Introduction

Many environments within the human body are acidic, such as the stomach, tumors, inflamed tissues, and acidic organelles (such as endolysosomes)1. This acidity has been used to provide pH-dependent drug release, to improve efficacy and the therapeutic index2,3,4. Recently, isopropenyl ethers (IPPEs) have emerged as critical intermediates to produce medicinally valuable asymmetric acyclic ketals that are highly acid-sensitive and derived from the metabolite acetone, and therefore are useful for the development of innovative stimulus-responsive drug delivery systems, including prodrugs for a large subset of hydroxyl-group-containing drugs5,6,7,8,9,10,11,12,13. Besides, IPPEs can be involved in many other transformations, including Diels-Alder reaction and construction of C(sp3)-rich α-tertiary dialkyl ethers14,15,16,17,18,19,20,21. Given that many small-molecule drugs (e.g., taxanes, sterols, prostaglandins, and nucleosides) contain a hydroxyl group, transisopropenylation of alcohols to create IPPEs can be expected to be useful in late-stage derivatization of these drugs. For instance, drug-derived IPPEs could be used to synthesize acyclic-ketal-linked prodrugs modularly for comprehensive optimization of prodrug performance. In the context, we hypothesized that a facile method for versatile synthesis of IPPEs with a broad substrate scope—including hydroxyl-group-containing drugs—would be highly desirable, with the goal of facilitating the development and application of medicinally valuable ketals.

Several methods are reported to synthesize IPPEs from alcohols or their derivatives: for example, acid-catalyzed pyrolysis of alkoxypropanes22,23,24; treatment of 2-methoxypropyl (MOP) ketal of sterically hindered alcohols with an excess of trimethylsilyl triflate (TMSOTf) and base25; olefination of acetates using Tebbe reagents26,27 and reaction of alcoholic potash with methylacetylene28. However, these reactions are usually carried out under harsh reaction conditions (e.g., high temperatures, use of strong bases), and some methods suffer from requiring expensive and toxic organometallic reagents and poor functional group compatibility (Fig. 1a). Numerous attempts, including transition-metal-catalyzed addition of alcohols to alkynes and cross-coupling reaction29,30,31,32,33,34,35,36,37, have been made to complement approaches mentioned above. Nonetheless, no methods—which are compatible with complex drugs (e.g., paclitaxel, PTX) and various functional groups (such as acrylate, ester, and maleimide) that are either present in some biodegradable polymers or useful for the development of ketal-based drug delivery systems—have appeared so far.

Fig. 1: Design of an organocatalytic transisopropenylation approach for the synthesis of IPPEs from alcohols.
figure 1

a Organocatalytic transisopropenylation enables the straightforward synthesis of IPPEs from hydroxyl-group-containing complex drugs (e.g., PTX), functional building blocks, and polymers under mild reaction conditions, complementing traditional approaches that are limited by the poor functional group compatibility and harsh reaction conditions. The resulting IPPEs will facilitate the synthesis of medicinally valuable ketals, such as modular ketal-linked prodrugs and biomaterials. b Organocatalytic transisopropenylation of alcohols, using 2-MPE with catalysis by Brønsted acid–base pair. Alcohol is transisopropenylated via the MOP ketal and transient oxocarbenium ion. c Optimization of the Brønsted acid–base pair organocatalyst. General conditions: 1a (1.0 mmol, 1.0 equiv., 0.5 M), 2a (16.0 equiv.), 0.5 mol % catalyst, THF, 25 °C, 1.0 h. The conversion rate of 1a and molar ratio of 3a to 3a′ were determined by NMR.

Full size image

Here, we report a straightforward transisopropenylation approach to the synthesis of IPPEs from alcohols and 2-methoxypropene (2-MPE), using a Brønsted acid–base pair as the organocatalyst (Fig. 1b). The method features mild reaction conditions, readily available and cheap starting materials and catalysts, easy workup, and a wide range of substrates; it is also compatible with complex small-molecule drugs, some polymers, and various functional groups and allows scalable synthesis. The transisopropenylation approach can facilitate the facile synthesis of various asymmetric ketal-linked prodrugs and functionalized ketalated biomaterials. An IPPE derivative of PTX is applied as the model IPPE to demonstrate the utility of drug-derived IPPEs for the modular preparation of acetone-based acyclic-ketal-linked prodrugs, including heterodimeric prodrugs containing both PTX and floxuridine (FUDR), two anticancer agents with different tumor-inhibiting mechanisms.

Results and discussion

Design rationale

Given that there are many small-molecule hydroxyl-group-containing drugs, a method for synthesizing IPPEs from alcohols using a mild catalytic transisopropenylation strategy might be useful. However, such catalytic transisopropenylation approaches have been scarcely reported38. Although there are several transition-metal-catalyzed methods for transvinylation of alcohols38,39,40,41,42,43, most of them do not apply to transisopropenylation of alcohols25,42,44. To the best of our knowledge, the only catalytic transisopropenylation of alcohols was achieved using isopropenyl acetate and [Ir(cod)Cl]2/Na2CO3 catalytic system38. Still, the method’s substrate scope is limited by harsh reaction conditions (high temperature).

