka's enone (AE)-180 96 based on conventional alkylations and aldol condensations from known benzazepines 101 165 and 166 obtained in four steps from aryl amine 83 (bearing a methylenedioxy substituent) or 158 (bearing two methoxy substituents), respectively (Scheme 11). 102 The C ring was elaborated via an N-and C-alkylating annulation by treatment of benzazepine 165 with 1,3-dibromopropane (167) Cs 2 CO 3 giving product 168 in one step. ABC dione 172 was prepared by alkylation either with allyl bromide followed by Wacker oxidation or more efficiently with 2-methoxyallyl bromide (170) followed by acidic hydrolysis (Scheme 11). Aldol condensation efficiently constructed the CET (1) core in (AE)-173. The dimethoxy groups were transformed into the required methylenedioxy group by the usual protocol 85 In an alternative synthesis, these authors also studied the use of methylenedioxybenzazepine 166 as starting material (Scheme 12).

Ring C was elaborated by acylation/alkylation of benzazepine 166 with 3-chloropropanoyl chloride (175) to furnish the ABC ketone 176. Treatment with sodium hydride and 2-methoxyallyl bromide (170) produced the enol ether 177 (74%) as a major O-alkylated product along with the desired C-alkylation derivative 178 (22%). Claisen rearrangement allowed conversion of 177 to 178, which was exposed to acid hydrolysis to give dione 179. Intramolecular aldol condensation provided the known amido enone (AE)-180 96 following Hanaoka' synthesis 96 in a total of 16 steps from homopiperonylic acid (89) and 2.5% estimated OY.

Bubnov et al. developed a convenient and practical methodology for the preparation of various spiro-b-amino alcohols. This procedure was applied to a formal synthesis of rac-CET (AE)-(1) involving allylboration and RCM to prepare the CD ring containing azaspirobicyclic compound 183 (Scheme 13). 104 2-Pyrrolidone (181) Starting from piperonyl amine 83, diester 189 was obtained in four steps via a BishlereNapieralski cyclization followed by hydrogenation and N-alkylation with ethyl 4-bromobutyrate. A second N-alkylation with allyl bromide led to the ammonium salt intermediate 190, which was transformed via a 2,3-sigmatropic Stevens rearrangement into diester 191. Eventually, compound 193 was obtained by a Dieckmann cyclization with subsequent decarboxylation followed by Wacker oxidation. Aldol condensation of the dione 193 provided the spirocyclic AD enone 194 (Scheme 14). CET (1) B and C rings were formed by reductive rearrangement in acetic acid via the aziridinium ion 195. The obtained ABCD ketone (AE)-197 (revised stereochemistry at C4 as drawn) was oxidized with phenyliodonium diacetate to give acetal (AE)-198, which, after hydrolysis, was submitted to autoxidative dehydrogenation with potassium t-butoxide and oxygen to provide the desmethylcephalotaxinone (AE)-12. Formally synthesis of rac-CET (AE)-(1) could be achieved according to Weinreb 103 using two additional steps, leading to a formal total synthesis in 1.8% calculated OY over 18 steps from piperonal (82) .

An alternative approach to ketone (AE)-197 was described from allyl ketone 192. Clemmensen reductive rearrangement of this ketone first led to benzazepine derivative 199 (Scheme 15). 100 After Wacker oxidation of compound 199, application of Polonovskie Potier reaction conditions to 200 gave the enamine 201. It was previously reported that cyclization of this intermediate to (AE)-174 (formally an endocyclic enamine annulation) could only be realized in low yield. Li found that an aerobic treatment of 201 with Fe 2þ salt in acetic acid promoted the cyclization into ABCD enone (AE)-174 in good yield. This cyclization can be seen as an unusual azo-Nazarov-type cyclization. A catalytic hydrogenation eventually gave ketone (AE)-197. This synthesis in six steps and 14.5%e17% OY from 192 is not advantageous when compared to the three steps and 40.5% OY version presented in Scheme 14. Thus, formal synthesis of rac-CET (AE)-(1) following this procedure could be theoretically achieved in 21 steps and 0.7%e0.8% OY from piperonal (82) .

In 2005, Li et al. described a formal synthesis of the DolbyeWeinreb enamine 103,106 203 from dione 193, with a mild ClemmenseneClemoe PrelogeLéonard reductive rearrangement of 193 as the key step (Scheme 16). 105 Dehydrogenation of the benzazepine 202 by Hg(OAc) 2 gave enamine 203, which could be likely transformed in rac-CET (AE)-(1) by Weinreb's strategy 103 in expected 0.4% OY for 17 steps from piperonal (82) . Reductive rearrangement of aminodione 193 also produced aminoketone 200 in 15% yield, which could be processed into demethylcephalotaxinone (12) and then into CET (1) through a nine-stage sequence. This constitutes another formal synthesis of CET (1) with an expected OY of 0.17% in 21 steps from piperonal (82) .

In 2005, Li reported a novel synthesis of the amino enone 194 involving annulation of the readily available aminodione 193 (Scheme 17). 107 BischlereNapierralski condensation of amine 83 with d-valerolactone (204) mediated by POCl 3 gave the iminium ion 205 in 74% yield. Treatment of 205 with aqueous NaOH generated a labile enamine 206, which was alkylated with bromoacetone to give the keto enamine 207 (20%) and the iminium bromide 208 (50%). Both were transformed by treatment with 70% perchloric acid into iminium perchlorate 209, which on exposure to NaOH furnished the unexpected dione 193 (25%) and annulated cis-cyclopentanone 210 100 193 based on an intramolecular Schmidt reaction of symmetric azido dione 221 (Scheme 18). 108 The synthesis started with compound 213 (ring A precursor) obtained from homopiperonylic acid (89) by reduction, bromation, and chlorination, 109 and iodocyclopentenone 215 (ring D precursor) prepared by iodination of cyclopentenone (214) (Scheme 18). 110 Cephalotaxus Alkaloids (Z)-isomer (14%). Mild hydride reduction led to hydroxy enone 232 presumably via the oxonium ion 231, which could have underwent rapid and effective 4p electrocyclization. Acetylation of the hydroxyl function of the resulting compound 232 followed by selective cleavage of the N-Troc group by zinc reduction led to the spontaneous formation of ABCD unit 233 as a single diastereomer by an intramolecular aza-Michael addition. Air-oxidation in the presence of potassium t-butoxide gave rac-demethylcephalotaxinone (AE)- (12) . Application of Weinreb's method 103 would allow to synthesize rac-CET (AE)-(1) in 13 steps from piperonal (82) and 0.4% OY. submitted to mild Sonogashira conditions to afford the alkyne 235. The intermolecular PausoneKhand reaction of alkyne 235 with ethylene could generate two regioisomeric products. Only the desired isomer 236 (58% yield) was obtained under optimized conditions using n-butyl methyl sulfide as a promoter with dimethyl sulfoxide as the oxidant. Oxidation of alcohol 236 followed by reductive amination produced the amine 238, which was advanced to the 10-membered ring compound 239 by intramolecular N-alkylation. Mariano's intermediate 113, 114 240 was prepared successively by deprotection and reprotection with a t-butoxycarbonyl group. This strategy led quickly and efficiently to macrocyclic precursor 240 in eight steps and 20% OY. With this compound 240 in hand, a formal synthesis of rac-CET (AE)-(1) was thus attained, and additional five-step sequence 113,114 would give rac-CET (AE)-(1) over 15 steps from homopiperonylic acid (89) and 4.3% estimated OY. Recently, Hong et al. reported two formal syntheses 115 involving an unusual C4eC5 disconnection b 116e118 for ring B formation (Fig. 16 ). The key steps were first an N-acyliminium ion cyclization, which established stereogenic centers C5 and C4 and the formation of B ring. Then an RCM furnished the spiro-ring D. The two syntheses started from iodo alcohol 90 prepared from homopiperonylic acid (89) in two steps.

Sonogashira coupling of aryl iodide 90 with propargyl trimethylsilane catalyzed by palladium(II) and copper(I) iodide delivered alkyne 242 in excellent yield (97%) (Scheme 21). After a catalyst screening, partial hydrogenation was carried out with "P-2 Nickel" and ethylenediamine (Brown catalyst) leading to (Z)-243 in quantitative yield (dr > 15/1). Mitsunobu reaction with succinimide, followed by allylation next gave the hemiaminal 246, which, treated with TiCl 4 at low temperature in mesitylene, allowed for the generation of an N-acyliminium ion that cyclized to furnish the ABC unit 247 (cis/trans 2.5/1). Zhan's catalyst-1B 119 was the most effective for the RCM reaction to deliver Kuehne intermediate 94 (AE)-123. Formally, synthesis of rac-CET (AE)-(1) could be achieved by Kuehne's protocol 94 in 12 steps from homopiperonylic acid (89) and 23.5% calculated OY.

Since iminium ion cyclization showed a moderate diastereoselectivity and the separation of the two diastereomers by silica gel chromatography failed, this route was less attractive. To improve it, authors investigated cyclization from the E-allylsilane 249 obtained by Suzuki coupling between compound 248 and aryl iodide 90 catalyzed by palladium (Scheme 22). As previously seen, a Mitsunobu reaction gave 250. Subsequently, 250 was treated with a Grignard reagent and cyclized by a HosomieSakurai reaction to 251.

Through a TiCl 4 -promoted iminium ion cyclization, the ABC pyrrolobenzazepine 251 was obtained with a ratio of 20/1 in favor of the trans-isomer. The RCM reaction of the trans-isomer of amide 251 did not happen whatever the catalysts used. Reduction of the amide led to a less rigid tertiary amine 252, suitable for the desired cyclization. Metathesis proceeded smoothly in the presence of camphor sulfonic acid (1 equiv.) to protonate the amine and Zhan-1B's catalyst (1 mol %). The ABCD ring system 253 was obtained with inverted relative stereochemistry at C4. Dihydroxylation and Swern oxidation afforded the known demethylcephalotaxinone (AE)- (12) . Application of an improved Weinreb's procedure 103 would generate rac-CET (AE)-(1) in 12 steps and 26.2% expected OY from homopiperonylic acid (89). to end up the synthesis of rac-CET (AE)-(1) (Scheme 23). 120 An unusual N9eC10 bond formation through an amidation reaction closed B ring (Fig. 17 ) . Aldehyde 256 derived in five steps from L-proline (140) 

Synthesis of optically pure CET (1) remains a challenge for organic chemists. Three contiguous asymmetric centers including a tetrasubstituted one on D ring have to be controlled. It was early noticed by Weinreb 103 that racemic syntheses led to only one diastereoisomer without any required separation. This means that the configurations of the three asymmetric centers are mutually controlled. Strategies using disconnection C4eC5 for B ring formation exemplify this phenomenon. During the cyclization step, the thermodynamically more stable BD ring junction is cis, which is the relative C4eC5 configuration found in natural CET. Moreover, Weinreb also showed that NaBH 4 reduction of cephalotaxinone (10) is stereoselective (Scheme 24), delivering the required relative stereochemistry of the hydroxyl function at C3 of CET (1), which is assumed to be formed as a result of hydride attack from the convex face of 10.

It is thus possible to control successively the C4 center via the C5 tetrasubstituted center; then the C3 center via the C4 stereocenter, and Cephalotaxus Alkaloids consequently, control of the tetrasubstituted C5 stereocenter is the only necessary one to synthetize enantiomerically pure CET (1). Most of asymmetric syntheses are based on the control of this spiro center followed by a stereoselective cyclization affording the B ring and C4 control, a final reduction establishing the desired configuration at C3. The asymmetric syntheses of CET (1) described below are classified according to the type of stereochemical control (asymmetric catalysis, use of chiral auxiliaries, incorporation of chiral building blocks, and resolution strategies) and deconnection of ring B, related strategies then in chronological order in each section. treatment with a large excess of Grignard reagent prepared from chloroalcohol 265 and tosylation (three steps, 40% OY). Chiral bromoacetate 268 corresponding to the future D ring was obtained by reduction with diborane in the presence of catalytic amounts of the chiral oxazaborolidine 267, followed by acetylation of the allylic alcohol. A 87% ee was measured by NMR using Eu(hfc) 3 . Formation of the B and C rings was achieved by the strategy described for rac-CET (AE)-(1), employing two successive Pd-catalyzed transformations. The optically active form of Mori's intermediate 85 (À)-88 was isolated. The synthesis of CET (À)-(1) could be likely achieved in 13-steps linear sequence, 3.1% calculated OY, and 87% expected ee from cyclopentane-1,3-dione (77). Application of the Tietze's procedure led to Mori's alkene (À)-88 in 80% ee determined by chiral HPLC. Surprisingly, during its purification by column chromatography on silica gel, authors found that enantiomers could be separated leading to (À)-88 in 35.5% yield with >99% ee. This phenomenon called "self-dispropornation of enantiomers" results from different affinities of aggregates of compound 88 on silica gel. 123 Tu thus accomplished a formal synthesis of CET (À)-(1) in 14 steps, 3.3% calculated OY and 80% ee (18 TNS, 2.5% OY). The high purity material (estimated 99% ee) could be obtained in 1.7% calculated OY from trimethylsilylacetylene (269) .

