Research Article
Synthesis and initial in vitro evaluation of PSMA-targeted ligands with optimized structural motifs.
https://doi.org/10.21203/rs.3.rs-1987200/v1
This work is licensed under a CC BY 4.0 License
You are reading this latest preprint version
An improved series of ligands targeting prostatic specific membrane antigen (PSMA) has been reported. Varying compounds and their biological parameters were due to changes in the linker structure. Highly selective compounds with nanomolar IC50 values were obtained. As an example, a conjugate with Sulfo-Cy5 and MMAE was obtained and pre-studied.
Prostate cancer
SulfoCy5 label
Monomethyl auristatin E
Drug delivery
prostate-specific membrane antigen
Cancer is one of the leading socially significant diseases. Prostate cancer (PC) is the second most common disease among men after lung cancer.[1] One of the protein markers of this disease is prostate-specific membrane antigen (PSMA). In tumor cells of the prostate its expression is dramatically increased in comparison with healthy tissues. The good interaction with urea-based vectors in the active center of PSMA makes it possible to create targeted high-affinity conjugates.[2, 3]
The linker has a great influence on the affinity of the ligands. Its functions include not only the binding of the conjugate into a one-molecule structure, but also the interaction with the funnel-shaped tunnel leading to the active center of the PSMA. Previous studies have shown that one of the effective modifications of the linker is the introduction of a dipeptide chain based on phenylalanine and/or tyrosine into its structure. The aromatic rings included in the side chains of amino acids have the greatest influence.[4 – 6] By introducing substituents of different nature into their structure, high-affinity rates can be achieved (Figure 1A). This study was an extension of the previously developed series of PSMA ligands (a complete table of ligands including the previously developed library is given in the supplementary data) (Table S2) [7, 8]
A series of 14 ligands was obtained based on the methodologies we developed and optimized earlier. The detailed synthetic scheme of these ligands is given in the Supplementary data. The choice between the liquid-phase or solid-phase synthesis method for each ligand was based on several factors. For example, the choice of the solid-phase method for ligands with N-terminal tyrosine in the dipeptide chain is due to the difficulty of synthesizing such a dipeptide in the liquid phase. The choice of the synthesis method was also influenced by the availability of the used amino acids and their Fmoc-analogs. Both methods allow to synthesise ligands of a given structure and configuration.[9 – 11] The affinity analysis of the obtained compounds was performed by inhibiting the cleavage reaction of N-acetylaspartylglutamate.[12]
According to the phenylalanine and glycine-based ligand series, the aromatic ring of the C-terminus amino acid has a greater effect on affinity (Table 1). When it is replaced by glycine, the IC50 deteriorates by a factor of 3. The absence of an aromatic fragment in the position close to the vector-molecule in general also negatively affects the affinity, but not as significantly. Therefore, -Br, -Cl, -NO2, and -OH substituted phenylalanine and tyrosine analogs were introduced into the C-terminal amino acid. While for the N-terminal amino acid, the Tyr/Phe series was extended.
Table 1. Results of in vitro evaluation of inhibition of the N-acetylaspartylglutamate cleavage reaction.
