Steroid-fullerene hybrids: a review

Among carbon-based materials, fullerenes have emerged as very fascinating nanocarbons on account of the interesting properties and applications of some of their derivatives in medicine or materials chemistry. The low solubility of C 60 , the most famous family member, has limited its use in medicinal chemistry but its covalent functionalization with different moieties such as porphyrins or other bioactive molecules changes its physicochemical properties, such as its solubility and biocompatibility. New hybrid compounds, bearing fullerene and steroid units in the same molecule can be regarded as promising functional chimeras with encouraging biomedical applications and have become a topic of considerable interest in the last few years. Recent developments in the syntheses of steroid-fullerene hybrids are described herein.


Introduction
Steroids constitute an extensive and important class of biologically active polycyclic compounds that are widely used for therapeutic purposes. 1-3 Even after decades of research, the total synthesis of steroid nuclei by improved strategies continues to receive considerable attention. Numerous methods 4,5 have been exploited for the total synthesis of steroids which are widely distributed in nature, and which possess practical medical importance. Research into steroid total synthesis continues to this day.
In recent years, hybrid systems have grown in importance due to their many applications in medicine and drug development. Hybrid systems are formed from different molecular entities, natural or not, to obtain functional molecules in which the characteristics of various moieties are amplified or modulated or lead to completely new properties. These hybrids can be made from carefully selected components, either through direct covalent bonds or through the integration of key functional characteristics. These promising molecules make it possible to generate new molecular entities by intelligently combining two or more different classes of compounds of synthetic or natural origin. The interest is to be able to mix the structural characteristics of two or more biologically active molecules in the same molecule or their covalent coupling can improve the characteristics of the individual components or reveal new properties. In all cases, this approach is interesting because it offers a multitude of possibilities for generating a wide range of molecules for applications in biology, medicine and in the field of materials. This article provides an overview of the various syntheses of steroid-fullerene hybrids along with interesting biological activities. To the best of our knowledge and much to our surprise, there are no reports on this subject. Scheme 1. The Bingel reaction.
In 1999, Schuster and coworkers 34 reported the synthesis of steroid-fullerene hybrids using the Prato reaction. 35 The Prato reaction is a particular example of the well-known 1,3-dipolar cycloaddition of azomethine ylides to olefins. The amino acid N-methylglycine reacts with paraformaldehyde when heated at reflux in toluene to provide an ylide which reacts with a double bond in a 6,6-ring position of the fullerene via a 1,3dipolar cycloaddition to yield a pyrrolidinofullerene (Scheme 2).

Scheme 2. The Prato reaction.
Here, the Prato reaction 35 was used to synthesize the steroidal fulleropyrrolidine precursors 1, 2, and 3 ( Figure 1). Union of these precursors with tetraphenylporphyrin carboxylic acid 4 was realized through a standard EDCI coupling to furnish the fullerene-steroid hybrids 5-7 ( Figure 2). In 2001, Yang et al. 36 described the synthesis of two new steroid-fullerene hybrids by a Diels-Alder reaction as the key step.
The synthetic pathway is depicted in Scheme 3. Treatment of the available steroid 3'-acetoxypregna-5,16dien-20-one 8 with LiOH led to 9 in 90% yield. This latter was easily transformed by reaction with tertbutyldimethylsilyl triflate into the desired silyloxydiene 10 in an isolated yield of 85%. Reaction of C60 with 10 in toluene at 90 °C for 2 h led to the desired molecule 12 in a yield of 30% (in two steps, 40% yield based on C60 consumed) after hydrolysis of the enol silyl ether and deprotection of the hydroxy group at C-3 using classic acid conditions.
A biological study on the cytotoxic effects of hybrid 12 has been carried out at the cellular level. The authors observed its effect on the transmembrane membrane of the reconstituted sarcoplasmic reticulum Ca 2+ -ATPase (SR Ca 2+ -ATPase) in soybean phospholipid liposomes. Thus, they were able to conclude that compound 12 can inhibit the enzyme. In addition, they also observed its effect on the survival of human lung adenocarcinoma A549 cells. The authors find that the survival of A549 cells after treatment with the steroid-C60 12 is reduced. Eukaryotic cells are rich in membrane structures. Steroid-fullerene hybrids could be the cause and lead to abnormal cell functions and therefore decrease the survival of A549 cells. The authors were able to observe the formation of product 21 containing two additional oxygen atoms resulting from the Bingel-Hirsch reaction between compound 20 and C60 (see scheme 5). Another Diels-Alder cycloaddition reaction with 1 O2, like that observed for ergosterol itself, 38 takes place at the 5,6-diene moiety of ring B. It is interesting to note that this cycloaddition is observed only when the diene system is in an environment containing oxygen and an effective triplet sensitizer, such as C60. Thus, molecular oxygen, generally in a triplet ground state, is excited in a singlet state and reacts as a dienophile with the diene unit of the steroid leading to an endoperoxide following a Diels-Alder cycloaddition reaction.
The proposed structures are confirmed by their electronic spectra as well as by cyclic voltammetry, which allows concluding to the formation only of mono adducts, without observation of any formation of biscycloadducts which would come from the intramolecular Diels-Alder cycloaddition of the steroid diene part of ergosterol at C-7 to the fullerene double bond.
These new hybrid derivatives 17, 18, and 21, consisting of both fullerene and steroid units, can be considered as promising functional systems. The potential biomedical applications of these hybrid molecules have not been reported to date.

