Tumor Targeting with an isoDGR-Drug Conjugate.

Herein we report the first example of an isoDGR-drug conjugate (2), designed to release paclitaxel selectively within cancer cells expressing integrin αV β3 . Conjugate 2 was synthesized by connecting the isoDGR peptidomimetic 5 with paclitaxel via the lysosomally cleavable Val-Ala dipeptide linker. Conjugate 2 displayed a low nanomolar affinity for the purified integrin αV β3 receptor (IC50 =11.0 nm). The tumor targeting ability of conjugate 2 was assessed in vitro in anti-proliferative assays on two isogenic cancer cell lines characterized by different integrin αV β3 expression: human glioblastoma U87 (αV β3 +) and U87 β3 -KO (αV β3 -). The isoDGR-PTX conjugate 2 displayed a remarkable targeting index (TI=9.9), especially when compared to the strictly related RGD-PTX conjugate 4 (TI=2.4).

Nowadays, the development of molecular devices able to selectively deliver chemotherapeutics at the disease site has gained a central position in cancer research. In particular, such targeting agents would allow circumventing the lack of selectivity observed when administering cytotoxic agents to patients. Due to this main limitation, traditional chemotherapy requires the use of high drug dosages, with consequent severe side effects that vitiate the overall efficacy of the therapy.1 A first approach consisted in the preparation of antibody–drug conjugates (ADCs), in which the use of monoclonal antibodies (mAbs) to target specific tumor antigens resulted in a clear discrimination of cancer cells from healthy tissues. However, this strategy presents several drawbacks, especially related to high manufacturing costs, poor pharmacokinetic properties and possible immune‐system‐induced alteration of drug efficiency.2 At this stage, small molecule–drug conjugates (SMDCs) arose as an alternative to ADCs: in this case, the targeting moiety is a small molecule, such as an oligopeptide, a peptidomimetic or a vitamin, capable of interacting selectively with particular proteins overexpressed by tumor cells. Unlike ADCs, the use of a small molecule ascribes improved pharmacokinetic properties to the entire conjugate, which in principle can be synthesized by easier and more affordable synthetic strategies.2

In the field of SMDCs, integrin αVβ3 represents a very interesting target to be exploited for the selective delivery of anticancer agents within the tumor site. As matter of fact, the expression of this transmembrane receptor is increased in a variety of human cancer types (e.g., breast cancer, glioblastoma, pancreatic tumor, prostate carcinoma) with respect to healthy tissues. The increased expression of αVβ3 integrin in tumor cells is associated with different pathological features: angiogenesis, tumor growth, apoptosis resistance, and metastasis.3 Integrin αVβ3 recognizes endogenous ligands by the tripeptide arginine‐glycine‐aspartate4 (RGD) and also by the related sequence iso‐aspartate‐glycine‐arginine5, 6 (isoDGR). In 2012, computational and biochemical studies showed that isoDGR‐containing cyclopeptides act as genuine αVβ3 antagonists, blocking the ligand binding site and inhibiting integrin allosteric activation.6a In contrast to the RGD ligands which in some cases may cause adverse paradoxical integrin activation effects,6a, 7 compounds based on the isoDGR motif could become a new generation of integrin‐binding drugs free from these drawbacks. For example, isoDGR ligand 1 (Figure 1) displays inhibitory effects on the FAK/Akt integrin‐activated transduction pathway and on integrin‐mediated cell infiltration processes, qualifying therefore as a true integrin antagonist.8

Figure 1

Structures of the integrin ligands cyclo[DKP‐isoDGR] (1) and cyclo[DKP‐RGD] (3), and of the corresponding SMDCs cyclo[DKP‐isoDGR]‐Val‐Ala‐PTX (2) and cyclo[DKP‐RGD]‐Val‐Ala‐PTX (4).

A variety of ligands containing the RGD sequence have been synthesized and reported in the literature so far, with some of them showing a very high affinity for the integrin receptor.9 Moreover, numerous RGD–drug conjugates have been developed for tumor targeting in the past two decades,10, 11, 12 while no example of isoDGR–drug conjugate has ever been reported. In fact, compared to the high binding affinity of the RGD ligands for αVβ3 integrin (IC50<15 nm),9 the isoDGR motif displayed much lower affinity (IC50≥43 nm),13 with a single notable exception (1, IC50=9.2 nm), see Figure 1.8

Herein we report the first example of an isoDGR–drug conjugate (2, Figure 1), based on ligand 1, which displays a high binding affinity for the purified integrin αVβ3 receptor (IC50=11.0 nm), see Table 1.