In 2012, Renaud et al. reported two variants for the formal synthesis of CET (À)-(1). 124 Both syntheses were based on a stereoselective carboazidation of methylene cyclopentanol derivatives. In the first approach, which will be described later (see Section 3.3.5.2, Scheme 43), the starting material was obtained from the racemate by resolution. The second approach was more efficient and started with (S)-cyclopentenol (284) obtained in 96% ee set up via enantioselective Noyori's hydrogenation of propargylic ketone 280 using (S,S)-1,1 0 -bi-2-naphthol 281 to give 282. Compound 282 was obtained in 64.5% yield from 4-pentenoic acid (277) (Scheme 27). 125 Dihydroxylation of silyl enolate (S)-285 afforded the diol 286 as a single diastereoisomer, which was protected as an acetonide. Desilylation, oxidation, and Wittig methylenation afforded the bicyclic alkene 287, which was submitted to radical carboazidation. The by reduction of the azide. First, the stereogenic center at C5 (CET numbering) of (S)-cyclopentenol 284 was used to control the configuration of the newly formed diol at C1eC2, and then it was destroyed on double bond formation in 287. This chiral center was eventually controlled during the carboazidation step leading to 289. This sequence corresponds to the concept of self-regeneration of stereogenic centers. 126 130 The phosphoramidite ligand 296 provided the chirality during the [3 þ 2] cycloaddition, delivering compound (þ)-297 in 91% ee (chiral HPLC). Transformation of compound (AE)-297 into rac-CET (1) had been previously reported by Ikeda, and formally, ent-CET (þ)-(1) (enantiomer of the natural product) could be obtained in 91% expected ee and 14% calculated OY over a 19 steps in the LLS from piperonal (82) . 

In 2002, Royer et al. described an asymmetric synthesis of azaspiro compound (À)-303, containing the CD ring fragment exploiting a stereoselective semipinacolic reaction as the key step. 131 In 2004, these authors developed an enantioselective synthesis of (À)-CET (À)-(1) from spirolactam ketal (À)-306, 132 relying on Kuehne 94 301 furnished the desired spiro compound (À)-303. The N-acyliminium ion 302 was attacked preferably anti to the naphtyl group, delivering compound (À)-303 with a good diastereoisomeric excess (de 80%). The major diastereoisomer was recrystallized and also easily recovered by chromatography, and thus azaspirononanedione (À)-303 was obtained in 70% yield and de surpassing 99%. The ketone function of (À)-303 was protected as a cyclic ketal, and then reductive cleavage of the N-benzylic bond provided spirolactam ketal (À)-304. N-alkylation of its sodium salt with the tosylate 305 of 3,4-methylenedioxypheny-2-ethanol 133,134 followed by ketal hydrolysis afforded the enantiopure ACD intermediate (À)-306 described by Kuehne in its racemic form. 94 Following Kuehne's procedure, compound (À)-306 was transformed into (À)-CET (1) with improved formation of enone (À)-307 using a stoichiometric amount of Pd(OAc) 2 without benzoquinone (83% yield). A further modification consisted in the use of AlH 3 96 at 0 C instead of LiAlH 4 in refluxing tetrahydrofuran (THF) for the reduction of (À)-308 to avoid any racemization of sensitive CET (1). CET (À)-(1) was thus obtained in 98.7% ee and 7.5% OY over a 16-steps linear sequence from 2,5-dimethoxy-2,5-dihydrofuran (299) using (S)-1-(1-naphthyl)ethylamine (298) as a chiral source.

After having explored ABC / ABCD synthetic strategies, 116, 117 in 2005, Dumas and d'Angelo et al. reported a novel enantioselective synthesis of the azaspiro unit (À)-304 containing CD rings described by Royer 132 based on an asymmetric Michael reaction involving a chiral enamine prepared from cheap and recyclable enantiopure (R)-1-phenylethylamine 309 (Scheme 30). 135 As a consequence of stereogenic control of the quaternary carbon center in the resulting (R)-ketodiester 311, the two asymmetric centers C3 and C4 of CET (À)-(1) were formed with their natural ACs. Similarly, use of the antipodal (S)-1-phenylethylamine (309) could afford ent-CET (þ)-(1), although it would lead to biogolically inactive esters. 136 The key step was a Curtius rearrangement of acyl azide 314, which allowed installation with perfect stereochemical fidelity and high efficiency of an a-nitrogen substituent at the tetrasubstituted center in ketodiester 311 to set the tetrasubstituted spirocyclic center.

The ketodiester (R)-311 was elaborated in 85% yield and an ee ! 95% through Michael addition of enamino ester (R)-310 to methyl acrylate. Sequential highly stereoselective Luche reduction of the ketone at low temperature followed by saponification and acidification provided an intermediary hydroxy diacid, which on spontaneous lactonization generated the diastereomerically pure lactone acid 313 obtained in 75% yield. The two ester functions of 311 were differentiated at this stage. The acid 313 was first converted into acyl azide 314, which was directly rearranged into the benzyl carbamate 315 by treatment with benzyl alcohol at 110 C. Saponification followed by oxidation with in situ-generated ruthenium tetroxide, and then esterification of the intermediate acid by diazomethane afforded the ketoester (R)-316. This compound was converted into ketal and then submitted to hydrogenolysis to deliver the target spiroketal amide (À)-304. By relying on Royer 132 and Kuehne's procedures, 94 

In 1999, Ikeda et al. developed an asymmetric version 137 of their racemic synthesis. 130 , 138 The intermediate ABCD ketone (þ)-149 was prepared from unnatural D-proline (R)-140 used to furnish the C ring with C4 atom and to control the C5 tetrasubstituted stereogenic center set in 319 (Scheme 31).

Compound 318 obtained from D-proline (R)-140 according to Seebach 139 was stereoselectively allylated to give oxalolidinone 319, which reacted with 5-lithiobenzodioxole (320), generated from 4-bromobenzodioxole (itself prepared from benzodioxole 330) and n-BuLi at À78 C. 

In 2008, Hayes et al. developed a formal synthesis of CET (À)-(1) based on an alkylidene carbene 1,5-CH insertion reaction as a key step to construct the 1-azaspiro [4.4] nonane CD ring system (338 / 339). 141 In a first approach, vinyl chloride 338 was chosen as a carbene precursor prepared in seven steps and 50% OY from L-proline (S)- (140) via N-Boc-proline methyl ester 331 (Scheme 33). DIBAL-H reduction of 331 followed by Wittig reaction with 1-triphenylphosphoranylidene-2-propanone (332) gave an intermediate E-enone. Catalytic hydrogenation and a second Wittig reaction with chloromethylenephosphorane (334) afforded the vinyl chloride 335. After deprotection of the amine, amide formation with the carboxylic acid 89 and reduction of the resulting amide 337 led to the pivotal alkylidene carbene precursor 338. Its treatment with potassium hexamethyldisilazide caused the key alkylidene carbene 1,5-CH insertion reaction leading to the desired azaspiro [4.4] nonane derivative 339 in good yield (65%e79%). Regioselective iodination of the aromatic ring system gave the Heck-cyclization precursor 340. Its X-ray crystal structure allowed the confirmation of its absolute stereochemistry, thus demonstrating that as expected the 1,5-CH insertion reaction had proceeded with retention of configuration. The Heck cyclization under modified Fu's conditions (Pd(OAc) 2 , HBF 4 $PtBu 3 , Cs 2 CO 3 ) delivered the ABCD olefin 341 in 61% yield, together with enamine 342 (30%). The last stage involved removal of the extra carbon atom at C3 on ring D by decarbonylation of aldehyde 347. Toward this end, alkene 341 was epoxidized using dimethyldioxirane in the presence of BF 3 $Et 2 O. The resulting epoxide 343 was submitted to a regioselective Yamamoto rearrangement to lead to an allylic alcohol protected as an acetate 345, which was transposed by treatment with aqueous HCl to primary alcohol 346 and then oxidized (Swern) into aldehyde 347, which, in the presence of Wilkinson's catalyst, afforded the key Mori's intermediate (À)-88. CET (À)-(1) could be likely obtained using Mori's protocol 85 in this way in 87% expected ee, 23 steps and 0.06%e0.2% calculated OY from L-proline (S)-140.

In their second formal synthesis of CET (À)-(1), an alkylidene carbene 1,5-CH insertion reaction was also used to establish the key tetrasubstituted stereocenter as previously described (Scheme 34). 142 However, as there is no additional carbon atom on the phosphorane 349, the previous tedious sequence to remove this very atom on ring D is no longer required (341 / (À)-88), seven steps, 6.3%e18.3% OY, Scheme 33). The carbene precursor 350 was derived from N-Boc prolinal (S)-348, obtained from L-proline (S)-140 in three steps and 94% yield. 143 The resulting spirocyclic product 352 of carbene insertion was converted by FriedeleCrafts cyclization into Mori's intermediate 85 363 (R ¼ Me) or (À)-88 (R ¼ -CH 2 -), which were then transformed into CET (À)-(1) following Mori's procedure (Scheme 34).

Ketone 350 was prepared by Wittig reaction of aldehyde 348 with phosporane 349. Protection of hydroxymethyl moiety with TBS was suitable for the carbene insertion, thus treatment of 350 with lithium trimethylsilyldiazomethane gave the azaspirocycle 352, which was converted to NBoc spiroamine 355 by successive deprotection, oxidation, and decarbonylation in good yield (63.2%e84.6% from 352). Because racemization occurred at the tetrasubstituted stereocenter on N-deprotection, compound 355 was first subjected to epoxidation. The epoxide 356 was not isolated and immediately treated with Lewis acid to provide 357 as the major product (2:1). Alcohol 357 was transformed into the CD amino alcohol 360 via tosylation, elimination with DBU to deliver the intermediate cyclopentene 

Three essential steps were used by Djaballah and Gin et al. for their synthesis of CET (À)-(1) (Scheme 36). 144 First, the nitrogen atom was introduced from the oxime 374 followed by Neber rearrangement. Secondly, construction of the benzazepine core relied on a strain-releasing rearrangement of Nvinyl-2-aryl aziridine 376 to 378. Eventually, assembly of the spirofused pyrrolidine core occurred via 1,3-dipolar cycloaddition of azomethine ylide derived from vinylogous amide 379. The chirality was introduced from the optically pure cyclopentenone 372 substituted on C1 and C2 (CET (1) numbering), prefiguring the D ring. Compound 372 was prepared from D-ribofuranose (365). The isopropylidene group allowed stereocontrol, so that only one diastereomer 381 was obtained. Synthesis began by double protection of D-ribofuranose (365), then C1 olefination followed by ParikheDoering oxidation furnished ketone 368, which was alkylated with with vinylmagnesium bromide to provide the allylic alcohol 369 (8:1 dr) (Scheme 35).

Cyclopentene 370 was prepared by an RCM (Grubbs II). Regioselective chloroselenylation of the alkene followed by selenide oxidation and elimination afforded the chlorocyclopentene and deprotection gave the vicinal diol 371. Eventually, the chiral b-chloroenone 372 [10 steps, 42.2% OY from D-ribofuranose (365)] was obtained by periodate-mediated oxidative cleavage of the diol 371.

After the synthesis of 372 featuring the D ring of CET (1) (Scheme 35), the aziridine 375 was prepared in a few steps (Scheme 36). Condensation of ketone 373 with hydroxylamine chlorhydrate furnished oxime 374, which was reacted with LiAlH 4 and diisopropylamine at 60 C affording the racemic aziridine 375 through a Neber-type reductive rearrangement. An additioneelimination sequence starting with the addition of aziridine 375 onto b-chloro-enone 372 furnished the N-vinyl aziridine 376 as a 1:1 diastereomeric mixture (Scheme 36). The dihydro-3-benzazepine 378 was prepared by heating the aziridine 376 with Cs 2 CO 3 , which generated an efficient rearrangement via opening of the aziridine. Compound 378 was N-alkylated with TMSCH 2 I to give the tertiary vinylogous amide 379. O-Activation of vinylogous amide group in 379 was carried out by pivaloyl triflate (generated in situ by combination of pivaloyl chloride and AgOTf), followed by desilylation with TBAT to give azomethine ylide 380. The dipolar cycloaddition with phenyl vinyl sulfone regioselectively afforded the spirofused pyrrolidine 381. In unexpected ways, only one diastereoisomer 381 was obtained; the C1eC2 isopropylideneketal was effective as a stereodirecting element. The desired R configuration at C5 was determined by single crystal X-ray diffraction analysis. Radical desulfurization followed by hydrolysis of pivalate and reacylation with benzyl chloroformate led to 382. After diol deprotection, differentiation of the C1 and C2 oxygen substituents was achieved by using Boc 2 O and Yb(OTf) 3 leading to a selective C1-O-carboxylation. The hydroxy group at C2 was oxidized to ketone using iodoxybenzoic acid (IBX). The with NaBH 4 (95%) concluded the synthesis of (À)-CET (1) in 25 steps from D-ribofuranose (365) and 1.0% OY (ee n.r.).