|
№ |
Linker structure |
IC50 ± SD, nM |
№ |
Linker structure |
IC50 ± SD, nM |
|
DCL* |
- |
546 ± 388 |
1j |
L-Phe-L-Phe(3-Br) |
3.1 ± 0.9 |
|
ZJ-43 |
- |
11.0 ± 2.6 |
1k |
L-Phe-L-Phe(2-Br) |
4.6 ± 1.3 |
|
2-PMPA** |
- |
80 ± 23 |
1l |
D-Phe-D-Phe(4-Br) |
12 ± 2 |
|
1a |
L-Tyr-D-Tyr |
99 ± 53 |
1m |
D-Phe-L-Tyr(3-OH) |
17 ± 3 |
|
1b |
L-Tyr-D-Phe |
54 ± 16 |
1n |
L-Phe-L-Tyr(3-Cl) |
3.6 ± 0.9 |
|
1c |
D-Tyr-L-Tyr |
41 ± 14 |
I |
L-Phe-Gly |
114± 48 |
|
1d |
L-Tyr-L-Phe |
2.6 ± 0.8 |
II |
Gly-L-Phe |
38 ± 10 |
|
1e |
L-Tyr-L-Tyr |
2.1 ± 1.2 |
III |
L-Phe-L-Tyr |
9.3 ± 3.0 |
|
1f |
L-Phe-L-Tyr(3-NO2) |
1.5 ± 0.9 |
XII |
L-Phe-L-Phe(4-NO2) |
132 ± 32 |
|
1g |
D-Phe-L-Tyr(3-NO2) |
1.6 ± 1.1 |
XIII |
D-Phe-D-Tyr(3-Br) |
80 ± 50 |
|
1h |
D-Phe-L-Tyr(3-Br) |
4.9 ± 2.6 |
XIV |
L-Phe-L-Tyr(3-Br) |
196 ± 102 |
|
1i |
L-Phe-L-Phe(4-Br) |
1.9 ± 0.8 |
XV |
L-Phe-L-Tyr(3-OH) |
892 ± 368 |
* DCL - (S)-2-(3-((S)-5-amino-1-carboxypentyl)ureido)pentanedioic acid, ** 2-PMPA - 2-Phosphonomethyl pentanedioic acid
The substituted aromatic fragment in the structures 1f-1n is in the C-terminal amino acid. The result for compounds XIII D-Phe-D-Tyr(3-Br) 80.0 ± 50.0 nM; XIV L-Phe-L-Tyr(3-Br) 196.0 ± 102.0 nM confirms that configuration has a strong influence on affinity. The two conjugates 1f L-Phe-L-Tyr(3-NO2) 1.5 ± 0.9 nM and 1g D-Phe-L-Tyr(3-NO2) 1.6 ± 1.1 nM containing a nitro group at the 3-position of the C-terminal tyrosine in the dipeptide showed the best results. Though, in the case of XII L-Phe-L-Phe(4-NO2) 132.0 ± 32.0 nM such a good result is not observed, which suggests that the hydroxyl group in the 4-position of the aromatic ring is most favorable, while the nitro group, in this case, is too bulky. It is worth noting that the introduction of bromine and chlorine into the C-terminal amino acid of the dipeptide chain resulted in nanomolar affinity values.
Another well-established ligand is 1e L-Tyr-L-Tyr 2.10 ± 1.24 nM, which also has an unmixed LL configuration and a substituted aromatic ring at the C-terminus. The introduction of the second hydroxyl fragment (see Supplementary data) (XV L-Phe-L-Tyr(3-OH) 892 ± 368 nM) dramatically decreases PSMA binding ability.
We can conclude that the linker structure in the N-terminus of the dipeptide benefits from the introduction of phenylalanine or tyrosine in the L-configuration. The substituents in aromatic fragments in the C-terminal amino acid of the linker can also have a significant effect. But it is inextricably dependent upon the configuration of the amino acid residue.
All structures of interest showed docking score values comparable with reference compound PSMA-617 (Table S2). The docking study revealed an ability to share almost the same binding mode for all proposed structures. In all structures, the core-head optimally fits a well-established tunnel-shaped binding site forming a net of hydrogen bonds with Arg210, Asn257, Arg534, Arg536, and Lys699 as well as providing a tight coordination bond with Zn2+ via its urea oxygen. These observations are in good agreement with earlier studies.[13] As an example of a typical ligand position, the binding mode of 1f is shown in Figure 1B. A flexible linker optimally serves for the placement of functional moieties into appropriate subpockets to form a high-affinity net of ligand-protein interaction. At the same time, the linker itself participates in several additional intermolecular contacts including a hydrogen bond with Arg511 and a water bridge with Arg534.