Table 1

Inhibition of biotinylated vitronectin binding to purified αVβ3 receptor.

Entry	Ligand	Structure	αVβ3 IC50 [nm][a]
1	1	cyclo[DKP‐isoDGR]	9.2±1.1
2	2	cyclo[DKP‐isoDGR]‐Val‐Ala‐PTX	11.0±0.2
3	3	cyclo[DKP‐RGD]	4.5±1.1
4	4	cyclo[DKP‐RGD]‐Val‐Ala‐PTX	13.3±3.6

[a] IC50 values were calculated as the concentration of compound required for 50 % inhibition of biotinylated vitronectin binding. Screening assays were performed by incubating the immobilized integrin αVβ3 with increasing concentrations (10−12–10−5  m) of the RGD or isoDGR ligands in the presence of biotinylated vitronectin (1 mg mL−1), and measuring the concentration of bound vitronectin in the presence of the competitive ligands.

Conjugate 2 has been designed in a way similar to the corresponding RGD‐drug conjugate (4, Figure 1),10g which contains the RGD integrin ligand 3.14 The isoDGR targeting moiety has been linked to the cytotoxic agent paclitaxel using the lysosomally cleavable dipeptide Val–Ala: this sequence showed high plasma stability, whereas it is rapidly cleaved by lysosomal cysteine proteases (such as cathepsins B and D) upon integrin‐mediated internalization by endocytosis.10g, 15, 16

In order to synthesize conjugate 2, first we prepared peptidomimetic 5 (Scheme 1), a derivative of the isoDGR ligand 1 bearing an amino functional group suitable for conjugation. Bifunctional diketopiperazine 6 10h was Boc‐deprotected and reacted with Cbz‐Arg(Mtr) to give carboxylic acid 9 upon allyl ester cleavage. Acid 9 was coupled with dipeptide 13, obtained starting from protected aspartic acid 10 (commercially available) and benzyl glycinate 11. The benzyl (Bn) and carboxybenzyl (Cbz) protecting groups of the resulting compound 14 were selectively removed by catalytic hydrogenolysis to afford amino acid 15, which was cyclized under high dilution conditions (1.4 mm). Mtr‐ (4‐methoxy‐2,3,6‐trimethylbenzenesulphonyl) and tert‐butyl ester removal on macrolactam 16 afforded the desired isoDGR peptidomimetic 5 after HPLC purification and freeze‐drying. The benzylic amine of compound 5 was coupled with 18, the N‐hydroxysuccinimidyl ester of carboxylic acid 17,10g to give compound 19 (Scheme 2). Treatment of 19 with trifluoroacetic acid in dichloromethane afforded amine 20 which was reacted with carbonate 22 to obtain the final isoDGR‐PTX conjugate 2.

Scheme 1

Synthesis isoDGR peptidomimetic 5. Reagents and conditions: a) TFA/CH2Cl2 1:2, RT, 2 h; b) Cbz‐Arg(Mtr)‐OH, HATU, HOAt, iPr2NEt, DMF, 0 °C to RT, overnight, 94 % over 2 steps; c) [Pd(PPh3)4], N‐methylaniline, CH2Cl2, 0 °C, 1 h, 88 %; d) HATU, HOAt, iPr2NEt, DMF, 0 °C to RT, overnight, 86 %; e) piperidine, DMF, 2 h, RT, 67 %; f) 13, HATU, HOAt, iPr2NEt, DMF, 0 °C to RT, overnight, 95 %; g) H2, 10 % Pd/C, THF/H2O 1:1, overnight, RT, 95 %; h) HATU, HOAt, iPr2NEt, DMF/CH2Cl2 1:1 (1.4 mm), 0 °C to RT, overnight, 79 %; i) TFA/TMSBr/thioanisole/EDT/phenol 70:14:10:5:1, 2 h, RT, 47 %. (TFA=trifluoroacetic acid, HATU=1‐[Bis(dimethylamino)methylene]‐1H‐1,2,3‐triazolo[4,5‐b]pyridinium 3‐oxid hexafluorophosphate, TMSBr=bromotrimethylsilane, EDT=1,2‐ethanedithiol).

