Advanced Pyrrolidine-Carbamate Self-Immolative Spacer with Tertiary Amine Handle Induces Superfast Cyclative Drug Release.

Amine-carbamate self-immolative (SI) spacers represent practical and versatile tools in targeted prodrugs, but their slow degradation mechanism limits drug activation at the site of disease. We engineered a pyrrolidine-carbamate SI spacer with a tertiary amine handle which strongly accelerates the spacer cyclization to give a bicyclic urea and the free hydroxy groups of either cytotoxic (Camptothecin) or immunostimulatory (Resiquimod) drugs. In silico conformational analysis and pKa calculations suggest a plausible mechanism for the superior efficacy of the advanced SI spacer compared to state-of-art analogues.

Introduction

The selective delivery of drugs to the site of disease represents a widely pursued research goal, aimed at improving the efficacy and tolerability of pharmacological interventions. In this context, the covalent conjugation of pharmaceutical ingredients to antibodies (resulting in Antibody‐Drug Conjugates, ADCs) represents one of the most validated technologies. Historically, ADCs have been developed to release cytotoxic agents at the tumour site and kill cancer cells selectively, sparing healthy tissues. More recently, the ADC technology was adapted to various classes of pharmaceutical agents, including antibacterial, anti‐inflammatory, and pro‐inflammatory drugs. In addition to antibodies, the covalent drug conjugation to different carriers (e. g. albumin, peptides, small ligands, polymers, etc.) is pushing the boundaries of targeted medicine. In most of these constructs, therapeutic effects are only displayed when the drug is effectively disconnected from the carrier. In this mechanism of action, a key role is played by the so‐called self‐immolative (SI) spacers, i. e. synthetic devices designed to undergo spontaneous disassembly in response to specific stimuli. In particular, different types of activation signals (see “Trigger Activation” in Figure 1) typically lead to the liberation of a reactive functional group in the SI spacer. The latter initiates a variety of intramolecular reactions (mainly electronic cascade in aromatic and π‐extended systems or cyclization of nucleophilic groups) that terminate with the release of thermodynamically‐stable end‐products (see “SI spacer Degradation” in Figure 1). In the prodrug context, not only SI spacers facilitate the drug disconnection from the carrier and its release in a pharmaceutically‐active form, but they also act as chemical adaptors for the whole prodrug assembly. Several conjugation strategies have been proposed to functionalize a variety of bioactive molecules, including amines, phenols, primary and secondary alcohols. In particular, the use of carbamate bonds is a practical and versatile strategy to connect a variety of hydroxy payloads to specific triggers and the ethylenediamine‐carbamate spacer (Sp1, Figure 1) has long represented a standard in drug conjugates and other stimuli‐responsive systems.

Figure 1

Schematic drug release mechanism of a generic covalent drug conjugate, consisting in the initial trigger activation and subsequent degradation of a self‐immolative (SI) spacer. The molecular structure and degradation mechanism of three cyclizing amine‐carbamate SI spacers (i. e. benchmark ethylenediamine‐carbamate Sp1, pyrrolidine‐carbamate Sp2 and its engineered derivative Sp3, described in this work) is shown: secondary amine cyclization leads to carbamate cleavage, to give the drug's free hydroxy group (1) and cyclic urea 2 (mono‐cyclic urea for Sp1, bicyclic urea for Sp2/3; R=Me for Sp1/2, R=(CH2)2NMe2 for Sp3). SI spacer cyclization rates increase in the order Sp1

Our group has recently investigated the structural modification of Sp1. In 2020, we reported that the pyrrolidine‐carbamate SI spacer Sp2 (Figure 1) undergoes cyclative cleavage and releases OH‐bearing drugs (1) and cyclic urea 2 at higher rates than Sp1. The incorporation of Sp1 and Sp2 into protease‐activable anticancer prodrugs provided experimental evidence that a fast SI spacer degradation augments the anticancer effects in vitro. In light of these data, we speculated that a rapid SI spacer degradation may be particularly important for the prodrug efficacy in vivo. Indeed, even if the trigger activation (Figure 1) occurs efficiently and selectively at the site of interest, a long timespan between this initial stimulus and the final drug release prolongs the drug survival in a modified and inactive form. This delayed activation would facilitate the drug migration from the site of disease, its excretion or, at worst, its activation in healthy organs.