In 2008, Ishibashi et al. reported an asymmetric synthesis 145 based on a previous construction of the CET (1) skeleton using as characteristic feature an elegant radical cascade for the simultaneous construction of the B and C rings. 146 Diethyl D-(À)-tartrate (384) was transformed into acetonide diol 385 whose tosylation and reaction with lithium phenylacetylide gave compound 387 (Scheme 37). After bromination, the acetonide group was removed and the intermediate bromodiol 388 was subjected to Bu 3 SnHmediated radical cyclization to afford the cyclopentane diol 389 in 47% yield from 384 (eight steps). The two hydroxyl groups were protected and subsequent ozonolysis afforded the desired chiral cyclopentanone 390, which was condensed with amine 392 in the presence of Ti(Oi-Pr) 4 to give an intermediary imine acylated with acryloyl chloride to afford enamide 393 in 50% yield (Scheme 37). Its treatment with Bu 3 SnH in the presence of 1,1 0 -azobiscyclohexanecarbonitrile (ACN) generated a radical cascade reaction building simultaneously the two rings B and C by a 7-endo and 5-endo selective cyclization, delivering the desired ABCD compound 394 in 27% yield. Deprotection of silylated hydroxyl groups by tetrabutylammonium fluoride gave diol 395, which on treatment with trifluoroacetic acid anhydride, dimethyl sulfoxide, and Et 3 N treatment afforded compound 396. Methylation of 394 and ultimate reduction using aluminum hydride 96,132 led to CET (À)-(1) in 99.6% ee, ending the total asymmetric synthesis in 1.3% OY and 17 steps from diethyl D-tartrate (À)-(384).

In 1997, Nagasaka et al. proposed the synthesis of CET (À)-(1) and ent-CET (þ) (1) via resolution of chiral actetals of the spiroketone (AE)-403 providing the C and D rings (Scheme 38). 133 This compound was used in the known procedure developed by Kuehne. 94 Alkylation of the Stork enamine of cyclopentanone by methyl acrylate gave the cyclopentanone ester 399. After formation of the vinyl acetate 401, a nitro group (nitrogen atom precursor) was introduced using ammonium nitrate and trifluoroacetic anhydride to yield to the nitro compound 402. Reduction of the nitro group and in situ lactamization allowed formation of the spirolactam (AE)-403.

Acetalization of (AE)-403 with (R,R)-2,3-butanediol 404 afforded two diastereoisomers (À)-405 and (þ)-406, which were easily separated by HPLC (100% combined). The optical purity was checked by chiral HPLC on compounds (À)-403 and (þ)-403 obtained by hydrolysis of their respective acetals. Each enantiomer was confirmed to be enantiomerically pure. The two diastereoisomers (À)-405 and (þ)-406 were alkylated with sulfonic ester 305 leading to Kuehne's intermediates (À)-306 and (þ)-306. Synthesis of CET (À)-(1) and ent-CET (þ) (1) in 99% estimated 

Cephalotaxus Alkaloids ee could be achieved employing Kuehne's procedure 94 in 16 steps from cyclopentanone (397) and 5.8% calculated OY for each enantiomer.

In 2002, El Bialy et al. 140 Recrystallization of the ammonium malate of spiro compound (AE)-408 gave the enantiomer (À)-408 which could be converted to unnatural ent-CET (þ)-(1). The mother liquor of malate salt afforded Ikeda's intermediate (þ)-408 130 (yield not reported), which was protected and oxidized to the known ketone (þ)-324 and then transformed according to their sequence 137 HPLC. Applying Hanaoka's protocol would allow the synthesis of CET (À)-(1) in 22 steps, <7.5% OY and 89.3% estimated ee from piperonal (82).

Stoltz et al. developed a formal asymmetric synthesis of the two enantiomers of CET (1) and the first total synthesis of (À)-dupracine (16) using the spirocyclic amine (AE)-110 first used in their synthesis of rac-CET (AE)-(1) (Scheme 5) to introduce CD rings. 90 Compound (AE)-110 was coupled with the chiral nonracemic hemiacetal 414 obtained by chemical resolution (Scheme 40). 90 3,4-Methylenedioxymandelic acid (R)-(413) (97.5% ee) obtained by chemical resolution with ephedrine (412) was protected as a 1,3-dioxolan-4-one prior to bromation then reduction to deliver compound 414 as a 1/1 mixture of diastereomers. This mixture reacted with spirocyclic amine (AE)-110 (see Scheme 5) under reductive amination conditions to give diastereomers 415 and 416 in a 1:1 ratio, which were smoothly separated by chromatography. Both diastereomers were submitted separately to Heck reaction conditions to produce benzazepines (þ)-417 and (À)-418. The pyrrolo-fused pyridinium salt 421 was irradiated to give the tricyclic aziridine 422, its regioselective ring opening with acetic acid, amine protection with Boc, then protection of alcohol as acetic ester, at last deprotection of the Boc group with TFA led to meso spirocyclic aminodiester 425 in five steps and 38% OY. Desymmetrization by enzymatic hydrolysis with electric eel acetyl cholinesterase afforded the unstable aminomonoalcohol 426 in the range of 80%e90% ee, which reacted with (3,4-dimethoxyphenyl)acetyl chloride (144) to furnish amidoalcohol 427. To remove hydroxy group at C1, compound 427 was transformed into oxalate 428, which underwent radical reduction by treatment with azobisisobutyronitrile (AIBN)/n-Bu 3 SnH generating the desired product 429 in a 10% yield along with the isomerized homoallylic alcohol 431 as the major product (65%). Mori's intermediate 430 was obtained by reduction of 429 with LiAlH 4 . It was also obtained through a five-step sequence (55% OY) from the rearranged homoallylic alcohol 431 involving alcohol deprotection and oxidation, a,b-unsaturated enone formation, and then ketone and amide reduction. This formal synthesis could allow the synthesis of CET (À)-(1) in 19/23 steps from 419 via ACD intermediate 430 with 0.2% expected OY (0.5% OY brsm) and 51% estimated ee. This medium ee indicates that some epimerization occurred during transformation of spirocyclic compound 426 into ACD unit 430.

The second approach by Mariano et al. described the synthesis of the ACD unit (À)-102, which is the key intermediate described by Suga and Yoshida 86,87 (Scheme 42). The azaspiroalcohol 433 was prepared earlier 128 in enantiomerically pure form by using a sequential photocyclization/aziridine ring opening/enzymatic desymmetrization sequence with >95% ee in four steps and 19% OY from pyridinium salt 421. Radical reduction with tributyltin hydride and AIBN of the mixed oxalate gave the homoallylic ester 434.

Hydrogenation of 434 with 10% Pd/C followed by hydrolysis with HCl generated the spirocyclic alkanolamine 435. Reaction with iodoarylacetyl chloride 432 furnished the spirocyclic amidoalcohol 436, which was transformed into the tosylate 437. Tosylate elimination with DBU afforded the unsaturated amide 438, which was reduced with aluminum trihydride to give the Suga and Yoshida intermediate (À)-102. This formal synthesis would give CET (À)-(1) in 95% expected ee over 19 steps in 0.5% calculated OY from alcohol 419.

In 2012, another formal synthesis of CET (À)-(1) was reported by Renaud. 124 Unlike the alternative approach (see Section 3.3.1.3, Scheme 27), an access to azaspiroamide 404 (Nagasaka's intermediate) 133 was achieved through a stereoselective carboazidation sequence as the key step (Scheme 43).

In this synthesis, the silylated derivative (S)-439 (96.5% ee) was prepared from racemic cyclopentanol 148 To conclude this part, it is interesting to compare the target intermediates in the formal syntheses of CET (1) as subtle changes in their structure may induce very different reactivities and OY of transformation into CET, either in its racemic form (AE)-(1) (Fig. 18 ) or optically active ones, the natural enantiomer (À)-(1) and the distomer (þ)-(1) (Fig. 19 ). In addition, this schematic view shows in some ways the interconnections between the different approaches.

Side chains of Cephalotaxus alkaloids are very important for the biological activity of the corresponding esters since it is well known that CET itself does not exhibit interesting activity. They are characterized by a relatively simple structure with a chiral tertiary alcohol moiety. Nevertheless, coupling of the side chain to CET remains an important challenge due to the steric Figure 18 Target intermediates in the formal syntheses of (AE)-cephalotaxine (1) . Names in boxes refer to a formal synthesis. Figure 19 Target intermediates in the formal asymmetric syntheses of cephalotaxine (1) . Names in boxes refer to a formal synthesis. side chain. It thus appeared that syntheses of the acidic side chains themselves are not useful processes to attain the Cephalotaxus esters. Nevertheless, numerous interesting syntheses of the acidic side chains have been reported. We will first report on the synthesis of the CET esters and then on the synthesis of the side chains. In each part, syntheses will be presented in chronologic order.

The previous review in this series 2 reported more than 15 syntheses of CET's esters. Due to the impossible coupling with the entire side chain 444, most of these syntheses dealt with the esterification of CET (1) with a precursor a-keto acid 443 as shown in Scheme 44. Formation of the aketoester 445 was followed by the introduction of the acetyl group through a Reformatsky reaction. Although this esterification/Reformatsky alkylation sequence is interesting, proceeded well and allowed the preparation of various HTs such as HT (3), HHT (2), deoxyHT (5) , and so on, the sequence also exhibits a major drawback: the Reformatsky reaction on the prochiral ketoester part of 445 continuously gave a mixture of two epimers in a 1/1 mixture leading thus to the loss of half part of the CET. This major inconvenience was avoided in only two syntheses of HHT (2): the synthesis reported by Hudlicky 149 (where the presence of the second keto group allowed a chelation with the Reformatsky reagent leading to the formation of a single stereomer) and the synthesis of Cheng 150 (using a chiral sulfinyl ester).

Since the previous review in this series, 2 only four articles dealing with syntheses of CET esters have been published.

The first one by Robin 21 in 1999 is quite interesting in opening new routes to these alkaloids. This strategy was based on the esterification of CET (1) with a cyclic precursor of the side chain, the tetrahydropyran rac-449 (Scheme 45), which was easily prepared in four steps from a-ketoester 446.

Ketoester 446 was treated with the lithium methyl acetate to give diester 447, which was fully saponified and then cyclized with formic acid to the diacid 448. The latter was selectively monoesterified with BF 3 $MeOH to furnish the acid ester 449. CET (1) was coupled with rac-449 using dicyclohexylcarbodiimide or mixed anhydride method, but once again the obtained ester was a mixture of two epimers in a 6/4 ratio and had to be separated by chromatography. While the authors showed that highly enantiomerically pure HHT (2) (99.8% ee) could be obtained in a large scale process, they also developed a diastereoselective sequence. For this purpose, rac-449 was resolved through its quinine ester. Eventually, coupling of CET (À)-(1) and (À)-449 gave anhydroHHT (450), which can also be hydrolyzed to enantiopure HHT (À)- (2) . This methodology using tetrahydropyran 449 was patented in 2005 and applied to synthetize various analogs. 1)) according to an efficient semisynthesis described in 1999 using enantiopure Robin's acid 449. 21 The synthetic strategy to obtain [ 14 C]-CET (À)- (1) Cephalotaxus Alkaloids alcohol obtained from piperonal (82) , the 17-steps sequence provided after preparative reverse phase HPLC a 100% enantiopure product with a radiochemical purity of 98.9% and a chemical purity of 98.5% in 4.475% OY.

The synthesis of deoxyHT (À)-(5) reported by Gin in 2006 153 was the most recent and also the most efficient method of synthesis of CET esters. This synthesis was based on the esterification of a cyclic acid, namely the b-lactone 456 obtained in a nonracemic form through an efficient sequence (Scheme 47). It started with acetal derivatization of D-malic acid D-(409) with pivalaldehyde to give dioxolanone 453. Diastereoselective alkylation of dioxolanone 453 with prenyl bromide was accomplished after deprotonation with LiHMDS to give the prenylated dioxolanone 454 in 66% yield. This sequence allowed the introduction of the chiral tertiary alcohol with complete diastereoselectivity following the studies of D. Seebach 126 and Tietze (see Scheme 51) . 154 Transesterification with acetal cleavage of 454 using benzyl alcoholate provided the tertiary alcohol 455, which was cyclized to the b-lactone followed by hydrogenolysis and hydrogenation to give the acid 456. Activation of the carboxylic function with 2,4,6-trichlorobenzoyl chloride allowed the acylation of CET (À)-(1) followed by transformation into deoxyHT (À)-(5) on methanolysis. 

Homoharringtonine, and Homodeoxyharringtonine (2008) Djaballah, Gin, and coworkers extended this study to the syntheses of others CET esters such as HHT (À)-(2), homodeoxyHT (À)- (22) , and anhydroHT (À)- (20) . 144 Preparation of anhydroHT (À)-(20) (Scheme 48) involved the intermediate (R)-455 of the previous synthesis 153 (see Scheme 47) , which, after transformation into the corresponding methyl ester, was cyclized to tetrahydrofuran 457 through alkoxymercuration. Deprotection of the benzyl ester and activation as a mixed anhydride (Yamaguchi's reagent) allowed the coupling with CET (À)-(1) to give anhydroHT (À)- (20) in excellent yield.