We noted that resonance-stabilized oxocarbenium ion equilibrates with ketal under acidic conditions45, and that the rearrangement of oxocarbenium ion can proceed to afford IPPEs in the presence of weak bases25. Thus, we envisioned that a Brønsted acid–base pair could be used as an organocatalyst to not only catalyze the reaction between alcohol and the IPPE 2-MPE to form MOP ketal, but also catalyze the oxocarbenium ion in the equilibrium to undergo rearrangement to provide new IPPE. Therefore, a simple, mild, and catalytic transisopropenylation approach to syntheses of IPPEs could be developed (Fig. 1b).

Reaction development

We started by carrying out the reaction of alcohols and 2-MPE (2a) in THF using 2-octanol (1a) as the substrate (Fig. 1c and Supplementary Tables 1–6). According to the assumption, Brønsted acid can catalyze alcohol and 2a to give MOP ketal, so we evaluated Brønsted acids with different acidity firstly. An excess of cheap 2a (16 equiv.) was used to drive the equilibrium toward the oxocarbenium ion production. Surprisingly, we found that the reaction using p-toluenesulfonic acid (pTSA) produced 3a. However, it also produced a significant amount of side products that were dimer and oligomers of 2a (Supplementary Figs. 1–2), caused by acid-initiated cationic polymerization of 2a. Because base plays a vital role in promoting the oxocarbenium ion in the equilibrium to undergo rearrangement to afford IPPE, we selected pTSA as the Brønsted acid for acid–base pair and prepared several p-toluenesulfonate catalysts to evaluate bases. As a result, p-toluenesulfonate catalysts notably inhibited the side reactions, and the yield of 3a was markedly affected by the bases’ strength. Overall, 2,6-lutidinium p-toluenesulfonate (LPTS) gave the most satisfactory results.

Synthetic scope

With the optimized reaction condition, reactions of 2a with various drug substrates were then examined (Fig. 2; see details in Supplementary Information). We started by carrying out the reaction using PTX (a diol with potent anticancer activity). We found that the reaction produced a mixture of PTX-diMOP and PTX-2′-MOP-7-IPPE and that the 7-position but not the 2′-position of PTX was selectively transisopropenylated (Supplementary Fig. 3). As PTX could not be selectively converted to the diIPPE product, instead a mixture of MOP and IPPE, 2′-hydroxyl of PTX was first protected by TBS (tert-butyldimethylsilyl), and then 7-hydroxyl of PTX was selectively transisopropenylated. As a result, we successfully synthesized PTX-2′-TBS-7-IPPE (3b) from PTX-2′-TBS with a yield of 85%. To the best of our knowledge, this is the first report of drug-derived IPPEs. However, no PTX-derived IPPEs were yielded when we tried traditional approaches (such as [Ir(cod)Cl]2/Na2CO3 catalysis method and TMSOTf-promoted dealkoxylation) (Supplementary Figs. 13–16). Note that we ruled out Tebbe methylenation because Tebbe reagent is nonselective and would have decomposed the PTX.

Fig. 2: Synthetic scope of alcoholsa.
figure 2

aReaction conditions: 1 (1.0 equiv.), 2a (16.0 equiv.), 0.5 mol % LPTS, THF, 25 oC, 12 h. b20.0 mol % LPTS, 48 h. c2.0 mol % LPTS, 24 h. d5.0 mol % LPTS, 48 h. Yields of isolated products are given. ePurity of 3 is provided, as determined by 1H NMR. fLarge scale synthesis: 50 mmol 1, isolated yields by distillation.