We describe herein an advanced SI spacer (Sp3 in Figure 1) in which a tertiary amine handle enables much faster cyclative drug release compared to previously described spacers of the same class. Thanks to this rapid carbamate cleavage, this accelerated spacer degradation holds promises for improved delivery of therapeutic agents.

Results and Discussion

We undertook the structural optimization of the pyrrolidine‐carbamate SI spacer Sp2, aimed at decreasing the carbamate half‐life (t 1/2) for a faster hydroxy cargo release. The molecular structures of the synthesized SI spacers are reported in Figure 2 (all synthetic procedures are reported in the Supporting Information). Here, the Sp2 pyrrolidine ring was replaced by either a piperidine (in Sp4) or an isoxazolidine (in Sp5) cycle. In particular, Sp4 was designed to assess the impact of the cyclic amine ring size on the SI spacer degradation. On the other hand, the use of an isoxazolidine ring in Sp5 aimed at evaluating the contribution of the “α‐effect” to the spacer reactivity. While Sp4 was prepared starting from racemic 2‐piperidinecarbaldehyde, the synthesis of Sp5 was inspired by Bode's route to 5‐oxaproline. Moreover, as we recently observed the inhibitory effects on the spacer cyclization given by a phosphate monoester group, we devised the SI spacer modification with a basic handle. Since at physiological pH (approximately 7.4) the pyrrolidine is mostly present as protonated species, a second amine handle may force the pyrrolidine deprotonation, thus lowering its pK a value and increasing its nucleophilic character. To this end, SI spacers Sp3 and Sp6/8 were endowed with a tertiary amine handle connected either at the carbamate N atom (in Sp3 and Sp7) or at the pyrrolidine ring (in Sp6 and Sp8). While the former spacers were rapidly obtained by reductive amination of Boc‐l‐prolinal, the latter were prepared by multistep synthesis, starting from oxygenated proline derivatives, such as 4‐oxo‐l‐proline and hydroxyproline. Firstly, the exocyclic secondary amine of SI spacers Sp3/8 was conjugated to the tertiary hydroxy group of the anticancer drug Camptothecin (CPT) via carbamate bond. As described recently,[ , ] the final spacer‐drug modules Sp3/8‐CPT (Figure 2) were isolated as trifluoroacetate salts. These compounds were dissolved in a DMSO/phosphate buffer (pH 7.5) mixture and incubated at 37 °C, followed by aliquot collection at different time points and CPT release analysis by HPLC. The percentage of intact carbamate calculated from peak integrals was plotted versus time, and the spacer cyclization rates were estimated in terms of carbamate half‐life (t 1/2). The results of this first screening are summarized in Figure 2.

Figure 2

Molecular structure and drug release activity of II‐generation cyclizing SI spacers (Sp3/8), connected to the tertiary hydroxy group of Camptothecin (CPT). Experimental procedures for the SI spacer‐CPT module synthesis and release studies are included in the Supporting Information.

Notably, replacement of the native pyrrolidine ring with a piperidine (Sp4) and an isoxazolidine (Sp5) led to a complete inhibition of the cyclative carbamate cleavage. On the other hand, the SI spacer derivatization with a tertiary amine led to some interesting results in terms of CPT release efficacy. In particular, carbamate Sp6‐CPT released CPT with lower rates than reference Sp2‐CPT (t 1/2=5.5 and 3.6 h, respectively), while the reactivity of SI spacers Sp7 (t 1/2=3.3 h) and Sp8 (t 1/2=3.2 h) was slightly improved, but very similar to the one of native Sp2. To our delight, carbamate Sp3‐CPT showed a half‐life of 0.9 h, i. e. four times shorter than reference compound Sp2‐CPT. All in all, these data indicate that the cyclative cleavage of pyrrolidine‐carbamate SI spacers can be accelerated by a tertiary amine handle, provided that the latter is in close proximity to the electrophilic carbamate bond. Indeed, the tertiary amine installation in a remote position from the carbamate (e. g. in Sp6 and Sp8) showed no significant impact on the CPT release rates compared to reference Sp2. The importance of the amine handle proximity is even more evident by considering the better CPT release performance of Sp3 (where a C‐2 alkyl chain connects the carbamate N atom to the tertiary amino group) compared to Sp7 (bearing a C‐3 alkyl chain).

