Linker Modification Strategies to Control the Prostate-Specific Membrane Antigen (PSMA)-Targeting and Pharmacokinetic Properties of DOTA-Conjugated PSMA Inhibitors

Article abstract of DOI:10.1021/acs.jmedchem.5b01210

Since prostate-specific membrane antigen (PSMA) is up-regulated in nearly all stages of prostate cancer (PCa), PSMA can be considered as a viable diagnostic biomarker and treatment target in PCa. This project is focused on the development and evaluation of a series of compounds directed against PSMA. The modifications to the linker are designed to improve the binding potential and pharmacokinetics for theranostic application. In addition, the results help to further elucidate the structure-activity relationships (SAR) of the resulting PSMA inhibitors. Both in vitro and in vivo experiments of 18 synthesized PSMA inhibitor variants showed that systematic chemical modification of the linker has a significant impact on the tumor-targeting and pharmacokinetic properties. This approach can lead to an improved management of patients suffering from recurrent prostate cancer by the use of one radiolabeling precursor, which can be radiolabeled either with ⁶⁸Ga for diagnosis or with ¹⁷⁷Lu or ²²⁵Ac for therapy.

Full text of DOI:10.1021/acs.jmedchem.5b01210

Subscriber access provided by RUTGERS UNIVERSITY
Article
Linker Modification Strategies to Control the Prostate-
specific Membrane Antigen (PSMA)-Targeting and
Pharmacokinetic Properties of DOTA-Conjugated PSMA Inhibitors
Martina Benesova, Ulrike Bauder-Wüst, Martin Schäfer, Karel D.
Klika, Walter Mier, Uwe Haberkorn, Klaus Kopka, and Matthias Eder
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.5b01210 • Publication Date (Web): 15 Feb 2016
Downloaded from http://pubs.acs.org on February 18, 2016
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Linker Modification Strategies to Control the
Prostateꢀspecific Membrane Antigen (PSMA)ꢀ
Targeting and Pharmacokinetic Properties of DOTAꢀ
Conjugated PSMA Inhibitors
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
§
§
§
†
‡
Martina Benešová, Ulrike BauderꢀWüst, Martin Schäfer, Karel D. Klika, Walter Mier, Uwe
Haberkorn, Klaus Kopka, and Matthias Eder*,§,⊥
‡
,ǁ
§,⊥
§
Division of Radiopharmaceutical Chemistry, German Cancer Research Center (DKFZ),
Heidelberg, Germany
†
Molecular Structure Analysis, German Cancer Research Center (DKFZ), Heidelberg, Germany
‡
Department of Nuclear Medicine, University Hospital Heidelberg, Heidelberg, Germany
ǁ
Clinical Cooperation Unit Nuclear Medicine, German Cancer Research Center (DKFZ),
Heidelberg, Germany
Since prostateꢀspecific membrane antigen (PSMA) is upꢀregulated in nearly all stages of prostate
cancer (PCa), PSMA can be considered as a viable diagnostic biomarker and treatment target in
PCa. This project is focused on the development and evaluation of a series of compounds
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
directed against PSMA. The modifications to the linker are designed to improve the binding
potential and pharmacokinetics for theranostic application. In addition, the results help to further
elucidate the structure–activity relationships (SAR) of the resulting PSMA inhibitors. Both in
vitro and in vivo experiments of 18 synthesized PSMA inhibitor variants showed that systematic
chemical modification of the linker has a significant impact on the tumorꢀtargeting and
pharmacokinetic properties. This approach can lead to an improved management of patients
suffering from recurrent prostate cancer by the use of one radiolabeling precursor, which can be
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
6
8
177
225
radiolabeled with either Ga for diagnosis or with Lu, or Ac, for therapy.
KEYWORDS
Prostate cancer, prostateꢀspecific membrane antigen (PSMA), lowꢀmolecular weight ligands,
structure–activity relationships (SAR), PET imaging, endoradiotherapy, theranostics
ACS Paragon Plus Environment