Full size image

We further demonstrated the versatility of this method by synthesizing IPPEs from more drugs: the diols fulvestrant (FUL, a sterol), FUDR (a nucleoside analogue), and tafluprost (TAF, a prostaglandin), as well as the monoalcohols zidovudine (ZDV, an anti-HIV agent), lovastatin (LOV, a statin), idebenone (IDBN, an anti-Alzheimer agent), podophyllotoxin (PPT, an anticancer agent), testosterone (TEST, an anticancer agent), and abiraterone (ABI, an anticancer agent). The diols were first treated with 2a, and then the products were analyzed by means of NMR spectroscopy to ascertain the sites of IPPE and MOP ketal formation. As a result, for the anticancer agent FUL, the IPPE formed at the secondary hydroxyl group (17-OH), whereas a MOP ketal formed at the phenol group (3-OH) (Supplementary Fig. 4). Similar to PTX, FUL could be not be selectively converted to the diIPPE product. Therefore, TBS protection of FUL at 3-position was first done, and then FUL-3-TBS-17-IPPE (3c) was synthesized from FUL-3-TBS. For FUDR and TAF, both hydroxyl groups could be converted to IPPEs (Supplementary Figs. 5–6). For drugs with two IPPE conversion sites, they could be either mono TBS-protected or used as is for the preparation of drug-derived IPPEs. For example, an IPPE derived from the anticancer agent FUDR, designated FUDR-5′-TBS-3′-IPPE (3d), was synthesized from FUDR-5′-TBS. In contrast, either no reaction was observed or decomposition occurred when we attempted to use [Ir(cod)Cl]2/Na2CO3 catalysis method and TMSOTf-promoted dealkoxylation to prepare FUDR-derived IPPEs. Besides, the antiglaucoma drug TAF was directly converted to a diIPPE, designated TAF-3,5-diIPPE (3e). For the monoalcohols, ZDV, LOV, ABI, PPT, TEST, and IDBN were transformed to ZDV-IPPE (3f), LOV-IPPE (3g), ABI-IPPE (3h), PPT-IPPE (3i), TEST-IPPE (3j), and IDBN-IPPE (3k), respectively. We also explored the method’s applicability to additional drugs with tertiary alcohols and/or aromatic alcohols (such as 7-ethyl-10-hydroxycamptothecin, 7-hydroxycoumarin, and estradiol). However, we found that the hindered tertiary alcohol did not react at all due to steric hindrance (Supplementary Figs. 7–8) and that phenols gave only the corresponding MOP ketals (Supplementary Figs. 7–11). Nevertheless, the phenol of vitamin E with electron-donating groups afforded vitamin E-IPPE (3l). These results indicate that our method for the synthesis of drug-derived IPPEs from hydroxyl-group-containing drugs has a broad substrate scope, and it has the potential to be useful for the synthesis of IPPEs derived from analogues of these drugs.

Having studied the drug scope exclusively, we probed this strategy for alcohols widely used in pharmaceutical science and biomaterials. To this end, two kinds of hydroxyl-bearing polymers (1m, mPEG-OH, Mn = 132, 550, 2000 g/mol; 1n, mPEG-PCL-OH, Mn = 2000–1800 g/mol), cholesterol 1o, oleyl alcohol 1p, and menthol 1q were subjected to this protocol. To our delight, polymers were successfully converted to corresponding IPPEs (3m and 3n), enabling the potential synthesis of acid-sensitive macromolecular ketal-linked conjugates. Moreover, we probed this strategy for alcohols that are widely used as functional building blocks. Various functional groups, including ester, halogen, acrylate, alkenyl, alkynyl, amide, benzyl, epoxy, nitryl, cyano, and maleimide (Fig. 2 and Supplementary Table 7), were well tolerated. These functional IPPEs may have broad applications in post-modification of ketal-based materials5,46,47,48,49,50,51,52. Besides, the method was amenable to other primary, secondary, and tertiary alcohols (Supplementary Table 7), with good to high yields. The scalability of our approach was verified using 1a and 1q on 50 mmol scale, and large quantities of 3a and 3q (yield > 85%) were obtained by distillation.

Encouraged by the above results, we further probed whether other enol ethers were subjected to this method (Fig. 3), and we found various 2-methoxy-1-alkenes reacted with drug 1j to produce corresponding enol ethers (5a–f), which would allow the synthesis of ketals with adjustable hydrolysis kinetics. Likewise, 2-methoxy-1-alkenes also reacted with other alcohols (such as primary alcohols 1k and 1w, and secondary alcohol dehydroepiandrosterone) to afford corresponding enol ethers in good yields (Supplementary Table 8).

Fig. 3: Synthetic scope of enol ethersa.
figure 3

aReaction conditions: 1j (1.0 equiv.), 2 or 4 (16.0 equiv.), 20.0 mol % LPTS, THF, 25 oC, 24 h. Yields of isolated products are given.

Full size image

The monitoring reaction process of 2a with alcohol revealed that alcohol was firstly converted to MOP ketal, and then MOP ketal was transformed to new IPPE (Supplementary Fig. 17a). Besides, isolated MOP ketal was reacted with 2a or 2-ethoxypropene under similar conditions to yield the same IPPE (Supplementary Fig. 17b), further confirming the proposed mechanism of the method (Supplementary Fig. 18). Other enol ethers should also be formed according to similar mechanisms.