Preparation of HHT (À)-(2) and homodeoxyHT (À)-(22) (Scheme 49) involved a synthesis based on b-lactone 458 obtained through a sequence, which mimicked the synthesis described in Scheme 47. A cross-metathesis provided the disubstituted alkene 460 in 61% yield (a dimeric bislactone was also formed). After cleavage of the benzyl ester, the activation of the carboxyl group (Yamaguchi's reagent) allowed the coupling with CET (À)-(1) to furnish 461, which was transformed into 462 on methanolysis. Preparation of each natural compound HHT (À)-(2) and homodeoxyHT (À)- (22) from this common intermediate is interesting. When allyl benzyl ether 462 was treated with H 2 (Pd/C) in MeOH followed by addition of AcOH at the last stage of the reaction, HHT (À)-(2) was isolated in 79% yield. On the other hand, 462 gave homodeoxyHT (À)- (22) in 69% yield when submitted to Pd/C catalyzed hydrogenation in glacial AcOH.

Though it has been clearly shown that coupling of CET (1) with entire side chains was not possible, syntheses of these side chains have been continuously prolific. All the syntheses reported since 1998 were asymmetric and are of interest due to their diversity and originality. Some of them included the preparation of already described intermediates in the synthesis of Cephalotaxus esters and then could be considered as formal syntheses of the natural products.

The first asymmetric synthesis of the side chain ester 473 of HHT (2) was reported by Dumas and d'Angelo. 155, 156 In this work, the formation of the chiral tertiary alcohol center was accomplished in a highly diastereoselective manner in an early stage of the sequence. The synthesis started with the preparation of the known chiral imine 465 (Scheme 50), which was subjected to a Michael reaction with acetoxyacrylonitrile (466), an equivalent of the acid side chain. The adduct 467 was proved to be stereochemically homogeneous and was hydrolyzed without purification to give the more stable ketone (R,R)-468. The keto group was protected as a dioxolane and the potential aldehyde function was directly transformed into the ester 469 on treatment with MnO 2 in MeOH in the presence of NaCN. The ketal of ester 469 was cleaved using CeCl 3 $7H 2 O and NaI, and the so-formed keto group was converted to the silylvinyl ether 470. Ozonolysis of 470 gave the sensitive aldehyde 471. Addition of MeMgBr to the crude aldehyde and esterification of the acid furnished alcohol 472. Oxidation of the alcohol 472 to a methylketone followed by new addition of MeMgBr gave the tertiary alcohol, which eventually on hydrogenolysis led to (R)-methyl 3,7-dihydroxy-3-methylcarbonyl-7-methoxyoctanoate (473), the methyl derivative of HHT side chain.

In 2005, Tietze reported the preparation of the chiral side chain acids 480e 483 of deoxyHT (5), homodeoxyHT (22) , nordeoxyHT (21) , and neoHT (19), respectively (Scheme 51) 154 based on Seebach's procedure for the alkylation of D-malic acid with self-regeneration of stereogenic centers. 126 The strategy was used later by Gin for the preparation of CET esters. 153 The dioxolanone 453 prepared from D-malic acid (D-409) was treated with two equivatents of LiHMDS followed by bromoallyl derivatives to give allylated dioxolanones 474e476 in high yield and diastereoselectivity. On hydrogenation on PtO 2 , these compounds furnished dioxolanones 477e479 whose acid hydrolysis permitted the isolation of 480e482, side chains of deoxyHT (5), homodeoxyHT (22) , and nordeoxyHT (21), respectively. When 453 was alkylated with benzyl bromide, the same sequence allowed the obtention of 483, the side chain of neoHT (19) .

An original approach was reported by Russel et al. in 2006. 157 Three compounds were described in this study: (R)-493 and (R)-496, the anhydroHT and HHT's side chains respectively, and (R)-449, the intermediate used by Robin for the synthesis of HHT (2) (see Scheme 46) thus representing a formal synthesis of HHT (2) . Dioxolanone (R)-487 was obtained in three steps and resolution in 11.4% OY from citric acid (484). 158 The acid function of (R)-487 was selectively reduced under 

Several Cephalotaxus alkaloid side chains were prepared by Royer through a rapid synthesis (fourefive steps). 159 hand, it was postulated that this epoxide could be a good candidate for esterification with CET (1) while no report of such coupling was available. The diester 501, easily obtained from commercially available monomethyl itaconate (499) and methyl mandelate (500), was epoxydized with mCPBA to give a 1/1 diastereomeric mixture of 502 in good yield. Flash chromatography permitted the isolation of (þ)-502 with the required (S,S) absolute configuration. Then, epoxide (þ)-502 was reacted with various organocuprates to give compounds 503e507. Hydrogenolysis of 503e507 allowed the obtention of (R)-508, (R)-496, (R)-509, side chains of deoxyHT (5), homodeoxyHT (22) , and neoHT (19) , respectively and analog (R)-510 in 12.2%e17.8% OY. Hydroxymercuration of 507 followed by spontaneous cyclization and then hydrogenolysis of the chiral appendage delivered (R)-494, the side chain of anhydroHT (20) in 10.8% OY.

In 2013, Yang reported the [2,3]-Meisenheimer rearrangement as a general strategy for the construction of chiral tertiary alcohols. 160 As an illustration of this method, the synthesis of HT and HHT's side chains is shown in Scheme 54. The synthesis began with known N-Boc ester (S)-512 of L-aspartic acid obtained through a two-step process (94%). The gem-dimethyl group was first installed by a Grignard reaction to give 513, which was transformed into the lactol 514 by IBX oxidation and concomitant cyclization. The Wittig reaction with phosphorane 515 followed by deprotection and N-bisbenzylation of the amino group furnished key allylamine (S)-516. The Meisenheimer rearrangement occurred on treatment with mCPBA on the latter to give hydroxylamine 517. This compound led to 518 on treatment with H 2 on catalytic Pd/C via double bond saturation and then NeO bond cleavage. Eventually, on oxidation of 518 the HHT's side chain acid (R)-481 was obtained whose methylation yielded the corresponding dimethyl ester (R)-473. Its optical rotation indicated a perfect stereochemical outcome of the rearrangement. A similar process allowed synthesis of a cyclic form of the HT's side chain. Starting from L-threonine (519), amino alkene 520 was prepared in seven steps. The Meisenheimer rearrangement occurred on treating 520 with mCPBA to give the sensitive hydroxylamine 521. Six steps were then needed to obtain (R)-522, a cyclic form of the HT's side chain acid. These authors patented the preparation of HHT's side chain acid (R)-481 by this methodology. 161 Through the same sequence, Yang and coworkers have also described the synthesis of the diacid side chain of deoxyHT (5) and homodeoxyHT (22) starting from L-valine and L-leucine respectively. 162 

Hung et al. 163 recently reported the synthesis of the esters (473) and (524) of the side chains of HHT (2) and homodeoxyHT (22) , respectively, by using Tietze 154 and Gin's methods 144 (Scheme 55). Allylation of 453 (see Scheme 51), esterification, and then cross-metathesis (see Scheme 49) with allyl alcohol 524 allowed the preparation of dioxolanone 523. The latter was subjected to different hydrogenation conditions to give after methanolysis (R)-473 or (R)-526, side chains' esters of HHT (2) and homo-deoxyHT (22).

In 2016, Tietze's method 154 (see Scheme 51) was also used by Mac et al. for the synthesis of a series of side chains analogs of HHT (2) (Scheme 56). 164 The dioxolanone 453 was diastereoselectively allylated to furnish compound (R,R)-523. A MizorokieHeck type reaction with various phenylboronic acid derivatives followed by hydrogenation and methanolysis gave arylated ester chain analogs 527aeh of HHT (2) in moderate to good yield.

The synthesis of analogs of bioactive natural products remains a broad area of research that can lead to very interesting new structures with improved activity. In the case of Cephalotaxus alkaloids, the exchange of the ester side chain appears to be the simplest way to access analogs and this strategy has been used and described in some patents. The structural modifications of the CET core seem more difficult but have been also scarcely addressed giving rise to interesting compounds and synthetic pathways. We will describe below the most significant reports. Works describing (R,R)-525

Scheme 56 Synthesis of homoharringtonine side chain's analogs by Mac (2016).

Cephalotaxus Alkaloids attempts to synthesize CET (1) or the preparation of compounds with a passing structural resemblance will not be considered in this section.

Since 1997, several reports on the synthesis of CET's analogs have appeared. Tietze et al. in 2000 165 described the preparation of four analogs rac-532e 535. For these compounds, variation of the size of the different rings B and C of the polycyclic core of CET (1) and ring A substitution was explored. These authors used their strategy developed for the synthesis of the natural product (Scheme 57).

Compounds 530aeh, with various chain lengths and ring sizes, were easily prepared in a few steps by alkylation of 84 and 528aec with 81 and 529aec using a standard S N reaction. Compounds 530aeh were then submitted to a Pd-catalyzed TsujieTrost cyclization to give 531aec,e,f,h Scheme 57 Synthesis of cephalotaxine analogs by Tietze (2000) .

in satisfactory yields (compounds 530d,g did not cyclize under these conditions). The final Heck cyclization required the use of HermanneBeller's palladacycle 87 and allowed the preparation of compounds 532e535 and 88. In this study, the final compounds required further reactions to install at least an alcoholic function on ring D to allow the coupling with side chains. This drawback was eliminated in a following study published in 2007 by the same authors. 166 Furthermore, Cephalotaxus esters analogs were prepared by coupling the CET analogs with homodeoxyHT side chain (Scheme 58). The synthesis was now based on a domino reaction giving a rapid and efficient access to ACD compounds 538 from amines 536, 528a, 84 and cyclopentenone ester 537.

The Pd-catalyzed reaction of 538aec allowed the formation of the tetracycle 539aec. While LiAlH 4 reduction of 539aec gave efficient access to amino alcohols 540aec, which are analogs of CET (1), these compounds did not couple with an activated form of the side chain of homodeoxyHT (22) . In contrast, it was observed that NaBH 4 reduction Scheme 58 Synthesis of deoxyharringtonine analogs by Tietze (2007) . of 539b,c gave the amidoalcohols rac-541b,c, which allowed coupling with the imidazolide 542 of homodeoxyHT side chain furnishing after further reduction the esters 543b,c, which are analogs of homodeoxyHT (22) . It should be noted that the presented synthesis is racemic and that optically pure side chains were used leading to a mixture of two diastereomers, which were separated.

In 2005, Bubnov et al. 167 published the synthesis of compound rac-548 exhibiting some structural analogies to CET (1) with ring B and D being bridged and not fused (Scheme 59). The synthetic strategy was simple and started with 2-pyrrolidone (181), which was diallylated, N N-protected, and transformed into spiro-pyrrolidine 545 through ring closure metathesis. Condensation of 545 with acyl chloride 546 gave access to 547, which was transformed to the desired product rac-548 by Heck cyclization. The same type of work was pursued by the authors in 2010 168 allowing the preparation of new products 549e552. protected lactam 554 through deprotonation using sec-BuLi and addition of MeI, BuI, or PhCH 2 Br to provide diastereomeric alkylated lactams 555 and 556 in a 4/1 ratio. The new chiral center of the major isomer has the (S)-configuration. Transformation of acetonide 555 led to alkylated CETs 557e559 in a four-step sequence in 15e28% OY.

In 2010, these authors reported a different type of CET analogs in which the A ring was changed for a phenyl or a naphthyl group (Scheme 61) using the same strategy as for CET's synthesis (see Scheme 34) . 169 N-alkylation of the spirolactam (L)-304 with the appropriate arylethyl derivative 560a,b gave 561a,b. The end of the synthesis paralleled those of the CET synthesis: functional group manipulations of 561a,b led to 562a,b. These spiro compounds 562a,b were cyclized to the ABCD compounds 563a,b, which were eventually transformed to the phenyl 564 and naphthyl 565 CET analogs.

In 2016, Chandrasekhar et al. used their aryne insertion methodology to synthetize three CET analog intermediates (Scheme 62). 120 In these compounds 570e572, the A ring is changed for a phenyl, dimethylphenyl, or dimethoxyphenyl ring. To conclude, it is noteworthy that if several analogs of CET (1) have been designed and synthesized, they have been coupled with side chains for only two products (in two reports), although their biological activity was not reported. treated with sodium hydride to yield synthetic CET ester 575a in 30% yield. This compound was the most active of the tested analogs in this study with nanomolar IC 50 (3.5 ng/mL) while HHT (2) was less potent (IC 50 14 ng/mL). 

Analogs 576 of HT (3), 579 of drupangtonine (33) , and 577 and 578 of HHT (2) were also synthetized by conventional methods (Fig. 20) .

This work delineated the ester groups that reduced the cytotoxic activity against the K562 leukemic cell line (Table 11) . Nevertheless, a few ones gave rise to improved activity as esters 575a (n-hexadienyl), 575b (n-hexyl), 575s (n-hepthyl), 575v (4-methylbenzyl), 575w (3-methylbenzyl), 575af (2-methylthiophene), 575ag (2-methylfurane).