Scheme 5. Synthesis of steroid-fullerene hybrid 21.
In 2013, the same authors 39 used epiandrosterone which is an important steroid to synthesize steroidfullerene hybrids. Indeed, this hormone is the natural metabolite of dehydroepiandrosterone via the enzyme 5α-reductase and is known for its low androgenic activity. In addition, it is present in most mammals, including pigs. 40 Prato's method was here used to synthesize potentially biologically active fullerene-steroid hybrids from formyl-substituted steroid derivatives. The fullerene-steroid conjugates 27-30 were prepared in a multistep synthetic procedure in which the C60 unit has been connected to the steroid unit by Prato reaction from pristine [60]fullerene and the respective formyl-containing steroids. Thus, the first step required the preparation of formyl derivatives 25 and 26 as depicted in Scheme 6. The convenient transformation of the hydroxyl group at C-3 of the epiandrosterone 22 by oxidation gave the corresponding 5α-androstan-3,17-dione 23, whereas acetylation of 22 afforded 3β-acetoxy-5α-androstan-17-one 24, both in yields like those previously reported. The corresponding formyl derivatives, 3-chloro-2-formyl-17-oxo-5α-androstan-2-ene 25 and 3β-acetoxy-17chloro-16 -formyl-5'-androstan-16-ene 26, were obtained respectively by a Vilsmeier-Haack reaction of 23 and 24 with phosphorus oxychloride and dimethylformamide in dichloromethane. The yields obtained are good (78% and 50%, respectively) and the products crystallize after basic hydrolysis with aqueous sodium acetate.
Finally, 1,3-dipolar cycloaddition reactions of azomethine ylides generated in situ with C60 according to the Prato protocol leads to the production of N-methyl-2-substituted pyrrolidino [3,4: 1,2] [60] fullerenes. Thus, the reaction is carried out by mixing the chloroformyl derivative (25 or 26), C60 and sarcosine (N-methylglycine), the whole brought to reflux in toluene under an argon atmosphere for 6 h (see Scheme 6). The formation of the products is confirmed by the color change of the solution from purple to brown.