Synthesis of modular acetone-based ketal-linked prodrugs

The use of prodrugs—in which a drug is covalently but reversibly linked to another moiety and can be released in its active form—has emerged as an essential strategy for modifying drug performance3,53,54. Because many diseased tissues are characterized by low pH, acid sensitivity would be a desirable characteristic of such prodrugs55. Therefore, acyclic ketals, which are more acid-labile than analogous acetals and cyclic ketals, have great potential for the development of acid-activatable, traceless prodrugs of hydroxyl-group-containing drugs11,12,56,57. However, the synthesis of acyclic-ketal-linked prodrugs is markedly hindered by a lack of effective synthetic methods, and only a few acyclic-ketal-linked prodrugs have been reported5,7,8,9,49. Given that IPPEs could readily react with hydroxyl-group-containing drugs to afford acyclic-ketal-linked prodrugs and having established the organocatalytic transisopropenylation method for synthesis of IPPEs, we further explored the utility of this method for accessing divergent asymmetric ketal-linked prodrugs (Fig. 4 and see details in Supplementary Information). As proof-of-principle, hydroxyl-group-containing dexamethasone (DEX), a hydrophobic drug, was reacted with a PEG-derived IPPE 3m (Mn = 2000 g/mol) under catalysis by dichloroacetic acid (DCA) to obtain a water-soluble O,O-ketal-linked PEGylated prodrug, i.e., PEG-K-DEX (Fig. 4a). In addition to O,O-ketal-linked prodrug, an O,S-ketal-linked prodrug and an O,ON-ketal-linked prodrug, which both are rarely reported and may have different stimuli-responsiveness than O,O-ketals58,59,60, were also enabled by using a lauryl alcohol-derived IPPE under similar reaction conditions. Specifically, an O,S-ketal prodrug of thiol-group-containing captopril methyl ester (CAPME) and an O,ON-ketal prodrug of hydroxylamine-group-containing bufexamac (BUF) (Fig. 4b) were synthesized. Moreover, IPPE 3w was similarly reacted with hydroxyl-group-containing DEX to obtain an intermediate, which was subsequently conjugated to human albumin serum (HSA) through maleimide-thiol reaction to yield a water-soluble O,O-ketal-linked protein prodrug conjugate, i.e., HSA-K-DEX (Fig. 4b and Supplementary Fig. 19).

Fig. 4: Ketal-linked prodrugs.
figure 4

a synthesis of PEG-K-DEX; b structures of LA-K-CAPME, LA-K-BUF, and HSA-K-DEX; c synthesis of PK5F; d structures of PTX-7-K-EG3, PK3F, and PK5E.

Full size image

To demonstrate the utility of drug-derived IPPEs for the modular preparation of ketal-linked prodrugs, we further synthesized four PTX prodrugs, designated PK5F, PTX-7-K-EG3, PK3F, and PK5E (Fig. 4c, d). As shown in Fig. 4c, PK5F was synthesized using PPTS-catalyzed reaction of PTX-2′-TBS-7-IPPE 3b with FUDR-3′-TBS and subsequent removal of the TBS group with tetra-n-butylammonium fluoride (TBAF). Similarly, PTX-7-K-EG3, PK3F, and PK5E were synthesized using 3b and the corresponding alcohols (HO-EG3, FUDR-5′-TBS, and EdU-3′-TBS) (Fig. 4d). Please note that conditions for the TBS deprotection were mild and compatible with the ketals, and no side products were observed in the last step in the synthesis of PTX-derived prodrugs. In addition, although ketal exchange reactions may occur and yield symmetric homodimers in synthesizing asymmetric ketal-based prodrugs, we did not note the formation of homodimers using our method. For PTX-7-K-EG3 with a short PEG chain, PEGylation of PTX via an acid-sensitive acetone-based ketal, which would be traceless, might address the solubility problem of PTX without causing side effects associated with Cremophor EL—the excipient of a commercial PTX formulation (Taxol). For PK3F and PK5F, the drug FUDR was used to synthesize heterodimeric prodrugs to improve antitumor efficacy, and they might show different hydrolysis kinetics. For PK5E, 5-ethynyl-2′-deoxyuridine (EdU), a nucleoside analogue that is widely used for detecting live cells via a copper-catalyzed fluorogenic click reaction, can be used to investigate the intracellular hydrolysis of the heterodimers.

In vitro and in vivo characterizations of PTX prodrugs

Hydrolysis experiments indicated that hydrolysis of PTX prodrugs yielded native PTX (Supplementary Figs. 20–21), thus confirming that the acetone-based ketal in the prodrugs acted as a traceless linker. At pH 5.0 (lysosomal pH), the t1/2 values (calculated from kinetic constant values; Supplementary Table 9) of PTX-7-K-EG3, PK3F, PK5F, and PK5E were 0.78, 3.41, 3.16, and 6.54 h (Fig. 5a and Supplementary Fig. 22a), respectively; whereas, at pH 7.4 (physiological pH), the corresponding values exceeded 200 h (Fig. 5b and Supplementary Fig. 22b). Note that the t1/2 of PK5F did not significantly differ from that of PK3F; although the fluorine atom is far from the ketal bond in the PKFs (i.e., PK3F and PK5F), the rate of hydrolysis of PK5E, which has an alkynyl group instead of a fluorine atom, was lower than that of PKFs. In addition, the t1/2 values of these heterodimers were larger than the values of PTX-7-K-EG3, probably because of the greater steric bulk of the promoieties in the former. The difference between the t1/2 values of the same prodrug at the two pHs was not only related to the difference between the two pH values (ΔpH), probably because ketal hydrolysis proceeds by general acid catalysis, the rate of which can be affected by multiple factors (e.g., the composition of the buffer)57,61,62.