Each conjugate of interest carries three aromatic rings attached to the linker. One of these rings occupies a part of the tunnel-shaped cavity and forms hydrophobic contact with Tyr700, Lys207, and Ser547. At the same time, this ring, if decorated by an appropriate functional group (carboxyl or nitro group), provides an opportunity to form polar contacts with the Lys207 residue. The other two rings fit small subpockets on the opposite end of the binding site. One of them forms π-cation interaction with Lys514. Both stereoisomers at this position retain this interaction, and neither of them confers a significant advantage in calculated binding energy. The second ring forms planar stacking with Trp541 and π-cation interaction with Arg511. Functionalization of this ring with polar substituents (i.e., hydroxyl, nitro group) seems to be reasonable due to the formation of direct contacts or water bridges with the Glu542, Ser501, and Asp465 residues.
For this substituent, stereospecificity plays a more vital role. In the case of the D-isomer at this position, the geometry of the linker hinders optimal binding to the targeted cavity, which leads to a decrease in the binding energy. For example, 1e L-Tyr-L-Tyr exhibits a superior docking score (-14.0 kcal/mol) than 1b L-Tyr-D-Phe (-13.2 kcal/mol). It should be noted that these observations are in accordance with the experimentally obtained activity values.
1f L-Phe-L-Tyr(3-NO2) 1.5 ± 0.9 nM and 1e L-Tyr-L-Tyr 2.10 ± 1.24 nM was taken for conjugation with the fluorescent label Sulfo-Cy5[14]. The general structure of the obtained conjugates is shown in Figure 2A. During the experiment, cell lines were incubated with 2a, 2b, and III conjugate solutions of various concentrations (4 nm, 40 nm). The percentage of stained cells was determined afterward. The obtained data are shown in Figure 2B.
As a result, for all conjugates, an increase in the percentage of stained cells was observed with the increasing concentration of the studied compound. It is also worth noting the correlation between the affinity of the initial ligands and the percentage of stained cells: thus, the highest proportion of stained cells at all concentrations was shown by conjugate 2a L-Phe-L-Tyr(3-NO2)-Sulfo-Cy5, which is based on the most prominent ligand 1f.
The 2a conjugate gave the best fluorescence signal accumulation. Accordingly, the L-Phe-L-Tyr(3-NO2) ligand was chosen to synthesize a therapeutic conjugate with the drug monomethyl auristatin E (Figure 3A)[15, 16]. Biological in vitro cytotoxicity tests of the obtained conjugate were performed on four cell lines: 22Rv1 (PSMA+), LNCaP (PSMA++), A549 (PSMA -) and PC-3 (PSMA-) by MTS-test. Cell line LNCaP with high PSMA expression is less sensitive to MMAE (monomethyl auristatin E) itself then PSMA-negative cell lines A549 and PC-3. After conjugation the compound 3 L-Phe-L-Tyr(3-NO2)-MMAE became slightly more toxic to PSMA-positive LNCaP (Table S3) cells. The therapeutic conjugate with monomethyl auristatin E PSMA-Val-Cit-PAB-MMAE obtained previously was used as a comparison conjugate.[17]
This article presents a previously undescribed series of high-affinity PSMA ligands in which IC50 values in the range of 1-2 nM were achieved. In terms of inhibition of NAAG cleavage, the obtained ligands are ahead of the previously studied analogs as well as such standard molecules as 2-PMPA, ZJ-43, and DCL. Due to the high affinity of the obtained ligands, they can serve as an optimized high-affinity platform for both therapeutic conjugates and conjugates with fluorescent dyes, as well as serve as the basis for radiopharmaceutical conjugates, which will be further studied.
Acknowledgements
The research was carried out with the financial support of Russian Science Foundation grant No. 22-15-00098 (PSMA-ligands design, synthesis and physico-chemical characterisation) and Russian Foundation for Basic Research, Grant № 20-34-90150 (inhibition of the N-acetylaspartylglutamate cleavage).
Conflict of interest
The authors declare that they have no conflict of interest.
Graphical Abstract
Reviewers agreed at journal
30 Aug, 2022
Reviewers invited by journal
30 Aug, 2022
Editor assigned by journal
23 Aug, 2022
First submitted to journal
22 Aug, 2022
You are reading this latest preprint version