Scheme 2

Synthesis of isoDGR–drug conjugate 2. Reagents and conditions: a) DIC, NHS, DMF, 0 °C to RT, overnight; b) 5, CH3CN/PBS 1:1; pH 7.3–7.6, 0 °C to RT, overnight; c) TFA/CH2Cl2 1:2, 1 h, RT, 55 % over three steps; d) 4‐Nitrophenylchloroformate, pyridine, CH2Cl2, −50 °C to −20 °C; 4 h, 69 %; e) 22, iPr2NEt, DMF, 0 °C to RT, overnight, 55 %. (PBS=phosphate‐buffered saline).

The tumor‐targeting ability of conjugate 2 was assessed in vitro against two isogenic cancer cell lines characterized by different integrin αVβ3 expression. U87 human glioblastoma cells were selected as the integrin αVβ3‐expressing cell line and, at the same time, used to generate the corresponding clone U87 β3‐KO, in which the expression of the gene encoding for the β3 integrin subunit was deleted using CRISPR‐Cas9 gene editing technology.17 Flow cytometry studies on U87 and U87 β3‐KO cells confirmed the absence of integrin αVβ3 in the U87 β3‐KO cell line (Figure 2).

Figure 2

Flow cytometry experiments on U87 and U87 β3‐KO cells to assess the different αVβ3 integrin expression. Cells were incubated with the secondary antibody (CF488A‐goat anti‐mouse IgG, Biotium 20 011), or with the anti‐αVβ3 antibody (clone LM609‐Millipore MAB 1976) followed by the secondary antibody, see the Supporting Information.

Clone U87 β3‐KO is not a perfect negative control, as the conjugates can still be actively internalized upon interaction with other integrins, for example, αVβ5. Therefore, the selectivity shown by the conjugates in cell viability experiments using U87 β3‐KO and U87 cells [IC50(αVβ3−)/IC50(αVβ3+)] should be considered as a minimum (conservative) value.18 Cells were incubated with increasing doses of isoDGR–PTX conjugate 2 for 144 h, before measuring the cell viability in culture (Table 2).

Table 2

Evaluation of anti‐proliferative activity of isoDGR‐PTX conjugate 2 in U87 and U87 β3‐KO.

Structure	IC50 [nm][a]		TI[b]
	U87 (αVβ3+)	U87 β3‐KO (αVβ3−)
Paclitaxel (21)	0.64	0.28	1
isoDGR‐PTX conjugate (2)	927.6	4003.0	9.9
RGD‐PTX conjugate (4)	550.0	581.7	2.4

[a] IC50 values were calculated as the concentration of compound required for 50 % inhibition of cell viability in culture, based on quantitation of the ATP present as estimated by CellTiter‐GLO; cells were treated for 144 h in 96‐well plates. [b] Targeting index (TI): [IC50(αVβ3−)/IC50(αVβ3+)]conjugate/[IC50(αVβ3−)/IC50(αVβ3+)]paclitaxel.

Parallel experiments were performed under the same conditions with the RGD–PTX conjugate 4 (Figure 1) and paclitaxel (PTX, 21). Paclitaxel itself appeared to be 2.3 times more effective on U87 β3‐KO cells, possibly because the β3‐integrin‐depleted cells divide faster. Taking into account the intrinsic selectivity shown by free paclitaxel, the isoDGR–PTX conjugate 2 displayed a remarkable targeting effect (TI=9.9), especially when compared to the strictly related RGD–PTX conjugate 4 (TI=2.4). Apparently, the isoDGR‐PTX conjugate 2 is recognized more specifically by integrin αVβ3 than the related RGD‐PTX conjugate 4 which can be effectively internalized also by other integrins expressed on the cell surface.

In conclusion, the first isoDGR–drug conjugate (2) has been developed for tumor targeting. Compound 2 displayed a low nanomolar affinity for the purified integrin αVβ3 receptor and a notable targeting ability when tested on two isogenic cancer cell lines expressing integrin αVβ3 at different levels. Fine tuning of the linker19 is in progress in order to improve the potency of the conjugate while retaining the high selectivity.

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.

Supplementary

Click here for additional data file.