Scheme 6. Synthesis of steroid-fullerene hybrids 27-30.
Flash chromatography makes it possible to easily separate the diastereoisomers and to obtain compounds 27, 28, 29 and 30, in the form of stable brown solids, with respective yields of 42%, 31%, 28% and 47%. The stereoselectivity, which provides diastereomers 27 and 30 as main products, is relatively modest. This result agrees with the fact that the electrophile, the C60 attacks preferentially on the Re face of the 1,3-dipole.
The absolute configuration of fulleropyrrolidines was easily determined using chiroptic properties with Cotton-typical effects in CD spectra of chiral adducts. Thus, the absolute C2 configuration for compounds 27-30 was determined to be: (2R) for 27, (2S) for 28, (2S) for 29 and (2R) for 30. Finally, the authors observed that the conjugation of epiandrosterone derivatives with C60 increases the solubility of these new hybrid systems (27)(28)(29)(30) in organic solvents such as chloroform, dichloromethane and dimethylformamide, among others. It is therefore undeniable that the presence of the steroid unit in the new hybrid compounds (27, 28, 29 and 30) improves their solubility. This result allowed the authors to consider other investigations on these structures, especially biological ones.
In 2014, the same authors 41 described the multistep synthesis of a fullerene hybrid dumbbell having two fullerene units connected through an epiandrosterone molecule. Thanks to the presence of a formyl group at the C-16 position of ring D of the starting product 26, the first C60 unit was introduced by a 1,3-dipolar cycloaddition reaction. And after the transformation of the acetate group attached to position C-3 in ring A into the corresponding malonate, a cyclopropanation reaction made it possible to introduce the second unit of C60 (see Scheme 7).
After hydrolysis with potassium carbonate in methanol of the acetate group in C-3 position of the 3βacetoxy-17-chloro-16-formyl-5α-androstan-16-ene 26, the 17-chloro-16-formyl-3β-hydroxy-5α-androstan-16ene 31 was isolated in good yield (92%) and allowed the incorporation of the malonate moiety by reaction of 31 with (ethoxycarbonyl)acetyl chloride in anhydrous dichloromethane and pyridine at 0 °C. This latter bearing a formyl group was obtained as a pale-yellow solid in 82% yield. The same year, Bjelakovic et al. 42 reported the synthesis, morphological study, and preliminary in vitro antioxidant capacity of new fullerene-steroid hybrids. The synthetic strategy to obtain the fullerene-peptide-steroid hybrids 40 and 42 are depicted in Schemes 8 and 9 and show firstly the synthesis of a steroid peptide dyad and then its conjugation with the fullerene subunit based on esterification and amidation reactions with protection-deprotection steps required. Two different activating systems, DCC/DMAP and EDC/HOBt in dichloromethane were used to optimize these reactions. The DCC/DMAP-assisted esterification of N-protected GABA 36 with sterol 37 led smoothly to the ester 38 in very good yield (80%), while in the presence of EDC/HOBt no product was isolated. The steroid peptide dyad 39 was coupled with Boc-Gly-OH 43 after a quantitative Boc-deprotection with TFA. This latter was isolated with an excellent yield (93%) and used for two different syntheses of steroidal hybrids. The reaction of 40 with Fp-GABA-OH led to fullerene-peptide-steroid hybrid 40 in a 63% yield and reaction with GABA afforded a steroid dipeptide dyad 41 presenting an elongation of the peptide chain (Scheme 9). Finally, triad 42 was isolated in a poor yield of 19% after Boc-removal and a subsequent coupling with a fullerene unit. The yield is lower than this obtained using a shorter peptide. The DCC/DMAP activating system seems to be somewhat more efficient for amidations including both Boc-Gly-OH and Fp-GABA-OH (reactions 3/4 and 4/5, respectively) than for esterification of Boc-GABA 36.
This synthetic approach can be used to generate different steroid-fullerene derivatives with potential biological activity. High antioxidant activity of these new hybrids, twice and 13-fold better than fullerene C60 and standard antioxidant agent vitamin C, respectively, was observed after an in vitro study using the Ferrous ion Oxidation-Xylenol orange (FOX) method. To obtain the target steroid-fullerene hybrids, the authors first synthesized the imine 43 by reaction of the previously reported 39 3-chloro-2-formyl-17-oxo-5α-androstan-2-ene 26 and the glycine methyl ester hydrochloride using triethylamine as a base and dichloromethane as solvent at room temperature (see Scheme 10). This imine was isolated in a good yield (77%) as a yellow oil. The coupling of this latter with C60 using anhydrous silver acetate along with the achiral ligand [dppe, 1,2-bis(diphenylphosphino)ethane] in toluene at room temperature led to a mixture of the two cis-diastereomers 44 and 45. It is worth pointing out that the stabilization of the N-metallated azomethine ylide is provided by the allylic moiety of the steroid unit. The two products were separated by flash chromatography to give 44 in a yield of 32% and 45 in a yield of 25% (see Scheme 10). We can see that the stereoselectivity is modest due to poor chiral induction of the steroid moiety on the attack to one of the faces of the dipole in the fullerene derivative. The same reaction was done with other catalytic systems to ascertain their stereocontrol in the 1,3-dipolar cycloaddition of azomethine ylides onto C60 even in the presence of a chiral moiety, such as the imine 43.
This is an efficient stereodivergent protocol to synthesize hybrid systems in a diastereoselective manner allowing access to a variety of optically active fullerene hybrids.
In 2018, the same authors 48 described the synthesis of new [60]fullerene-steroid hybrids by Bingel-Hirsch protocol of the corresponding steroid containing malonates with C60.