In 2008, Djaballah and Gin synthesized four CET-ester analogs, benzyldehydroHHT 580, bisnorhomodeoxyHT 581, bisnorhomodeoxyHT b-lactone 582, and deoxyHT b-lactone 583 (Fig. 21 ), using the protocol described in Section 4.1.3 (Scheme 47). 144 They exemplified up to 43 Cephalotaxus esters analogs, some of which being dimers of CET (1) linked by a central oxygenated chain by the C3 hydroxyl group. 172 These HT structural analogs were tested against human hematopoietic and solid tumor cell lines (Tables 12 and 13) , and compared to HHT (2), deoxyHT (5), homodeoxyHT (22) , and anhydroHT (20) , which are among the most active Cephalotaxus alkaloid esters.

HHT (2) shows decreased activity by a factor of 125 (resistance index (RI) ¼ 125), whereas esters deoxyHT 5, homodeoxyHT (22), benzyldehy-droHHT (580), and bisnordeoxyHT (581) have much lower RI of 11, 3, 19, and 12 respectively, indicating that they are significantly less sensitive to multidrug resistance (MDR) and tumor cell lines can therefore be Cephalotaxus Alkaloids considered to be sensitive to these products. One possible explanation for the greater susceptibility of HHT (2) to MDR is its low lipophilicity due to the structure of the side chain, making it a good substrate for the efflux pumps. Compounds having a ClogP value greater than 1.2 generally lead to a low susceptibility to MDR (i.e., IR ¼ 19 to the esters 5, 22, 580, and 581). The exception is HHT (2), which has a relatively low value of cLogP (0.95, relatively polar) reflecting its susceptibility to MDR (i.e., IR ¼ 125). All these compounds show a broad spectrum of cytotoxic activity, with the exception of the retinoblastoma cell line Y79 (Table 13) , which can be attributed to the overexpression of the MDR gene in this line.

In 2013, Lai et al. applied for two patents, one on preparation of aminated HHT derivatives 584 173 (Fig. 20) and the other on acylated HHT derivatives 585 and 586 (Fig. 22) . 174 Compounds 584a-w were prepared from HHT (2) via saponification of the methyl ester to HHT acid (31) and amidation with various amines R 1 R 2 NH in the presence of HATU/DIPEA in 6%e44% yield (Fig. 22 ). Groups R 1 and R 2 could be H, C 1e18 alkyl, C 2e18 alkenyl, C 2e18 alkynyl, Cephalotaxus Alkaloids C 3e7 cycloalkyl or cycloalkenyl, aryl, heteroaryl, or saturated heterocyclic ring systems, etc. Twenty-three HHT aza analogs were tested against human hematopoietic and solid tumor cell lines (Table 14) , and compared to HHT (2) . Analogs 584c, f, n, o, and u were inactive (IC 50 ! 16 mg/mL). Compound 584b and 584i were the most active of the tested analogs in this study with nanomolar IC 50 (up to 42 and 57 ng/mL) while HHT (2) was more potent (IC 50 6 ng/mL). The most promising analogs were tested against HeLa (cervical), CaES-17 (esophageal), CNE (nasopharyngeal), Hep2 (liver), MGC 803 (gastric), PC-3 (prostate), and SK-OV-3 (ovary) human cancer cell lines (Tables 15 and 16) , and however, none of these compounds were more active than HHT (2). Monoacyl-HHT analogs 585aei and diacyl-HHT analogs 586aef were prepared with activated HHT and carboxylic acids, anhydrides, or acyl chlorides via esterification in 6%e99% yield (Fig. 23) .

The acylated analogs were compared to HHT (2) and tested against human hematopoietic (Table 17 ) and solid tumor cell lines (Tables 18  and 19 ) including K562/adr (myelogenous leukemia resistant to DOX), Kasumi-1 (acute myeloid leukemia (AML)), NB4 (acute promyelocytic leukemia (APL)), Jurkat (childhood T acute lymphoblastic leukemia (ALL)), H9 (cutaneous T-cell lymphoma), RPMI8226 (plasma cell myeloma), Huh7 (hepatoma), A549 (lung adenocarcinoma), PANC-1 (exocrine pancreas carcinoma), Becap37 (breast), MG63 (osteosarcoma), RKO (colon), U87MG (glioma), HeLa (cervical), CaES-17 (esophageal), CNE (nasopharyngeal), Hep2 (liver), PC-3 (prostate), SK-OV-3 (ovary), HHT (2) (Tables 17, 18 and 19) . Compounds 585f and 585g were more potent than HHT (2) against 18 and 17 over the 20 cell lines tested, respectively.

In conclusion, the study of many syntheses of CET (1) published to date can draw a number of conclusions about the key points of its structure and the way chemists have overcome the problems posed by its synthesis. The formation of the B ring often remains a crucial point in building the CET skeleton. The C4eC13 disconnection remains by far the most studied (>30 syntheses of 50). The cyclization reactions used (electrophilic aromatic substitutions, organometallic coupling, etc.) perform generally with good yields but seem susceptible to cyclization intermediates and electronic richness of the aromatic ring. This disconnection is also extremely interesting from a stereochemical point of view providing the correct relative configuration at C4 and C5, avoiding difficult separations. Other disconnections do not always give such good results and are not completely diastereoselective. The formation of the spiro CD system is also a key point in these syntheses; in fact, the control of center C5 at the start of synthesis greatly simplified stereocontrolled strategies and enabled enantioselective approach to CET (1) by asymmetric synthesis of the CD system. Subsequent stereoinduction of the C5 center on C4 center and C4 center on the C3 center facilitate the production of optically active CET (1). However, the operating conditions used after the synthesis of the spiro system must be chosen carefully because of the risk of racemization of CET (1) and certain synthetic intermediates by breaking the N9eC5 bond (See Figs. 18 and 19 , the conversion of Mori's intermediates (AE)-88 and (À)-88). Last but not least, from an industrial point of view, initially the Merck Group took over the Weinreb racemic synthesis improving performance (11.7% by eliminating some purification steps). The optically active CET (1) in its natural configuration may be obtained by resolution of racemic cephalotaxine rac-CET (AE)-(1) with tartaric acid, by incorporation of chiral sources, or by enantioselective synthesis. The high-purity HHT (2) used for clinical trials was obtained by extraction (NCI batches) and by semisynthesis (ssHHT) from natural CET (1) then purified by HPLC to remove impurities and minor diastereoisomers first by the French company Oncopharm SA later based in Houston (USA). A number of HHT (2) analogs were prepared; however only the monoacylated analogs 585a, 585f, and 585g showed slight increase in activity agains both hematopoietic and solid caner cell lines. The main advantage was to prepare hydrosoluble stable forms. Figure 24 Simplified diagram of protein synthesis showing the A site where homoharringtonine (HHT) binds to inhibit protein synthesis. 178 In the cytoplasm of the cells, ribosomes (green) bind to the mRNA containing the code translated from the switched on genes into the DNA. The 20 different amino acids carried by transfer RNA (tRNA) are recruited according to a three nuclear base code into the ribosome A site. The protein chain elongation includes the aminoacyl tRNA entry, its proofreading, the peptidyl transfer, and the ribosomal translocation along the tRNA chain. The cartoon on the right explains how HHT (2) inhibits protein synthesis through inhibition of initiation elongation step. the structure and function of the ribosome." They have all generated 3D models that have shown how different antibiotics, including HHT (2), 176 bind to the ribosome. Also, they demonstrated by X-ray (at high resolution up to 2.9 Å) the crystals of the 80S ribosome in complexes with 12 eukaryote-specific inhibitors and four broad-spectrum inhibitors in Saccharomyces cerevisiae. 177 Both the ring system and the tail of HHT (2) participate in interactions with the ribosome that contribute to drug binding. HHT (2) interaction with the ribosome is stabilized by hydrophobic interactions and hydrogen bonds (Fig. 25) . Its six-membered aromatic ring stacks on the base of Cytosine 2452 for Haloarcula marismortui (2487 for Escherichia coli), and it fills the A-site cleft completely, the five-membered dioxolane ring attached to it. When HHT (2) is bound to the ribosome, 73% of its surface area becomes water inaccessible. The tertiary amine of HHT (2) 176 One of the main actions of HHT (2) is induction of apoptosis (Fig. 25 ). HHT (2) has a different mechanism of action from that of tyrosine kinase inhibitors, a class of drugs used in for the treatment of CML. Accordingly its activity in CML is independent of the mutation state of BCR-ABL cells (particular CML cells that contain a fusion gene called BCR-ABL1, generated by translocation of sequences from the c-ABLgene on chromosome 9 into the BCR gene on chromosome 22). 179 The mechanism of apoptosis induced by HHT (2) A study was conducted on SHI-1 cell lines to prove that mitochondria are involved in the apoptosis induced by HHT (2) . 180 Although HHT (2) induces apoptosis mediated by mitochondria (intrinsic pathway), it also induces apoptosis through death receptors (extrinsic pathway). The intrinsic pathway is induced by cleaving the proactive forms into active ones. HHT (2) activates some apoptosis mediators; it inhibits the apoptotic inhibitors and upregulates the release of other mediators through the action on genetic level. For instance apoptosis was studied in QGY-7703 cell line and 78 individual mRNAs were identified as having their transcription levels significantly changed. Those genes were mostly oncogenes, tumor suppressors, enzymes, kinases, and partially related to apoptosis, 68% being upregulated and 32% were downregulated. The following signaling apoptosis pathways were activated: TGF-b, tumor necrosis factor (TNF), FAS, p38MAPK, and p53. 181 The head protein of intrinsic apoptosis cascade is p53. It was activated by HHT (2) in human hepatoma QGY-7703 cells 181 , but cell line with p53 mutation was more resistant to HHT (2) . 182 A study demonstrated rapid apoptosis of U937 and HL60 cell lines on treatment with HHT (2); this was measured by increased annexin V binding capacity, caspase-3 activation, and the cleavage of poly(adenosine 5 0 -diphosphate (ADP)ribose) polymerase (PARP). Also, the expression of Bax was upregulated during cell death induced by HHT (2), while the expression of bcl-2 was only slightly decreased. 183 On treatment of human myelodysplastic syndrome (MDS) MUTZ-1 cell lines with HHT (2), Ca 2þ translocated from endoplasmic reticulum (ER) pool to cytosol, the mitochondrial membrane potential (Dwm) decreased, and Bid protein translocated from ER to mitochondria. Ca 2þ and Bid proteins were released from ER pool to cytosol. The activation of ER stress-associated proapoptotic factor, CHOP and ER chaperones BiP and XBP1 genes, was followed by cleavage of caspase-3 but not caspase-4. 184, 185 mRNA expression levels of survivin were investigated in various hematopoietic cell lines in relation with apoptosis induced by HHT (2) . It was concluded that the apoptotic effect of HHT (2) on the hematopoietic cell lines is associated with decreased level of survivin expression. 186 In a study of apoptosis using ssHHT (2) on HL60 and HL60/MRP (cells lacking Baxes) cell lines, ssHHT (2) was effective in a time-and dose-dependent manner, and independently of the expression of Bax. Additionally, ssHHT (2) induced apoptosis through the decrease of Dwm and the release of cytochrome c. This was explained by decreased myeloid cell leukemia-1 (Mcl-1) level and cleavage of Bcl-2, one of the proteins that governs the outer membrane of the mitochondria and has antiapoptotic activity (Fig. 26 ). 187 HHT (2) inhibited the telomerase content of HL-60 cells effectively and induced apoptosis. 188 HHT ( (2) (2)

(2) (2) TGFβ α Figure 26 Main pathways involved in apoptosis mechanism induced by homoharringtonine (2) . Yellow: apoptosis factor; Blue: apoptosis intermediator; Turquoise: dual role as intermediator and factor; Green: apoptosis inhibitor.