Page 3 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
INTRODUCTION
Prostate cancer (PCa) is the most commonly diagnosed malignancy in men worldwide and
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1–3
remains one of the leading causes for their cancerꢀrelated deaths. Despite significant efforts,
4
,5
currently available diagnostic and therapeutic strategies are often ineffective, particularly for
6
hormoneꢀrefractory PCa resulting in a very poor prognosis for these patients. With regard to this
situation, further development of a biomarkerꢀspecific detection, accurate staging of primary as
well as recurrent disease, and highly efficient endoradiotherapy are of utmost clinical interest.
Prostateꢀspecific membrane antigen (PSMA) is a type II membrane glycoprotein with an
extensive extracellular domain (44–750 amino acids) and plays a significant role in prostate
7
carcinogenesis and progression. For nearly two decades PSMA has been a wellꢀcharacterized
8
viable biomarker and treatment target in PCa. Additionally, lowꢀlevel PSMA expression was
9
10
shown in primary renal tumors and transitional cell carcinoma of the bladder. Moreover,
PSMA is presented on the neovasculature within multiple, nonꢀprostate solid tumors and may
1
1,12
also be a promising molecule for targeting tumorꢀassociated blood vessels.
Diverse chemical scaffolds have been investigated as potential ligands of PSMA including lowꢀ
13–16
17–19
molecular weight inhibitors,
various generations of antibodies,
RNA aptamer
2
0–22
23–25
conjugates,
and several combinations with targeted nanoparticles.
Originally, PSMA was
1
11
identified as the antigen of the monoclonal antibody (mAB) 7E11ꢀC5. Its Inꢀlabelled version,
commercially known as ProstaScint, is the only FDA (Food and Drug Administration) approved
26,27 111
radiopharmaceutical for the PSMAꢀtargeted treatment of PCa.
InꢀLabeled ProstaScint
recognizing the intracellular domain of PSMA became the most widely studied radiolabeled
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
PSMAꢀtargeting mAB. However, the large size of this antibody tends to prolong clearance and
access to all portions of the tumor is typically restricted, especially for viable cells with intact
2
8,29
membranes.
A second generation genetically engineered humanized monoclonal antibody is
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
huJ591, which recognizes the extracellular portion of PSMA and was recently successfully
9
0
111
131
177
30,31
translated into clinical trials. Yꢀ,
Inꢀ,
Iꢀ, and
Luꢀlabeled huJ591
have all been
8
9
described thus far. Another example of a successful candidate for clinical testing is ZrꢀDfꢀ
IAB2M, an antiꢀPSMA minibody (Mb), for imaging primary prostate cancer and soft tissue and
3
2
8,15,33–36
bone metastases. Lowꢀmolecular weight PSMAꢀtargeting radioligands
are also shown
to detect metastatic or recurrent PCa with high specificity and sensitivity and seem to be more
potent than antibodies with respect to tissue penetration, halfꢀlife of circulation, and productionꢀ
related costs.
2+
The active site of PSMA consists of a binuclear Zn active site and the S1 and S1' subpockets.
2
+
Besides the primary interactions of urea and carboxylate groups at the Zn active site and the S1'
subpocket of PSMA, there are potential lipophilic interactions within the S1ꢀaccessory
hydrophobic pocket. Prospective compounds lacking any one of these interactions have shown
3
7–42
significant attenuation in their binding properties.
However, the glutamate–urea–lysine
4
1,43
binding structural motif is a building block well suited for preparing potent PSMAꢀinhibitors.
In an attempt to improve both the detection by scintigraphic technologies such as positron
emission tomography (PET) and the treatment of PCa by PSMAꢀtargeting systemic radionuclide
therapy (endoradiotherapy), we synthesized a series of ureaꢀbased DOTAꢀconjugated PSMAꢀ
inhibitors (Figure 1) and evaluated their in vitro and in vivo properties. To retain the binding
affinity of the PSMA inhibitors, various linker moieties (Figure 2) were investigated between the
pharmacophore and DOTA, introduced correspondingly as carboxamide, which is able to
ACS Paragon Plus Environment