With the aim at confirming the superior efficacy of the new SI spacer Sp3, we investigated its ability to release a different payload, namely the immunostimulatory drug Resiquimod (R848). This imidazoquinoline (IMD) is a potent agonist of Toll‐like receptors (TLR) 7 and 8, intracellular proteins expressed by several types of immune cells and involved in the host defence from viral infections.

By mimicking single‐strand RNA fragments, IMDs induce TLR homodimerization and activate downstream pro‐inflammatory pathways. Due to these pharmaceutical effects, tumour‐targeted IMD prodrugs are being increasingly investigated to selectively activate the immune system at the site of disease, thus improving the outcomes of immunotherapy regimens.[ , ] As previously done with CPT, SI spacers Sp1, Sp2 and Sp3 were connected to the tertiary hydroxy group of R848 via carbamate bond. The resulting Sp1/3‐R848 adducts were dissolved in a DMSO/phosphate buffer (pH 7.5) mixture and incubated at 37 °C, following drug release analysis as described above. As shown by the HPLC traces and the chart in Figure 3, R848 carbamates proved generally more stable than the analogous CPT constructs. In particular, the native pyrrolidine spacer Sp2 showed a very slow carbamate cleavage, with a half‐life of 40 h calculated for the Sp2‐R848 construct. As expected, the drug release activity shown by Sp2 was superior than the benchmark ethylenediamine‐carbamate spacer Sp1, which released only traces of drug after eight‐hour incubation. As observed with the CPT adducts, the tertiary amine‐bearing spacer Sp3 showed the highest drug release activity of the series, as the half‐life of Sp3‐R848 carbamate (t 1/2=7.6 h) resulted approximately five times shorter than that of the Sp2‐bearing analogue. In this experiment, LC‐MS analysis of the Sp3‐R848 adduct upon eight‐hour incubation confirmed the formation of bicyclic urea 2 during Sp3 degradation (see Figure S1 in the Supporting Information). These CPT and R848 release data confirmed the superior performance of the advanced Sp3 spacer compared to both Sp2 and, even more dramatically, Sp1 references.

Figure 3

Drug release activity of SI spacers Sp1/3, connected to the tertiary hydroxy group of Resiquimod (R848). HPLC traces relative to the stability analysis of carbamates Sp1/3‐R848 at t=0, 8 h are shown (peak of free R848 is highlighted in red) together with stability curves. r.t.: retention time, Abs.: UV absorbance. Experimental procedures for the SI spacer‐R848 carbamate synthesis and release studies are included in the Supporting Information.

To qualitatively investigate the structural basis for the Sp3 spacer exceptional reactivity, we performed conformational analyses by computational methods of both Sp2 and Sp3 structures, connected to a generic alcohol (tert‐butanol, “tBu”) through a carbamate bond. In particular, Monte Carlo/Energy Minimization (MC/EM) conformational searches were performed at the molecular mechanics level (OPLS3 force field) on carbamates Sp2‐ and Sp3‐ in their main ionization state at pH 7.5, corresponding to positively‐charged amino groups (i. e. protonated pyrrolidine in Sp2‐, protonated pyrrolidine and tertiary amine in Sp3‐). Representative minimum‐energy conformations selected from the molecular mechanics calculations were optimized at the DFT B3LYP/6‐31G* level of theory. Solution phase energies of the obtained stationary points were computed at the same level of theory by single‐point energy calculations including the water/PBF solvent model. Finally, pK a values of the protonated species were calculated on DFT minimum energy structures displaying the lowest solution phase energies. According to these calculations, carbamate Sp2‐ adopts a preferential conformation (referred to as “Sp2‐tBu anti” in Figure 4A) in which the pyrrolidinium ion engages the carbonyl sp2‐hybridized O atom in a hydrogen bond, forming a seven‐membered ring. On the other hand, a similar intramolecular interaction of the pyrrolidinium ion with sp3‐hybridized O atom leads to the “Sp2‐tBu syn” conformation, 2.74 kcal/mol higher in energy than the “anti” counterpart.

Figure 4

A) Molecular structures of representative conformations for Sp2/3‐, optimized at the DFT B3LYP/6‐31G* level. Relative energy differences are calculated from the corresponding solution phase energies (water PBF). Calculated pK a values for N‐H+ protons are reported (orange). Additional conformations of Sp2/3‐ are shown in Figure S2 in the Supporting Information). B) Plausible mechanism of carbamate cyclative cleavage carried out by SI spacer Sp3.