The intrinsic apoptosis using transforming subfamily of growth factorb-inducible Sp1-like proteins called transforming growth factor-beta inducible early protein gene (TIEG1) was studied. TIEG1 induced apoptosis through upregulation of Bax and Bim and downregulation of Bcl-2 and Bcl-XL, release of cytochrome c, and caspase-3 activation. 190 The apoptotic effect of HHT was studied in K562 cell line using gene chip technology, it was found that 17 genes were upregulated and 27 were downregulated. Most of them were found to be related to apoptosis, oncogenes, or tumor suppression. Several genes with altered gene expression were found, such as TIEG, vitamin D3 upregulated protein one gene (VDUP1), RNA binding motif protein 4 gene (RBM4), and v-myc myelocytomatosis viral oncogene homolog (C-MYC). The activated transforming growth factor-b and TNFesignaling pathways play an important role in apoptosis according to the dynamic gene expression pattern in these apoptotic cells. TIEG was significantly modified after induction of apoptosis, which is critical for apoptosis signal transmission. 191 A Chinese team tried to identify different genes whose expression is modified during apoptosis induced by HHT (2) on leukemic HL-60 cell lines; they demonstrated that TGF-a and TNF apoptosis signaling pathways were initiated but not the Fas signaling pathway. Overexpression of TIEG and VDUP1 could induce apoptosis in some kinds of cells and they were upregulated. Other genes, such as BRD2, ACVRL1, and RBM4 were also upregulated and are probably involved in apoptosis signaling pathways. 192 The efficacy of HHT (2) against ponatinib-resistant Ph-positive cells Ba/F3 (Ba/F3 ponatinib-R) was shown by decreased protein expression levels of BCR-ABL and Crk-L. On the other hand, caspase-3 and cleaved PARP levels were significantly increased. HHT (2) treatment reduced the expression of BCR-ABL, as well as heat shock protein (HSP) 90, which stabilizes the BCR-ABL protein. It was found that HHT (2) reduced the expression level of the antiapoptotic protein Bcl-2 and of c-Myc. HHT (2) was examined on ponatinib-resistant primary Ph-positive ALL and CP CML cell lines. The synthesis of BCR-ABL, HSP90, and Bcl-2 was downregulated and was independent of caspase activity. 193 HHT (2) antagonizes specifically the p-eIF4F function, which is specifically required by cancer cell for their survival and not the eIF4F, which is required for normal cells. 194 HHT (2) reduces the levels of prooncogenic or prosurvival proteins (important proteins for tumorigenesis) having short half-lives, such as Mcl-1, cyclin D1, or c-Myc. This was demonstrated on a well-studied transgenic mice model Em-myc lymphomas harboring lesions in Pten, Tsc2, Bcl-2, or eIF4E. 175 Cytotoxicity of HHT (2) and cell-growth activity were investigated in three AML cell lines, HL-60, NB4, and U937, and in three CML cell lines, K562, KU812, and KCL22. AML cells showed more sensitivity than CML cells to HHT-induced cytotoxicity. Using HL-60 cells, it was revealed that HHT (2) decreased the levels of Mcl-1, X-linked inhibitor of apoptosis protein (XIAP), survivin, and B-cell lymphoma 2 (Bcl-2) homology domain 3 (BH3)-only proteins as well as the Dwm. U937, K562, KU812, and KCL22 cells that expressed B-cell lymphoma-extra-large (Bcl-xL) (antiapoptotic transmembrane protein in the mitochondria) were less responsive to HHT-induced apoptosis than HL-60 cells. It was suggested that Bcl-xL plays a more important role than Bcl-2 and Mcl-1 in protecting against HHT (2)-induced apoptosis. 195 In addition, it was demonstrated in leukemic cell lines K562 that HHT (2) inhibits the synthesis of vascular endothelial growth factor, an important mediator of angiogenesis in the bone marrow. This work thus demonstrated a possible antiangiogenic activity for HHT (2) . 196 A study to identify novel biomarker that might contribute to variation in response to HHT (2) by using a panel of various human lymphoblastoid cell lines was performed. Genome-wide association analysis was applied using single nucleotide polymorphism (SNP) and mRNA expression data. The integrated analysis identified four unique SNPs that were associated with both expression and area under the curve. Functional validation using siRNA knockdown in leukemia cell lines showed that knocking down CCDC88A, CTBP2, SOCS4 genes in U937 and K562 cell lines significantly altered HHT (2) cytotoxicity. 197 The overexpression of EphB4 contributed to IM resistance in CML line cells. Cell lines from a patient with relapsed CML were established at the time of first diagnosis and after being relapsed. EphB4/RhoA may be a new marker for IM resistance. This is supported by the fact that HHT (2) with IM yielded more treatment advantages than IM alone by blocking EphB4/RhoA pathways. 198, 199 Apoptosis varying intervals in K562 cell line treated with HHT (2) was studied. The results showed that the viability of K562 cells reduced gradually from day 1 to day 5 and ascended from day 6 to day 8 after HHT (2) treatment. At the same time, the cleaved caspase-3 expression level in K562 cell lines increased gradually from day 1 to day 7 but decreased at the day 8 (P < .05). From day 1 to day 8 after HHT (2) treatment, the BCL-2 expression level declined first and then went up (P < .05). Autophagosome was also increased at day 8 after HHT (2) treatment. It is concluded that the apoptosis level of K562 cells after being treated with HHT (2) increased at the beginning of the treatment and then it declined, which may be associated by higher autophagy level in the late stage of HHT (2) treatment. 200 A recent study demonstrated that HHT (2) binds to the DNA with an intercalation mode. 201 Also, HHT (2) binds to the telomere G-quadruplex, which makes it a good model for drug design. 202 Analysis of proteins expression profiles in K562 cell line (a CML cell line) in response to HHT (2) treatment showed that nine proteins were downregulated and four upregulated. Aside from alterations in apoptotic proteins and proteins associated with transcription and translation, changes in oxidative stress response and redox reactionerelated proteins, such as HSPs, DJ-1, and thioredoxin were revealed. Specifically, these proteins were decreased after HHT (2) treatment in K562 and primary CML cell lines. 203 Also, it was found that HHT (2) is an effective inhibitor of unwanted angiogenesis. 204, 205 6.1.2 Potential Use for the Treatment of Other Tumors Beside its use in the treatment of CML, HHT (2) is being investigated for other type of tumors alone and in combination with other drugs: HHT (2) significantly inhibited the proliferation of human MM cell lines and tumor cells from patients with relapsed refractory MM in a dosedependent manner. This apoptotic process was associated with the activation of caspase-8, caspase-9, caspase-3, and PARP. 206 The effect of HHT (2) was demonstrated on MM, on the following cell lines AMO-1, IM9, RPMI8226, EJM, U266, and OPM-2; HHT (2) significantly reduced Mcl-1, a crucial protein involved in myeloma cell survival, in all the myeloma cell lines that were examined. Also, HHT (2) reduced the levels of c-FLIPL/S, activated caspase-8, and induced active reduction of Bid. It follows that HHT-induced apoptosis appears to be mediated via both intrinsic and extrinsic apoptosis pathways. The resultant imbalance between BH3-only proteins and Mcl-1 may be pivotal for apoptosis by HHT (2) . Also, HHT (2) treatment resulted in reduced levels of b-catenin and XIAP proteins, which also contribute to disease progression and resistance to chemotherapy in MM. HHT (2) enhanced the effects of melphalan, bortezomib, and ABT-737 when given in combinations with these agents. These results suggest that HHT (2) could constitute an attractive option for MM treatment through its ability to simultaneously target multiple tumor-promoting molecules. 207 HHT (2) and Bortezomib synergistically inhibit SKM-1 MDS cell proliferation and induce apoptosis in vitro. 208 Transmembrane receptor of the tyrosine kinase KIT oncoprotein is an important therapeutic target as part of the type III tyrosine kinases. Some mutations in this receptor (V560G, D816V, or D814Y) give it the ability of auto-dimerization and auto-phosphorylation without activation by its ligand, which can eventually lead to uncontrolled cell proliferation and a resistance to apoptosis. The gain of function associated with these mutations of the KIT receptor plays a critical role in the pathogenesis of SM and gastrointestinal stromal tumors (GISTs). The hypothesis that HHT (2), as an inhibitor of protein synthesis, decreases the protein level of KIT thereby reduces the level of phospho-KIT and repeals its constitutive signaling downstream. HHT (2) is effective both in vitro and in vivo demonstrating that it effectively inhibits tumor growth in mice and induced apoptosis of tumor cells carrying the V560G (sensitive to IM) and D816V (insensitive to IM) mutations of KIT. In addition, HHT (2) significantly prolonged the mouse survival time with aggressive SM (35 days instead of 7) though a rapid resistance to HHT (2) was observed in this case. 209 Potential activity of HHT (2) against clear celletype renal cell carcinoma (RCC) with mutant von HippeleLindau gene was revealed by highthroughput simultaneous screen and counterscreen. 210 In addition, HHT (2) has the potential for treatment of the Gefitinib-resistant nonesmall cell lung cancer, according to an in vivo study. 211 

A signal oncogene is important for some tumors for their continued activity of their malignant phenotype. The BCR-ABL fusion protein in CML is the best-studied example. HHT (2) that inhibits protein synthesis thus reduces the level of the BCR-ABL protein level. The kinase activity of BCR-ABL is directly inhibited by IM and flavopiridol (a transcription inhibitor). The median-effect method was used to evaluate combinations of flavopiridol/HHT (2) and flavopiridol/IM. The resulting analysis indicated synergistic decrease in the clonogenicity. Synergy increased when HHT (2) and IM were administered sequentially as opposed to simultaneously. Transfected IM-resistant Ba/F3 cells that expressed the E255K and T315I mutations of BCR-ABL were not cross-resistant to flavopiridol and HHT (2) . These results provided a rationale for the combination of inhibitors of transcription and/or translation with specific kinase inhibitors. 212 The combination of HHT (2) and DOX had ratio-dependent synergistic activities, with very strong synergism observed at a molar ratio of 4/1 (DOX/HHT) within double-loaded liposomes. According to this study liposomal codelivery of DOX and HHT (2) could synergistically potentiate antitumor effects. 213 HL60/ADR, Kasumi-1, and primary AML cells were pretreated with Decitabine and then exposed to HAA regimen (HHT (2), aclarubicin, and cytarabine (Ara-C)). Decitabine increased cytotoxic effect of HAA regimen and enhanced apoptosis in chemoresistant AML cells. 214 A synergistic effect of Sulforaphane in human CML cell line K562 with chemotherapy drugs HHT (2) and Ara-C was found and indicated clinical potential. 215 The combination of HHT (2) and YM155 was investigated on K562 cell lines, and it was found to be synergistic. The mechanism was related to the apoptosis induced through downregulation of survivin. 216 TNF-related apoptosis-inducing ligand (TRAIL) is a proapoptotic ligand from the TNF-a family. HHT (2) represents a very efficient enhancer of TRAIL-induced apoptosis with potential application in TRAIL-based, anticancer combination therapy. The agonistic anti-TRAIL receptor antibody is a potential antitumor agent. HHT (2) was found to be a very effective (in nano-molar dose) enhancer of TRAIL-mediated apoptosis/growth suppression. Cotreatment of TRAIL-resistant RKO or HT-29 cells with HHT (2) and TRAIL led to the effective induction of apoptosis and the complete elimination of the treated cells. HHT (2) suppressed the expression of the antiapoptotic proteins Mcl-1 and cFLIP and enhanced the TRAIL-triggered activation of JNK and p38 kinases. 217 Another study suggested that HHT (2) can replace toxic substances such as cycloheximide because it activates necroptosis through the same signaling pathways (involving RIPK1/RIPK3/MLKL) to sensitize human cancer cells to TRAIL-induced necroptosis. 218 HHT (2) was combined with antiproliferatives such as alkylating agents, intercalating agents, metal coordination complexes, pyrimidine nucleosides, purine nucleosides, inhibitors of nucleic acideassociated enzymes and proteins, and agents affecting structural proteins and cytoplasmic enzymes, for the treatment of a host with a cellular proliferative disease, particularly a neoplasia. 219 When matrine combined with HHT (2) was tested on the proliferation of human AML cells HL-60, a synergistic effect on apoptosis was noticed. 220,221

The main mechanism of HHT (2) resistance is through a pump that removes HHT (2) from tumor cells. This transporter is called drug-efflux pump P-glycoprotein, also known as MDR1. Through the action of this pump the intracellular drug levels may remain sublethal and, in the case of upregulation of MDR1, cancer cells may survive. The HHT-resistant cell line SKM-1/HHT was established and the prominent overexpression of MRD1 may be the main cause for multidrug resistance. 222 ssHHT (2) derivatives that bypass MDR1 might improve treatment efficacy.

Short hairpin RNAs (shRNAs) were designed, constructed, and transfected into HT9 leukemia cell line, for specific silencing of the MRD1 gene. Resistance against HT (3), DOX, and curcumin was decreased alongside the decreased expression of MRD1. The study indicated that shRNA recombinant plasmids could modulate MDR in vitro. 223 Human MDR B-cell lymphoma cell line BJAB/ADR was found to be resistant to HHT (2), daunorubicin (DNR), etoposide, and mitoxantrone (MTX). 224 A study on the pharmacological effect of HHT (2) on the molecular level was conducted. CET (1) and HHT (2) were tested for resistance on the MDR cell lines. These were the 55 cell lines of the Developmental Therapeutics Program of the National Cancer Institute (NCI, Bethesda, Md., USA) that overexpress the conferring P-glycoprotein/MDR1 (CEM/ ADR5000), which can expel HHT (2) of the cell and consequently which can attenuate its activity. In CEM/ADR5000 cells, a threefold increase to HHT (2) resistance was observed, though the other MDR cell lines did not show cross-resistance. 225 6.1.5 Mechanism of Allergic Reaction HHT (2) was screened and identified as a potential allergic factor. Histamine 1 receptor/cell membrane chromatography (H1R/CMC) model was applied for capturing membrane-retained components, which could induce H1R activation. Retention components were enriched and analyzed by H1R/CMC-HPLC/MS. HHT (2) was recognized, separated, and identified in HHT (2) medication. Ca 2þ flux assay and p-IP3R (channels that most often mediate Ca 2þ release from intracellular stores) expression found that HHT (2) retained by the H1R/CMC model increased phosphorylation of IP3R and promoted cytosolic free Ca 2þ elevation in a dose-dependent manner, which further verified the activity of HHT (2) in activating H1R. This provides an indication that a patient with overexpressed H1R should be aware of possible allergic reaction when applying HHT (2) injection. 226 6.1.6 Other Pharmacological Applications From the chloroform extract of the aerial parts of the only Cephalotaxus harringtonia L. specimen grown in Egypt, five known alkaloids CET (1), HHT (2), HT (3), isoHT (4), and deoxyHT (5) were identified by HPLCeESIeMS. 227 This extract was tested in vitro for its cytotoxicity against HCT116, HepG2, and MCF-7 cell lines and resulted with IC 50 values of 4.77, 12.9, and 17.5 mg/mL, while DOX had IC 50 values of 3.74, 4.57, and 2.97 mg/mL against these cell lines, respectively.