Page 5 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
incorporate diagnostic as well as therapeutic radionuclides. The relative ease of solidꢀphase
synthesis also facilitates structure–activity relationship studies. Overall, our approach can serve
as a useful platform for structureꢀaided design and further development of these inhibitors with
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
44
favorable tumorꢀtargeting and pharmacokinetic properties.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
RESULTS
The multistep solidꢀphase synthesis provided the compounds in isolated overall yields of 32–
8%. Subsequent purification provided final products with purities >98% by analytical RPꢀ
4
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
HPLC. Analytical data for the compounds are presented in Table 1. All compounds were found
to be stable for at least 6 months as lyophilized powders at −20 °C, and all compounds are stable
for at least 3 months in d ꢀDMSO solution at −20 °C with only slight decomposition evident
6
after an extended period of time. However, at ambient temperature, compound stability is limited
to less than 1 month in d ꢀDMSO solution as confirmed by NMR spectroscopy, RPꢀHPLC, and
6
MALDIꢀMS.
The PSMA inhibitors synthesized (Figure 2) were divided into four sets each based on the
chemical similarities in the structural changes of the linker. The first set of compounds (11–16)
included different numbers and arrangement of aromatic units (aminobenzyl) in the linker
region. Compound 11 employs as a first building block linear aminohexanoic acid and then an
aromatic amino acid (4ꢀaminomethylbenzoic acid, AMBA). The linker in 12, 13, 14 and 15
contains either one, two, three or four AMBA building blocks, respectively. Additionally, 16
contains biphenylꢀ
NHꢀCOꢀNHꢀLysꢀ2ꢀnaphthylꢀ
comprises 2ꢀnaphthylꢀ ꢀalanine as the first building block which is more flexible compared to
aminobenzyl used in the first set. The cyclohexanoic acid creates the second building block in 17
PSMAꢀ617). Compound 18 differs in conformation as it employs 2ꢀnaphthylꢀ ꢀalanine. 19, on
the other hand, differs in the naphthylic structure as it uses 1ꢀnaphthylꢀ ꢀalanine. Moreover 20
Lꢀalanine as a linker. The second set of compounds (17–20), inclusive Gluꢀ
4
4
L
ꢀAlaꢀcyclohexaneꢀDOTA (commercially known as PSMAꢀ617 ),
L
(
D
L
demonstrates simple sequence interchange of the building blocks compared to 17. The third set
ACS Paragon Plus Environment

Page 7 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
of compounds (21–24) describes the influence of the second building block while maintaining
the 2ꢀnaphthylꢀLꢀalanine building block as a constant in the linker (21), while 22 contains
additionally an aromatic unit, 23 aminohexanoic acid and 24 both an aromatic and cyclohexane
unit. The fourth set of compounds (25–28) examines the influence of the cyclohexane moiety as
the linker contains one cyclohexane ring (25), two cyclohexane rings (26) and three cyclohexane
rings (27). Moreover, 28 employs one cyclohexane and one aromatic ring in the linker area.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
13
Characterization of the compounds was conducted extensively by H and C NMR. Analysis of
the compounds using 2D NMR techniques (COSY, HSQC, and HMBC spectra) to assign the
1
13
signals from the 1D H and C NMR spectra was focused on compounds 12, 14, 17, and 19 as
these inhibitors represented the most promising candidates for success in concert with the stated
aims of the work and the unequivocal assignment of signals was accomplished for these
structures. Assignments could then be applied to the remaining compounds by analogy together
with the use of incremental shift calculations to effect partial or tentative assignments. DEPT
1
3
spectra were used for the assignment of the C multiplicities but were also effective for the
direct observation of solventꢀoverlapped signals. The dynamic effects arising from the slow
motional averaging of the macrocyclic ring on the NMR timescale leading to severe broadening
of signals in many cases, particularly those associated with the macrocyclic ring, were attenuated
1
3
for C NMR spectra by the acquisition at lower field and by the acquisition of spectra at
elevated temperatures (generally 70 and 90 °C). All compounds were deemed to be fully
consistent with their assigned structures by NMR in terms of proton count, carbon count, and
carbon multiplicities. For NMR data and spectra, see Supporting Information.
177
Radiolabeling with
Lu (15 min at 98 °C) provided labeling efficiency >98% for all
6
8
compounds, even for low amounts of used precursor (0.5 nmol). Radiolabeling with Ga
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
provided labeling efficiencies >95% (5 nmol) or >90% for low amounts of precursor (0.5 nmol)
after 15 min at 98 °C. However, 24 demonstrated instability and degradation at higher
temperatures as confirmed by NMR and was thus neither tested in vitro nor in vivo.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The potential binding to the serum proteins was performed for seven selected compounds, viz.
11, 14, 17, 22, 26, and 28. These representative compounds tested through gel filtration did not
demonstrate any transfer of activity to proteins contained in human serum. The distribution
coefficients determined via distribution in a twoꢀphase system (nꢀoctanol and HEPES buffer at
pH 7.4) are presented in Table 1. The logD values ranged from –2.12 ± 0.33 for 15 to –0.91 ±
0.49 for 25 (p = 0.0005). To compensate for differing solubilities, all compounds were diluted to
10 mM solution in DMSO for subsequent experiments.
All compounds revealed high inhibition potencies for PSMA for the human prostate cancer cell
line LNCaP. In this library of 18 compounds, significant differences in PSMA inhibition
potencies were found, e.g. K = 0.50 ± 0.63 nM for 16 and 28.60 ± 13.10 nM for 21 (p = 0.0206).
i
Four sets of compounds were defined with each set representing chemical structure related
similarities. The first set is based on the different number and arrangement of aromatic rings in
the linker area and demonstrated overall high inhibition potencies except for 11 (K = 15.31 ±
i
5.10 nM). The linker area of this compound contains two building blocks, the first one is
composed of aminohexanoic acid and the second one contains a benzene ring. The removal of
the aminohexanoic acid led to the compound 12 with significantly improved binding affinity (Ki
=
1.97 ± 0.76 nM; p = 0.0110). The compounds with a contiguous line of an increasing number
of aromatic rings, 13 (K = 7.51 ± 6.12 nM), 14 (K = 0.53 ± 0.47 nM), and 15 (K = 6.19 ± 1.51
i
i
i
nM) showed fluctuation in their binding affinities. A high binding affinity was demonstrated also
ACS Paragon Plus Environment