Fig. 5: In vitro and in vivo characterizations of PTX prodrugs.
figure 5

Hydrolysis of PTX prodrugs at (a) pH 5.0 and (b) pH 7.4. Data of (a) and (b) are means ± SD (n = 4 independent samples). c Representative transmission electron microscope image of PK5F-loaded micelles; scale bar = 50 nm. Experiments were performed three times independently, and one representative image is shown. d Sizes of heterodimer-loaded micelles, determined by dynamic light scattering. e IC50 values of drugs against HCT116 cells and 3T3 cells. f The extent of intracellular hydrolysis (%) of heterodimers after incubation of heterodimer-loaded micelles with HCT116 cells and 3T3 cells at 20 μM for 6 h. Data of (e) and (f) are means ± SD (n = 3 independent experiments), and unpaired two-tailed Student’s t-test was used for statistical analysis. g Confocal microscopy images of intracellular hydrolysis of PK5E. Untreated cells and cells incubated with EdU, PTX, and micelle-encapsulated PK5E at 5 µM for 6 h at 37 °C. Nuclei were stained with Hoechst 33342 (blue). α-Tubulin was stained with α-tubulin (11H10) Rabbit mAb (green). EdU was stained using Alexa Fluor 555 azide via click chemistry (pink). Scale bar = 20 μm. Experiments were performed three times independently, and representative images are shown. h Evaluation of in vivo efficacy of FUDR, PTX-loaded micelles, PTX-loaded micelles plus FUDR, and PK5F-loaded micelles against HCT116 xenograft tumors. Temporal dependence of tumor volume after intravenous injection of the four formulations on days 0, 3, 6, and 9 at the following doses: FUDR, 11.7 μmol/kg; PTX-loaded micelles, 11.7  μmol/kg; PTX-loaded micelles plus FUDR, 11.7 μmol/kg for each drug; PK5F-loaded micelles, 11.7  μmol/kg. Data are means ± SD (n = 6 biologically independent animals). i Accumulation (% injected dose/gram of tumor) of drugs at 4, 12, and 24 h after a single intravenous injection. Free PTX indicates the liberated PTX from PK5F. Data are means ± SD (n = 4 biologically independent animals). Unpaired two-tailed Student’s t-test was used for statistical analysis in (h) and (i). Source data of (a), (b), (e), (f), (h) and (i) are provided as a Source Data file.

Full size image

The use of PKFs for co-delivery and simultaneous release of PTX and FUDR—which have different tumor-inhibiting mechanisms and non-overlapping adverse effects—might result in improved anticancer efficacy than either agent alone63. We were unable to prepare excipient-free self-assembled prodrug nanoparticles, perhaps because paclitaxel is too hydrophobic. Thus, we encapsulated the PK5F, PK3F, and PK5E, which are hydrophobic and poorly water-soluble, in methoxy polyethylene glycol-poly(D,L-lactide) (mPEG-PDLLA, Mn = 5000–3000 g/mol) micelles (about 20 nm, drug loading: ~8.9 wt%, loading efficiency: ~74%) (Fig. 5c, d, Supplementary Table 10, and Supplementary Fig. 23), for following biological evaluations. The IC50 values of micelle-encapsulated PK3F and PK5F in a human colon cancer cell line HCT116 were lower than that of PTX and FUDR, and were 159 and 90.4 nM, respectively (Fig. 5e and Supplementary Fig. 24). Even though the t1/2 values for free PK5F and PK3F hydrolysis did not differ significantly, micelle-encapsulated PK5F exhibited higher toxicity than micelle-encapsulated PK3F, indicating that other processes, such as enzymatic hydrolysis, may have been involved, to different extents, in intracellular hydrolysis of the heterodimers. The degree of hydrolysis of micelle-encapsulated PK5F in HCT116 cells was 68% at 6 h, and the value for micelle-encapsulated PK3F was slightly lower (Fig. 5f and Supplementary Fig. 25). Intracellular hydrolysis of heterodimers was also examined using confocal microscopy using PK5E (Fig. 5g). Cells treated with micelle-encapsulated PK5E showed notable EdU staining and microtubule bundles, confirming that the extent of intracellular hydrolysis of micelle-encapsulated PK5E was high and that the released EdU was incorporated into the DNA of the proliferating cells. However, micelle-encapsulated PKFs displayed lower toxicities than PTX and FUDR in the healthy 3T3 cells (Fig. 5e), and the extent of hydrolysis of micelle-encapsulated PKFs was much lower in the 3T3 cells than in the HCT116 cells (Fig. 5f), perhaps because the two types of cells have different proliferation and metabolism modes. Our findings indicate that the heterodimers were activated more efficiently by tumor cells than by healthy cells.