Chikungunya virus (CHIKV) is a mosquito-transmitted virus that has reemerged as a significant public health threat in the last decade. HT (3) displayed potent inhibition of CHIKV infection (EC 50 0.24 mM) with minimal cytotoxicity. HT (3) acts on the postentry stage of the CHIKV replication cycle and strongly interferes with the process of viral protein synthesis. 228 Cryptosporidium parasites cause cryptosporidiosis, a diarrheal disease usually, which can result in fulminant diarrhea and death in AIDS patients and chronic infection and stunting in children. A library of drug repurposing candidates were screened against Cryptosporidium parvum among which HHT (2) showed promising in vitro activity, its IC 50 activity was 0.86 mM. 229 HHT (2) and CET (1) were tested against the hepatitis B virus (HCV), using the bovine viral diarrhoea virus (BVDV) as surrogate for HCV. The activity relative to BVDV was determined by reduction of BVDV-RNA production and protection of infected embryonic bovine trachea (EBTr) cells. Activity against HBV was investigated by HBsAg release and HBV-DNA release from hepatoblastoma HepG2 2.2.15 cells infected with HBV. HHT (2) showed high activity against HBV (IC 50 was 2 mM against HBV, but it has a very low therapeutic index). 230 CET (1) and HHT (2) induced toxicity in EBTr cells and had no protective effect against BVDV. In contrast, they were able to inhibit HBV production at concentrations 10-to 100-fold lower than other molecules inducing cell toxicity, which suggests that they are potentially useful in combined therapy against hepatitis B.

A library of 290 compounds was screened for antiviral activity against Middle East respiratory syndrome coronavirus (MERS-CoV) and severe acute respiratory syndrome coronavirus (SARS-CoV). Selection of compounds for inclusion in the library was dependent on current or previous FDA approval or advanced clinical development. Some drugs that had a well-defined cellular pathway as targets were included. In total, 27 compounds with activity against both MERS-CoV and SARS-CoV were identified. The EC 50 activity of HHT (2) against MERS-CoV was 71.8 nM. 231 HT (3) was extremely effective in inhibiting human retinal glial cell proliferation. Its activity was comparable to other antiproliferative drugs such as colchicine, daunomycin, and 5-fluorouracil. HT (3) is therefore a potential drug for the treatment of proliferative vitreoretinopathy (PVR). 232 In 2009, the inhibitory effects of HHT (2) on CML and B-ALL (B cell ALL) cell line was evaluated in vitro and showed that over 90% of leukemic stem cells were killed after treatment with HHT (2). In contrast, less than 9% or 25% of leukemic stem cells were killed after treatment with IM or dasatinib, respectively. 233 HHT (2) has demonstrated activity against coronavirus with an IC 50 ranging from 11 nM to 1.2 mM. 234 Also, HHT (2) when combined with the saponin digitonin showed activity against Trypanosoma brucei brucei. 235 

HHT (2) or OM (INN) provides a treatment option with a distinct mechanism of action for patients who have not achieved an optimal outcome with TKIs, including patients with the T315I mutation. 236 Recent reviews compiled clinical trials conducted first in China with mixtures of HHT (2) and HT (3), then in Western countries using purified high purity ssHHT (2) reported in 1998 by Robin (OncoPharm founder) also known as OM allowing to confirm its efficacy in CML and put the first European Medicine Agency (EMA) application in 2004 leading to its orphan drug status in 2009, and in the United States with a stable source of hydrosoluble omacetaxine reported in 2001 by Brown (ChemGenex founder). 12,237e240 6.2.1 Approved Clinical Applications HHT (2) was approved in 2009 by the EMA for t-Philadelphia chromosome positive CML in patients who have the T315I BCR-ABL kinase domain mutation and who are resistant to prior IM therapy. 241 It was approved by the FDA in October 2012 for the treatment of CML in CP or AP after failure of two or more TKIs. 242 

To evaluate the pharmacokinetics of HHT (2), sensitive HPLC-tandem mass spectrometry (LC-MS/MS) assays for the quantification of omacetaxine and its inactive 4 0 -desmethyl (4 0 -DMHHT) and cephalotaxine metabolites in human plasma and urine were developed and validated, according to the latest FDA and EMA guidelines. 243 Deuterated isotopes were used as internal standards and these assays were considered very suitable to support clinical pharmacologic studies of OM. Since omacetaxine is mainly metabolized by esterases, the plasma samples were immediately stabilized after collection with an esterase inhibitor and stored at a nominal temperature of À80 C. The validated plasma assay quantifies all analytes in the concentration range of 0.1e100 ng/mL and the urine assay in the range of 0.1e50 ng/mL. At all concentrations, the accuracies were within AE15% of the nominal concentrations and precisions were 15%. The developed methods have successfully been applied in a human mass balance study of omacetaxine.

Nearly 70%e90% of patients with CML in the early chronic phase (CP) who were treated with IM achieved complete cytogenetic response. With a median follow-up period of 5 years, the annual rate of resistance or progression is 3%e4%, and the mortality is 1%e2%. 244 In 2001 a study was conducted on 37 patients 21 years old with recurrent refractory AML, 28 patients were evaluable for response. Four patients responded completely and one responded partially (response rate 5/28 18%); the median duration was roughly 62 days. This study showed potential use of HHT (2) for pediatrics with tolerable toxicity. 245 Two cases of CML patients acquired the T315I mutation after IM treatment; one was treated with recombinant interferon-alpha (IFN-a), and the patient was in complete hematologic response (CHR) 10 months later, while the other patient treated with HHT (2) started at 1.25 mg/m 2 subcutaneously twice daily for 5 consecutive days every month and achieved a CHR after 5 months. 246 The outcomes of 176 patients of AP CML who received treatment with IM were reviewed and compared with the outcomes of 213 historic control patients treated with other drugs including HHT (2) . Treatments included INF-a (100 patients), HHT (2) (35 patients), DNR plus cytarabine toxicities were the major side effects (79% and 73% for CP-CML and AP-CML, respectively), with discontinuation due to toxicity in 10% of CP patients and in 5% of AP patients. 250 The efficacy and safety of HHT (2) in patients with AP-CML previously treated with tyrosine kinase inhibitors were assessed in two phase II trials. 51 patients in AP-CML and 44 in myeloid blast phase (BP-CML) received subcutaneous (SC) omacetaxine 1.25 mg/m 2 twice daily, days 1e14 every 28 days until hematologic response/improvement or any cytogenetic response, and then days 1e7 every 28 days until disease progression. An MHR was 37% in patients with AP-CML and 9% in patients with BP-CML (22% and 5% in those were with a history of T315I mutation). 251 In 2007, a 63-year-old male CML patient was reported harboring a T315I BCR-ABL mutation while on IM treatment that completely disappeared when he was treated with HHT (2). 252 In 2013, HHT (2) produced good clinical responses with acceptable tolerance levels in patients with CML-CP who had been treated with two or more TKIs. These Data were analyzed from two phase II clinical trials of 81 patients with CML-CPs. The median age of those in the analysis was 59 years (range, 26e83 years). Eight patients completely responded. The median duration for the study was 17.7 months. Fifty-six patients (69%) achieved and/or maintained hematologic response for at least 8 weeks. The overall survival was 34 months. 253 In 2015, a study was carried on 31 CML patients with complete cytogenic response to compare the immunological effect on T lymphocytes between HHT (2) (15 patients) and IM (16 patients); the study indicated that both have the same effect. 254 

A recent review discussed the use and effectiveness of HHT (2) in the treatment of intractable CML, with coverage of its pharmacology, mode of action, and pharmacokinetics. 237 A SC formulation demonstrated efficacy and safety in phase 1/2 trials. HHT (2) at 1.25 mg/m 2 SC was administered (twice daily) BID, days 1e14 every 28 days for two cycles. This mode of administration was developed because in early clinical studies short intravenous (IV) infusions (over 60e90 min) was used, but life-threatening hypotension and tachycardia were observed at dose levels above 3e4 mg/m 2 . 255 A stable aqueous solution for injection of HHT (2) was prepared with tartaric acid at pH around 4. 256 It is only recently however that the site of protonation, although obvious, was unambiguously seen at N9 of CET (1) or HT (3) as established through X-ray diffraction studies of various organic salts, 76 an observation in concordance with Steitz 176 (see Fig. 25 ) and Takano's 60 results implying that N9 should be available for protonation to ensure its biological activity. Therefore, HTs seem to be active via their protonated form. The maximum tolerated dose was estimated to be 5 mg/m 2 . 257 It is provided in vials that contain 3.5 mg HHT (2) as a sterile, preservative-free, and white to off-white lyophilized single-use solution for injection. 241 In a recent patent HHT (2) was incorporated in phycobiliprotein nanoparticles. 258 HHT (2) can be given as aqueous solution preferably at a pH of about 4. Also it can be provided orally. 259e263 The protonated form of HHT (2) was patented for potential application in oral and IV dosage forms. 264 MM remains an incurable disease in most patients. Newly developed high-proportion polyethyleneglycol (H-PEG) of long-circulating liposomes (LCL) loaded with HHT (2) (HHT-LCL-H-PEG) may be regarded as a promising nano-device to deliver anti-MM drug HHT (2) for treatment of MM patients. 265 This HHT-LCL freeze-dried powder can significantly improve storage stability, can make HHT-LCL freeze-dried powder convenient to dilution administration, make drug in liposome-sustained release, and extend the circulation time in blood, to improve the utilization rate of medicament effective component, and reduce toxic and side effects.

The major metabolites of HHT (2) are CET (1) and 4'-desmethylHHT (31), they are detected by HPLC. 266 In another study using 14 C HHT and detection by HPLC mass spectrometry (MS) (high resolution) in combination with offline radioactivity determination, six metabolites were detected in the plasma, urine, and feces, unchanged HHT (2) was predominant. 267,268

Trials have been carried on combining HHT (2) with other anticancer drugs to synergize and potentiate the activities of these drugs and improve therapy outcome. In some cases, the associated drugs can be efficiently codelivered in the form of liposomes. 269 

In 2002, the efficacy and safety of a combination regimen of simultaneous HHT (2) and IFN-therapy was proved with 37 patients with CP-CML who were not treated previously by either agent. The CHR rate was 85%. 270 In 2009, a clinical trial phase II carried on 44 previously untreated patients with Philadephia-positive chromosome (Phþ) CP-CML, the treatment consisted of HHT (2) (2.5 mg/m 2 per day) and cytarabine (7.5 mg/m 2 per day). Both drugs were given together via continuous IV infusion for 7 days. 36 of 44 patients (82%) achieved a CHR. Although HHT/cytarabine was generally well tolerated, no significant amelioration was observed for patients with interferon-refractory CML and result was not nearly as high as association with IM in newly diagnosed patients. 271 In 2010 in China, 12 patients with CML in blast crisis who have failed prior single-agent therapy with IM were enrolled in clinical trial to investigate the induction therapy of granulocyte colony-stimulating factor (G-CSF) and low-dose HHT (2) as well as standard-dose IM (GHI). The outcome of this study suggested that the GHI regimen may overcome disease-poor response to conventional doses of IM followed by allogeneic hematopoietic stem cell transplantation (allo-HSCT). 272 Ninety patients with Phþ CML in early CP received the triple regimen IFN-a, ara-C, and HHT. After a median duration of 16.5 months of therapy, 78 patients had their therapy changed to IM, 85 patients (94%) achieved CHR, and 67 patients (74%) had a cytogenetic response (Ph suppression to 90%). 273 14 patients were treated with a regimen that combines HHT (2) and IM among which six cases failed and eight cases sub-optimally responded to IM therapy. This trial proved the efficacy and safety of combining both drugs. 274 