“Anti” rotamers in carbamates are typically favored over the “syn” counterparts by 1.0–1.5 kcal/mol, and the 2.74 kcal/mol increased stability of “Sp2‐tBu anti” versus “Sp2‐tBu syn” indicates that rotational equilibrium of amine‐carbamate modules can be dramatically influenced by intramolecular H bonding. Concerning Sp3‐, two simultaneous H bonds can be formed by two donors (i. e. pyrrolidinium and trialkyl ammonium ions), which can individually engage either O atoms (Figure 4A). In contrast to the Sp2‐ data, the preferred Sp3‐ conformation features the pyrrolidinium ion engaging the sp3‐hybridized O atom, whereas the sp2‐hybridized O atom binds the trialkyl ammonium ion. This “Sp3‐tBu syn” conformation is favored by 1.16 kcal/mol over the “Sp3‐tBu anti” counterpart, in which the pyrrolidinium and trialkyl ammonium ions engage the sp2‐ and sp3‐hybridized O atoms, respectively (Figure 4, for additional conformations of Sp2‐ and Sp3‐, see Figure S2 in the Supporting Information). Interestingly, the two pyrrolidinium protons in the “Sp3‐tBu syn” conformation proved more acidic (calculated pK a=8.9, 8.4) than the trialkyl ammonium species (pK a=11.0).

These in silico data suggest a possible explanation for the observed SI spacer reactivity. Firstly, at the onset of SI spacer degradation, it is reasonable to assume that the nucleophilic attack of the uncharged pyrrolidine N atom to the carbonyl group occurs with a N−C−O bond angle >90°, following the well‐known Bürgi‐Dunitz trajectory. The three‐dimensional analysis of the Sp‐carbamate modules indicates that this geometry of attack is only accessible by the carbamate syn rotamers. Secondly, considering the perturbed pK a of the pyrrolidinium ion in the “Sp3‐tBu syn” structure, it is possible that the pyrrolidine nucleophilic attack in Sp3 is also facilitated by a preferential proton dissociation in aqueous medium.

In summary, the tertiary amine proximity to the carbamate group may facilitate the SI spacer degradation at different stages of the cyclization reaction, following the mechanism proposed in Figure 4B. In particular, upon pyrrolidine deprotonation (Step 1, equilibrium governed by the pH of the aqueous medium), the intramolecular hydrogen bond between the trialkylammonium ion and the sp2‐hybridized O atom in intermediate I may facilitate the pyrrolidine cyclization (Step 2) to give the tetrahedral intermediate II. Here, the basicity of the tertiary amine may favor the restoration of the sp2‐hybridized C center (bicyclic urea formation) and the liberation of the hydroxy group (Step 3). This impact of neighbouring groups on bond cleavage kinetics is reminiscent of “catalytic triads” in the active site of hydrolytic enzymes, which have inspired the development of synthetic enzyme mimics. Similarly, a positively‐charged lysine ϵ‐amine group in an ADC construct was recently found to act as acid‐catalyst, and exploited to induce modifications of acetal and succinimide labels connected to the antibody core.

Conclusions

The present work describes a highly reactive pyrrolidine‐carbamate SI spacer (Sp3) in which a superfast carbamate cyclative cleavage is induced by the presence of a tertiary amine handle in close proximity to the carbamate bond. This advanced spacer showed better performance than Sp2 in the release of both the cytotoxic agent CPT and immunostimulatory drug R848. In silico conformational analysis and pK a calculations allowed us to propose a plausible rationale for the superior efficacy of the advanced SI spacer compared to reference Sp2. Considering the very slow R848 release observed with state‐of‐art SI spacers Sp1 and Sp2, the new spacer Sp3 may be pivotal for the design of cleavable R848 conjugates, alternative to previous strategies for IMD derivatization. Moreover, as we reported for the Sp2 reference,[ , ] the Sp3 module can be easily installed into functional drug delivery systems. For instance, the Sp3 connection to a para‐aminobenzyl carbamate (PABC) spacer will access to protease‐activable prodrugs, which represent the backbone of drug delivery technologies, including marketed ADCs. In general, Sp3 application in different types of stimuli‐responsive materials can be envisioned as a valid alternative to current strategies for hydroxyl cargo delivery.[ , , ]

Conflict of interest

The authors declare no conflict of interest.

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supporting Information

Click here for additional data file.