In vivo anticancer efficacy study in an HCT116 xenograft model revealed that PK5F-loaded micelles, which had a lower IC50 than PK3F-loaded micelles, exhibited better efficacy than PTX-loaded micelles plus FUDR (Fig. 5h and Supplementary Figs. 27–30), probably because PK5F-loaded micelles improved co-delivery of PTX and FUDR to the tumor. Although the carrier vehicle is the same for PTX and PK5F, PK5F showed higher accumulation in the tumor than did PTX at all time points (Fig. 5i). We speculate that different accumulation values of drugs may be ascribed to different compatibility between drug and mPEG-PDLLA, which influences complex interactions between biological components, drugs, and micelles64. The extent of PK5F hydrolysis in tumors was also determined, and the degree of hydrolysis had increased from 55% at 4 h to 82% at 24 h after intravenous injection, indicating that intratumoral hydrolysis of the heterodimer was efficient (Supplementary Fig. 26). We note that the hydrolysis rate of the prodrug in the tumor tissue was slower than the in vitro intracellular hydrolysis rate, possibly because the tumor’s microenvironment is less acidic than endolysosomes1. Moreover, no significant loss of body weight was observed for any treatments (Supplementary Fig. 30). These results indicate that loading heterodimeric prodrugs into polymeric micelles can enable co-delivery of two drugs with different physicochemical properties and is a promising strategy for improved cancer therapy.

Synthesis of ketal-linked biomaterials

In addition to prodrugs, IPPEs could have tremendous utility in the development of novel biomaterials. For example, Fréchet et al. have modified dextran, a highly water-soluble natural biopolymer, as pH-sensitive and hydrophobic ketalated dextran using 2-MPE (2a, the simplest IPPE), and demonstrated their vast potential in biomaterials and drug delivery11,12,65,66,67,68. Nevertheless, the lack of diversity and availability of IPPEs limits the widespread applications of such ketalated dextran. To demonstrate the utility of IPPEs in creating functionalized ketalated dextran, IPPEs 3t and 3r were reacted with dextran (Mw = 10,000 g/mol) to afford acrylate- and alkyne-functionalized ketalated dextran derivatives (Fig. 6 and see details in Supplementary Information), respectively. Ketalated dextran with mixed ketals of acyclic ketals and cyclic ketals was yielded because acyclic ketals were partially reacted with its neighboring hydroxyls to form more stable cyclic ketals through a ketal exchange reaction. However, the hydroxyl substitution degree, the ratio of acyclic ketals to cyclic ketals, and hydrolysis kinetics of ketalated dextran could be well-tuned by adjusting the reaction conditions (e.g., reaction time)11. The functionalized ketalated dextran derivatives would allow versatile modification on demand by Michael addition reaction or click reaction. In addition, the drug-derived IPPE 3k was reacted with dextran to obtain a pH-sensitive dextran-drug conjugate.

Fig. 6: Ketal-linked biomaterials.
figure 6

Synthesis of functionalized ketalated dextran, using addition reaction of IPPEs to hydroxyls of dextran under catalysis by pyridinium p-toluenesulfonate (PPTS). DS denotes the hydroxyl substitution degree; AK denotes the acyclic ketals; CK denotes the cyclic ketals.

Full size image

In summary, we have developed an organocatalytic transisopropenylation approach to synthesize IPPEs that are difficult to access so far, using Brønsted acid–base pair as the organocatalyst. To the best of our knowledge, we accomplished the first syntheses of drug-derived IPPEs. We used an IPPE derivative of PTX to demonstrate the late-stage derivatization of drug-derived IPPEs for implementing the synthesis of modular ketal-linked prodrugs, enabling the development of polymeric micelles loaded with a heterodimeric PTX/FUDR prodrug—which showed higher potent anticancer efficacy than PTX. The prodrug strategy reported herein opens up an avenue for the development of modular prodrugs from which active drugs can be released tracelessly upon exposure to acid. In addition, functionalized ketalated dextran derivatives were also readily obtained using the functional group-containing IPPEs. The straightforward synthesis and late-stage derivatization of IPPEs will facilitate the development and practical application of acid-sensitive prodrugs, drug delivery systems, and biomaterials.

Methods

Syntheses

All details regarding the synthesis and characterization of IPPEs, relevant intermediates, prodrugs, and ketalated dextran derivatives are provided in the Supplementary Information.

Hydrolysis of prodrugs determined using HPLC

To investigate the hydrolysis kinetics, prodrugs were dissolved in ACN and diluted 19 times using buffers (30 mM, pH 5.0 acetate buffer, pH 7.4 phosphate buffer) to final concentration of 20 μM in 200 μL buffers of each sample. Hydrolysis experiments were performed in a constant temperature shaker (Zhicheng Inc. CHN) at 200 rpm and 37 oC (n = 4). At predetermined intervals, samples were added with 200 µL PB solution (200 mM, pH 8.0) and 400 µL ACN immediately and stored at −20 °C before measurement. The concentration of PTX was analyzed by HPLC. HPLC analyses were performed using Agilent 1260 instruments (Agilent, USA) with C18 column (Agilent, Poroshell 120 EC-C18, 4.6 mm × 100 mm, 2.7 µm). For PTX-7-K-EG3, samples were injected at a flow rate of 1.0 mL/min, column temperature at 25 °C, and UV detection wavelength at 227 nm. An aqueous solution of ACN/H2O (65/35) was used as the mobile phase. For dimeric prodrugs, samples were injected at a flow rate of 1.0 mL/min, column temperature at 25 °C, and UV detection wavelength at 227 nm. HPLC eluent condition: 20% [v/v] ACN, 1.0 min 20% [v/v] ACN, 4.0 min 80% [v/v] ACN, 7.0 min 80% [v/v] ACN, 7.5 min 20% [v/v] ACN, 10 min 20% [v/v] ACN.