The synthesis started with the preparation of compounds 52, 53, and 54 from commercially available epiandrosterone 48, in which functional groups in C3 and C17 positions were transformed. Thus, the steroid derivative 48 was treated with acetic anhydride in pyridine and led to the corresponding 3β-acetoxy-5αandrostan-17-one 49 in a good yield. As expected, the reduction of compound 49 with sodium borohydride in methanol afforded with stereoselective control the 3β-acetoxy-5α-androstan-17β-ol 50. In the following step, 5α-androstan-3β,17β-diol 51 was isolated after basic hydrolysis of the previous compound in a similar yield as this reported. 49,50 The sterol derivatives 48, 50, and 51 were transformed into the corresponding malonates 52-54 by reaction with the commercially available (ethoxycarbonyl)acetyl chloride (Scheme 12). The reaction was carried out in anhydrous dichloromethane and pyridine at 0°C and led to the derivatives 52, 53, and 54 in 80, 72, and 75% yields, respectively.
The steroid-fullerene hybrids 52, 53, or 54 were synthesized using the Bingel-Hirsch conditions by reaction of C60 and the malonyl steroids 52, 53, or 54, with CBr4 and DBU at room temperature (see Scheme 13). The cyclopropanation process was confirmed by the change of the color of the solution (purple to brown) leading to the [6,6] mono-adduct. In the case of compounds 55 and 56, the reaction was finished in 2 h and furnished after purification by flash chromatography, compounds 55 and 56 in 68 and 58% yields, respectively, as stable brown solids. While, the reaction between C60 and dimalonate 54 led to three hybrid systems 57, 58, and 59 in 42%, 12%, and 11% yields, respectively. It is interesting to note that the DHEA moiety in the [60]fullerene systems 63, and 64 increases the solubility in organic solvents, such as chloroform, dichloromethane, dimethylformamide, and dimethylsulfoxide. This fact might be useful since biological studies could be realized in some of these solvents. Indeed, in general, in vitro tests with the HIV-1 protease are performed in DMSO 52 due to the poor solubility of fullerene derivatives in water.
The TEM analysis showed that steroid-fullerene hybrids in water give nano-sized spherical vesicles with highly heterogeneous size populations. Moreover, DFT theoretical calculations revealed that hydrogen bonds play a major role in the geometry of these hybrid systems. And finally, these derivatives could have an application as HIV-1 protease inhibitors as suggested by molecular docking studies. All the characteristics of C60-steroid derivatives (including polarity, lipophilicity) constitute promising results to further explore the biomedical aspects of these molecular chimeras. The formation of the five-membered ring occurred with the formation of a new stereogenic center on the C-2 of the pyrrolidine ring by a stereospecific syn process. Since the configuration of the stereogenic centers of the steroid always remains the same, the reaction led to a mixture of diastereoisomers of derivatives 66 and 67, which were separated by flash chromatography. Thus, products 66 and 67 were isolated as stable brown solids with yields of 68% and 13%, respectively. The diastereoselectivity of the reaction could be explained by a preferential electrophilic attack of C60 on the Re face of the 1,3-dipole. As a result, the steric interactions due to the presence of the methyl group at C-20 are at the origin of the attack on the Si face. Thus, the 5:1 ratio (Re:Si) observed is explained by an attack on the [60]fullerene on the Re face of the 1,3-dipole due to lower steric hindrance.
Here too, it is worth noting the increase in the solubility of derivatives 66 and 67 in organic solvents. This fact is very interesting to consider possible biological studies on materials. Additionally, these hybrids can potentially be used as HIV-1 protease inhibitors. These are very encouraging results that could be the source of many applications (biology, materials, etc.). Recently, the same authors 54 reported the synthesis and characterization of a fullerene-steroid hybrid that contains H2@C60 and a dehydroepiandrosterone moiety synthesized by a cyclopropanation reaction.
The authors observed that the encapsulation of a hydrogen molecule does not in any way modify the reactivity or the stereoselectivity of the C60 hollow fullerene. 24 The synthesis is depicted in Scheme 16. The new endohedral steroid-fullerene hybrid 69 was prepared by applying the Bingel-Hirsch protocol between 3β-ethyl malonate-5-androsten-17-one 52 51 and H2@C60. The H2@[60]fullerene, compound 52, CBr4, and DBU mixture produced at room temperature very quickly leads to a change in the color of the solution from purple to brown, identical to that observed when the reaction is carried out using unmodified C60 51 and indicative of the success of the reaction. Thus, the [6,6]-closed mono-adduct 69 was isolated in a good yield (76%) as a stable brown solid. The endohedral hybrid is obtained with yields and reaction times similar to those obtained during hollow 60]fullerene reactions with steroids. 51 It is therefore interesting to note that the presence of the H2 molecule inside the cavity of the fullerene does not affect the reactivity of the hollow fullerene. Furthermore, the authors have shown that in the hybrid system 69, the hydrogen molecule does not present a clear perpendicular disposition to one of the hexagonal rings of the fullerene cage.

Scheme 16. Synthesis of steroid-fullerene hybrid 69.
Theoretical studies were considered on the new endohedral steroid-fullerene hybrid 69 to evaluate its biological potential. It was thus observed that it would have higher antiviral properties, in particular against Covid-19, than its hollow counterparts. Of course, these preliminary and encouraging results will have to be verified by future in vitro and in vivo tests.

Conclusions
The present review offers up-to-date literature on the syntheses of steroid-fullerene hybrids reported during the last years. The coupling of two or more natural products to make hybrids makes it possible to obtain new types of compounds with various structures, at the origin of improved properties compared to the starting molecules or at the origin of new properties. The biological aspect and all the properties of these molecules are still not well-known. But it is expected that the synthesis of fullerene derivatives will continue in the years to come as well as the still current interest in the chemistry of steroids and their derivatives given the applications (biological or in the chemistry of materials) inherent to this family of compounds.