In 2009, 87 AML patients were treated with HHT (2), Ara-C, and DNR (HAD) regimen, the complete remission rate of the 87 cases was 80/87 (92%). 275 Six hundred and twenty three newly diagnosed Chinese patients with AML were treated with induction chemotherapeutic regimens based on idarubicin, DNR, and HHT (2). Total complete remission rate was 66.5%, median survival was 18 months, and estimated survival at 3 years was 30.8%. The third year relapse rate was 55.1%. 276 A group of 254 Chinese patients with de novo AML with glutathione S-transferase theta 1 (GSTT1) and glutathione S-transferase mu 1 (GSTM1) mutations were administered in a clinical study. As induction therapy, the patients were treated with HHT (2) 2.5 mg/m 2 intravenously on days 1e7, Ara-C 150 mg/m 2 per day continuous infusion on days 1e7, and DNR 45 mg/m 2 intravenously on days 1e3 (HAD regimen). The second course of HAD induction therapy was applied to patients who were not in complete remission after the initial course of induction therapy. Postremission therapy consisted of HA regimen (HHT (2) and Ara-C) (course 1 and 4), DA regimen (DNR and Ara-C) (courses 2 and 5), and MA regimen (MTX and Ara-C) or HAM (Ara-C days 1e4, combined with MTZ 8 mg/m 2 , IV on days 5e7) (courses 3 and 6). Overall survival and disease-free survival of complete remission patients in GSTT1 and GSTM1 double present group was better than in GSTT1and/or GSTM1-group. 277 A regimen was investigated to explore the effect of low dose of HHT (2) and Ara-C combined with G-CSF on patients with relapsed or refractory AML: 35 out of 67 (52.2%) patients achieved complete remission and 8 out of 67 (11.9%) achieved partial remission. The overall response rate was 64.1%. 278 In 2003 a clinical trial of combined regiments HHT (2), Ara-C, and etoposide (VP-16) was conducted on 12 pediatric patients, 9 patients achieved complete remission within 20e45 days after one course of therapy. One patient had partial remission this clinical trial gave good indication for this combination. 279 HAA regimen, was evaluated in 46 patients with refractory and/or relapsed AML for efficacy and toxicity. 80% of patients achieved complete remission. 280 b-Elemene emulsion had a curative effect in treatment of refractory/ relapsed AML in combination with HAA regimen. This was clinically proven by comparing two groups: one using the HAA alone and one using b-elemene emulsion with HAA. The total effective rate in the b-elemene emulsion plus HAA group was 80.8%. However, the total effective rate in the HAA only group was 52.9%. 281 Sixty-seven patients with relapsed or refractory AML were enrolled in a clinical trial. They received the following treatment: HHT (2), Ara-C, and G-CSF (HAG). All the patients were treated with HAG regimen. 35 patients (52.2%) achieved complete remission and 8 (11.9%) partial remission. The overall response rate was 64.1% 282 One hundred and ninety six untreated de novo AML patients were recruited in China for a clinical trial using HHT (2) . Patients were treated with HHT (2) and Ara-C (HA)-based induction therapy composed of three chemotherapeutic drugs (HAD/M, D-DNR, M-MTX). The complete response rate was 153 of all cases (78.1%). 283 One hundred and seventy-one pediatric patients were recruited in a trial. They received a regimen including HHT (2), cytarabine and DNR with a median age of 7.58 years. Complete response was obtained in 140 of 171 (81.9%) cases within 60 days. The 5-year event-free survival was 52.75%. 284 Fourty-eight patients median aged 35 (14e57) years were enrolled into a clinical study to evaluate the efficacy and toxicity of HAA regimen as an induction therapy in the treatment of de novo AML. The HAA regimen was a well-tolerable and effective induction regimen for patients with de novo AML. 285 A case was reported with primary refractory FLT3-ITD-positive AML was treated with combination of low dose HHT (2) and Sorafenib (salvage treatment); the therapy was successful and the patient was in complete remission. 286 Ninety elderly patients with AML were chosen and divided into CAG (low-dose cytarabine and aclarubicin in combination with G-CSF) group (52 cases) and HAG (HHT (2) and Ara-C combined with G-CSF) group (38 cases). In the CAG group were 38 cases with complete remission (73.08%) and total efficiency was 84.62%; 40 cases with complete remission in HAG group (36.84%) and total efficiency was 73.68% in the HAG group. CAG regimen was more efficient than HAG in elderly patients with AML. 287 In a recent clinical study for elderly patients with de novo AML, 56 patients were enrolled in this study. They were treated with HHT (2) and cytarabine in combination with G-CSF (HAG). Updated results showed good overall survival. 288 A recent metaanalysis provided an overview of the effectiveness of HHT (2) combination regimens to treat AML. Twenty-one studies were analyzed including 1310 patients, 229 pediatric, and 216 elderly. HHT (2) was given in combination with cytarabine, DNR, idarubicin, aclacinomycin, MTX, or G-CSF. This retroanalysis showed the superiority of HHT (2)-containing regimens in AML treatments.

Clearly, other potential clinical applications are the other subtype of leukemia. Consequently different clinical trials with different regiments took place:

Thirty-one adult patients suffering from idiopathic thrombocytopenic purpura (ITP) were treated with low-dose combination chemotherapy consisting of cyclophosphamide and prednisone on days 1e7, combined with vincrinstine on day 1, and/or azathioprine on days 1 to 5 or 7 and/or HHT (2) on days 1, 3, 5 or on days 1e7 and/or etoposide on days 1e7. 289 The output was as following: 13 of 31 (41.9%) patients had complete remission, 9 (29.0%) had partial remission, and 9 (29.0%) had no response.

A 45-year-old female patient was reported with de novo acute megakaryoblastic leukemia (rare type of AML) with BCR/ABL rearrangement and der(16)t(1; 16)(q21; q23) translocation but negative for t(9; 22) Ph chromosome. On HAD induction chemotherapy, the patient achieved partial hematological remission. The patient was then switched to IM plus one cycle of CAG regimen (low-dose cytarabine and aclarubicin in combination with G-CSF). She achieved complete remission, the disease recurred after 40 days and the patient eventually died of infection. 290 Thirty-four myelodysplastic syndromes patients with MDS-RAEB and RAEB-T were recruited for a clinical trial of two regimens therapy of either HAA regimen or Idarubicin Ara-C (IA) regimen. It was suggested that the HAA regimen is more effective than IA and well tolerated. 291 Thirty-three elderly patients were enrolled in a clinical trial, 21 patients with high-risk MDS, and 12 with MDSeAML to evaluate the efficacy and toxicity of CHG (cytarabine, HHT (2), and G-CSF) induction chemotherapy (low-dose cytarabine, HHT (2) with G-CSF priming) for elderly patients with high-risk MDS or AML transformed from MDS (MDSe AML). As outcome of this study the overall response rate was 66.7% after one course of the CHG regimen with 19 patients achieved complete remission (57.6%) and 3 patients achieved partial remission (9.1%). 292 In a long trial of 9 years, 53 patients were enrolled to evaluate the effect of combination therapy of HHT (2) plus all-trans-retinoic acid (ATRA)e based therapy for the treatment of APL as induction therapy followed by three courses of consolidation chemotherapy and 2 years sequential therapy. All patients achieved complete remission. 293 A retrospective study on high risk 132 patients with higher risk MDS was done to compare two treatment combination therapy of low-dose CHG and Decitabine. Complete remission was not significantly different between the groups (27.1% with Decitabine vs. 30.6% with CHG). 294 In a clinical study of total of 98 patients at the acute phase of CML with invasive pulmonary aspergillosis (IPA) were treated with Voriconazole plus a low-dose of HHT (2) and IM. This study suggested that this combination could be used as the first-line therapy for IPA patients with CML. 295 The 49 cases in observation group were treated with the combination while the control group was treated with fluconazole, low-dose of HHT (2), and IM. The effective rates of the observation group and the control group were 95.92% and 91.84%, respectively with no significant differences. The study concluded that Voriconazole and low-dose HHT (2) and IM could be used as the first-line antifungals for the treatment of IPA patients with CML.

Thirty-three patients were enrolled in a clinical study to investigate the effectiveness of HA regimen for CML-MBC, overall improvement cytogenetic response was 60.1%, 7 patients (21.2%) had partial response. The study concluded that it is an effective treatment and that CD34 þ CD7 þ cells may be one of the valuable clinical parameters to assess treatment effectiveness. 296 

The major side effect is myelosuppression (a condition in which bone marrow activity is decreased, resulting in fewer red blood cells, white blood cells, and platelets). Myelosuppression is a side effect of some cancer treatments, 257, 271, 272, 279 with gastrointestinal manifestation including diarrhea occurring as well in some cases. 252 Those side effects were explained due to the effect of HHT (2) on the epithelial permeability of intestinal Caco-2 cell monolayers. 297 6.3 Natural Extracts From Cephalotaxus sp. and Natural Chinese Medicine

The essential oil of C. griffithii inhibited proliferation and migration of human cervical cancer cells. 298 The oil from C. griffithii needle extracted with petroleum ether was tested for its effect on proliferation of cancer cells (MTT assay on HeLa, ZR751, and HepG2 cell lines). The extract showed important phyto-components with multiple cellular targets for control of breast cancer. 299 Patents based on Chinese traditional medicine were filed containing Cephalotaxus as an ingredient combined with different plants' extracts and are used for the following indication: treatment of phlegm stagnation type osteoarthritis, 300 liver cancer, 301 ischemic intestinal colic, 302 for phthisis, 303 hepatitis syndrome, 304 for leukemia, 305 traumatic bleeding consists, 306 hookworm disease, 307 ovarian cancer, 308,309 lump-resolving agent in gastric cancer, 310 treatment of gastric cancer, 311 liver cancer, 312 collateral cancer, 313 rectal cancer, 314, 315 phlegm and fluid-retention lung-obstructing type cardiac insufficiency, 316 treating liver-depression qi-stagnation type breast cancer, 317 spleen-kidney-qi-deficiency bladder cancer, 318 and lung cancer. 319 A Cephalotaxus plant preparation inhibits the adhesion and migration of breast cancer, colorectal cancer, prostate cancer, lung cancer, gastric cancer, osteosarcoma, and other cancer cells, 320 liver fats. 321 (Table 20) .

A patent for dermatological application reported that HHT (2) was an active inhibitor of matrix metalloproteinase-1 that can be used for reduction and prevention of wrinkles. 322 Another preparation relates to a veterinary drug for treatment of bovine mastitis. 323 Eventually, patents were filed using preparations that contain HHT (2) for treatment of water and purifying it from cyanobacteria. 324, 325 Note Added In Proof During the preparation of this manuscript, additional synthetic studies 326, 327 and a review were published. 328 Traumatic bleeding consists 306 8 Hookworm disease 307 9 Ovarian cancer, 308, 309 10 Lump resolving agent in gastric cancer 310 11 Gastric cancer 311 12 Liver cancer 312 13 Collateral cancer 313 14 Rectal cancer, 314, 315 15 Phlegm and fluid-retention lung-obstructing type cardiac insufficiency 316 16 Liver-depression qi-stagnation type breast cancer 317 17 Spleen-kidney-qi-deficiency bladder cancer 318 18 Lung cancer 319 19 Adhesion and migration of breast cancer, colorectal cancer, prostate cancer, lung cancer, gastric cancer, osteosarcoma and other cancer cells 320 20 Liver fats 321

This result indicates the disappearance of the mutated clone. 248 In 2009, a year-long study on long-term treatment with HHT (2) was carried out. A year-long study on long-term treatment with HHT (2) was in the control group, 31 were treated with IFN-a, and 48 with hydroxycarbamide. Seventeen patients who were treated with IFN-a failed to achieve cytogenetic response within 12 months and subsequently received HHT (2) as a salvage treatment

18% of CP-CML patients achieved a major cytogenetic response (MCyR) with a median duration of 12.5 months. Responses were maintained for 12 months in three of forteen responders, and the median overall survival (OS) was 40.3 months. Among patients with AP-CML, 14% achieved or maintained a major hematologic response (MHR) for a median of 4.7 months, MCyR was not achieved, and the median OS was 14.3 months

The Alkaloids

The Alkaloids

Anticancer Agents from Natural Products

Anticancer Agents from Natural Products

Abstracts of Papers

Faming Zhuanli Shenqing CN 104193755

The IUCN Red List of Threatened Species

Faming Zhuanli Shenqing CN 102976924 A

This work has been released into the public domain by its author

Faming Zhuanli Shenqing CN 104667289 A

Zhonghua Shiyan He Linchuang Ganranbing Zazhi

Faming Zhuanli Shenqing CN 104306792

Faming Zhuanli Shenqing CN 103990002

Faming Zhuanli Shenqing CN 103721167 A

Faming Zhuanli Shenqing CN 103610743 A

Faming Zhuanli Shenqing CN 103550547 A

Faming Zhuanli Shenqing CN 103520324 A

Faming Zhuanli Shenqing CN 105147940 A

Faming Zhuanli Shenqing CN 105193994 A

Faming Zhuanli Shenqing CN 105106469 A

Faming Zhuanli Shenqing CN 104324356 A

Faming Zhuanli Shenqing CN 104800457 A

Faming Zhuanli Shenqing CN 104784401 A

Faming Zhuanli Shenqing CN 104784351 A

Faming Zhuanli Shenqing CN 104474475

Faming Zhuanli Shenqing CN 104784406

Faming Zhuanli Shenqing CN 104784423

Faming Zhuanli Shenqing CN 104721400

Faming Zhuanli Shenqing CN 104667025

Faming Zhuanli Shenqing CN 104623242

Faming Zhuanli Shenqing CN 105012613

Korean Kongkae Taeho Kongbo KR

Faming Zhuanli Shenqing CN 105012651

Faming Zhuanli Shenqing CN 103947691

Faming Zhuanli Shenqing CN 105016437

Studies in Natural Products Chemistry

Authors are thankful to BioCIS, Centre National de la Recherche Scientifique (CNRS), Labex LERMIT (ANR-10-LABX-33), Université Paris-Sud, Université Paris Descartes, Sorbonne Paris Cité, and Université Paris-Saclay (France) for their financial support during the time devoted to this contribution. A.R. thanks ISSAT (Syria) for financial support.