The hydrolysis (%) of prodrug was calculated by the following formula:

$${{\mbox{Hydrolysis (\%) =}}}\frac{{C}_{n}}{{C}_{{{\mbox{total}}}}}{{\times }}100 \%$$
(1)

where Cn represents the concentration of PTX in the nth sample and Ctotal represents concentration of PTX after complete hydrolysis of prodrug at pH 3.0.

The data were fitted with “Nonlinear regression (curve fit)-One phase decay” model using GraphPad Prism 7 (GraphPad Software, CA), and the rate constant was determined by a “Linear regression” model of ln(100−100 × Cn/\(C_{\mbox{total}}\)) vs time. t1/2 = ln2/k.

Preparation and characterization of drug-loaded micelles

Nanoprecipitation technique was utilized to load drugs into mPEG-PDLLA micelles. Briefly, mPEG-PDLLA (10 mg) and drug (1 mg) was dissolved in 0.5 mL acetone. The solution was slowly dropped into 10 mL phosphate buffer (10 mM, pH 7.4) under stirring (500 rpm). The residual solvent was removed by evaporation at room temperature for 2 h. After that, drug-loaded micelles were filtrated with 0.45 μm membrane filter, lyophilized to dryness, and stored in the freezer before use.

The yield of drug-loaded micelles was weighed. The hydrodynamic diameter, polydispersity index, and zeta potential of drug-loaded micelles were analyzed by DLS. Sizes of drug-loaded micelles were also observed by TEM (Talos F200C). The drug loading and loading efficiency of drug-loaded micelles were determined by HPLC. An aqueous solution of ACN/H2O (65/35) was used as the mobile phase. Drugs were detected by UV absorbance at λ = 227 nm and quantified using corresponding standard curves. The stability of drug-loaded micelles was studied using DLS and HPLC. The results showed that drug-loaded micelles were stable in the aqueous solution without prodrug hydrolysis and notable size changes in a week. In addition, the lyophilized drug-loaded micelles could be easily re-dispersed.

The drug loading (DL) and loading efficiency (LE) of drugs in micelles were calculated by the following formulas:

$${{\mbox{DL}}}\,{{\mbox{(wt\%)}}}\,=\,\frac{{{\mbox{actual}}}\; {{\mbox{measured}}}\; {{\mbox{mass}}}\; {{\mbox{of}}}\; {{\mbox{drug}}}}{{{\mbox{mass}}}\; {{\mbox{of}}}\; {{\mbox{drug}}}\; {{\mbox{loaded}}}\; {{\mbox{micelles}}}}{{\times }}100{{{{{\rm{wt}}}}}} \%$$
(2)
$${{\mbox{LE (\%)}}}\,=\,\frac{{{\mbox{actual}}}\; {{\mbox{measured}}}\; {{\mbox{loading}}}\; {{\mbox{of}}}\; {{\mbox{drug}}}\; {{\mbox{in}}}\; {{\mbox{drug}}}\; {{\mbox{loaded}}}\; {{\mbox{micelles}}}}{{{\mbox{theoretical}}}\; {{\mbox{(feed)}}}\; {{\mbox{loading}}}\; {{\mbox{of}}}\; {{\mbox{drug}}}\; {{\mbox{in}}}\; {{\mbox{drug}}}\; {{\mbox{loaded}}}\; {{\mbox{micelles}}}}{{\times }}100 \%$$
(3)

Cell cytotoxicity assay

Standard Cell Counting Kit-8 (CCK-8) assay was followed to evaluate the cytotoxicity of free drugs and micellar formulations or combinations (FUDR:PTX = 1:1, mol/mol) in HCT116 cells and 3T3 cells. Cells were seeded at a seeding density of 2500 cells/well in a 96 well plate, and cultured in 100 µL DMEM (or RPMI 1640 medium) containing 10% fetal bovine serum (FBS), 100 IU/mL penicillin, and 100 µg/mL streptomycin in 5% CO2 incubator at 37 °C for 24 h (n = 3). PTX was dissolved in DMSO, whereas, FUDR and drug-loaded micelles were in PBS (6.7 mM, pH 7.4). The medium in each well was replaced by 100 µL fresh medium containing drug. For HCT116 cells, drugs were ranging from 1 nM to 5 μM; for 3T3 cells, drugs were ranging from 10 nM to 20 μM. The cells were further incubated at 37 °C for 72 h. The medium in each well was aspirated, and 100 µL fresh medium plus 10 µL CCK-8 solution was added, followed by incubation at 37 °C for another 1.5 h. Absorbance was measured by a SpectraMax i3x (Molecular Devices, CA) at 450 nm. The data were fitted with “Dose-Response – Inhibition” model using GraphPad Prism 7 (GraphPad Software, CA), and the IC50 values were reported using equation of “log(inhibitor) vs. response—Variable slope (four parameters)”. Experiments were performed in triplicate.

Cellular uptake and intracellular hydrolysis studies

HCT116 cells or 3T3 cells were seeded at a density of 2 × 105 cells in a six-well plate with DMEM medium containing 10% fetal bovine serum (FBS), 100 IU/mL penicillin, and 100 µg/mL streptomycin. Cells were grown in a 37 °C incubator with 5% CO2 overnight to allow to attach (n = 3). Drugs with an equivalent concentration of 20 μM PTX were prepared as procedures above. The cell medium was replaced by 2 mL drug-containing medium. After 6 h incubation at 37 °C, the medium was aspirated and the cells were washed twice with cold PBS. Cells were de-attached from the six-well plate by trypsinization at room temperature. The cell pellets were collected by centrifugation at 150 × g for 5 min. The cell pallets were resuspended by a mixture of ACN/H2O (1:1) and homogenized by sonication (VC 505, Vibracell Sonics, Newton, USA). Then, 500 μL of cell homogenates were deproteinized by the addition of three volumes of cold ACN which contains 25 μL docetaxel solution (4 μg/mL) as internal standard (IS). The mixture was centrifugated at 19,800 × g for 10 min to remove precipitates. Three mL methyl tert-butyl ether was added to the supernatant to extract drugs, followed by vortex and centrifugation. The organic layer was aspirated and dried by an Eppendorf™ Concentrator Plus vacuum concentrator (Eppendorf, GER). The concentrations of PTX and dimers were analyzed by HPLC as mentioned above.

To investigate intracellular hydrolysis and distribution of PK5E, HCT116 cells were seeded onto 13-mm coverslips (5 × 104 cells per coverslip) in 24-well plate with DMEM medium containing 10% fetal bovine serum (FBS), 100 IU/mL penicillin and 100 µg/mL streptomycin. Cells were grown in a 37 °C incubator with 5% CO2 overnight to allow to attach. The cell medium was replaced by 500 μL drug-containing medium (5 μM). After 6 h incubation at 37  °C, the medium was removed, and the cells were washed twice with PBS and fixed with 4% paraformaldehyde solution. Following standard protocols, cells were first stained using BeyoClick™ EdU-555 Kit (Beyotime Biotechnology, CHN) and then stained with α-Tubulin (11H10) Rabbit mAb (Alexa Fluor® 488 Conjugate) (Cell Signaling Technology, USA). The nucleus was stained with Hoechst 33342 solution (Beyotime Biotechnology, CHN). Slides were sealed using Vectashield mounting medium (Vector Laboratories Burlingame, CA, USA) and observed using CLSM (Leica Microsystems, Wetzlar, GER).

In vivo antitumor evaluation

All experiments were carried out following the guidelines of the Beijing Laboratory Animal Center, and approved by the Ethical Commission at Nankai University. Female BALB/c nude mice (18–20 g) were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. and maintained in a 12 h light-dark cycle at 25 °C, 40% relative humidity with free access to food and water. Five million HCT116 cells in 100 µL DMEM medium were subcutaneously injected into the right flank of mice. After the tumor volume reached ~100 mm3, mice were randomly divided into five groups (n = 6). HCT116 tumor-bearing mice were intravenously injected with PBS, FUDR, PTX-loaded micelles, PTX-loaded micelles plus FUDR and PK5F-loaded micelles at a dose of 11.7 μmol/kg at day 0, 3, 6, and 9. The weight and tumor volume of each mice were measured every three days. The volume of the tumor was calculated using the formula: V = (length ×  width2)/2. Three weeks post injection, mice were sacrificed. The blood was drawn for blood count test and serum biochemical analysis. Major organs and tumors were collected, followed by fixed in 4% paraformaldehyde solution and embedded in paraffin. Organ tissues were sectioned for H&E staining, and tumor tissues were sectioned for H&E staining, Ki67 immunohistochemical staining, and TUNEL assay.

HCT116 tumor-bearing nude mice (n = 4) were received single injection of PTX-loaded micelles and PK5F-loaded micelles at the dose of 11.7 μmol/kg respectively. At predetermined time intervals (4, 12, and 24 h), mice were sacrificed and their tumors were dissected to measure accumulation of PTX and PK5F and hydrolysis degree of PK5F by HPLC.

Statistic analysis

Data were reported as means with standard derivations. Statistical analysis was conducted by Student’s t-test using GraphPad Prism version 7 for Windows (GraphPad Software, CA). Differences were considered statistically significant if p < 0.05.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.