Page 9 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
by the final member of this set, 16, with a biphenylꢀ
nM).
Lꢀalanine building block (K = 0.50 ± 0.63
i
The second set is based on different arrangements of the naphthyl alanine and the cyclohexane
moiety. Compound 17 which contains 2ꢀnaphthylꢀ ꢀalanine and cyclohexane showed a K of 2.34
± 2.94 nM, 18 with 2ꢀnaphthylꢀ ꢀalanine and cyclohexane demonstrated a K of 8.86 ± 3.64 nM,
and 19 with 1ꢀnaphthylꢀ ꢀalanine and cyclohexane exhibited a K of 12.78 ± 6.41 nM. The last
compound of this set, 20, which also consists of 2ꢀnaphthylꢀ
as an inverse arrangement as compared to 17, showed the lowest binding affinity (K = 17.70 ±
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
L
i
D
i
L
i
Lꢀalanine and cyclohexane, however
i
1.41 nM; p = 0.0012).
The third set is characterized by having 2ꢀnaphthylꢀ
Lꢀalanine as a constant in the linker.
Compound 21 with 2ꢀnaphthylꢀ ꢀalanine demonstrated the lowest binding affinity of all
L
compounds (K = 28.60 ± 13.10 nM). When the second building block, a benzene ring (22) or
i
aminohexanoic acid (23), was added, the binding affinity increased significantly (K = 1.81 ±
i
0
.80 nM for 22; p = 0.0241 and K = 6.31 ± 2.22 nM for 23; p = 0.0439).
i
Finally, the last set is based on compounds with a contiguous line of an increasing number of
cyclohexane moieties. No significant differences in binding affinity for these compounds were
observed (K = 9.33 ± 28.39 nM for 25; K = 20.96 ± 29.65 nM for 26; K = 19.40 ± 5.18 nM for
i
i
i
27). Compound 28 contains one cyclohexane and one additional aromatic moiety (a benzene
ring) in the linker (K = 24.71 ± 24.99 nM).
i
In addition, the specific internalization into LNCaP cells was determined for all compounds
synthesized in this study. The first set of compounds demonstrated overall high internalization
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
6
6
(
5.11 ± 2.46 %IA/10 LNCaP cells for 12; 3.66 ± 1.31 %IA/10 LNCaP cells for 13; 12.49 ±
6
6
1
.38 %IA/10 LNCaP cells for 14; 6.86 ± 1.59 %IA/10 LNCaP cells for 15) except for 11 (2.11
6
6
±
0.02 %IA/10 LNCaP cells) and 16 (2.22 ± 1.01 %IA/10 LNCaP cells). These compounds
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
with a contiguous line of an increasing number of aromatic rings, 12, 13, 14, and 15, showed
fluctuation regarding their internalization into LNCaP cells. The optimal number of benzene
rings in the linker based on binding affinity appears to be three aromatic moieties, as represented
by 14.
Compound 17 is the most promising compound of the second set based on the inhibitory
6
potency, showing overall the highest internalization (16.17 ± 3.66 %IA/10 LNCaP cells). With
more apparent changes in the linker area of 17, the internalization strongly decreased (7.18 ±
6
6
3
.47 %IA/10 LNCaP cells for 18, 0.90 ± 0.07 %IA/10 LNCaP cells for 19, 0.13 ± 0.01
6
%
IA/10 LNCaP cells for 20).
The third set of compounds revealed the importance for the presence of aromatic building blocks
6
in the linker as 22 showed the highest internalization of this set (7.17 ± 2.82 %IA/10 LNCaP
6
cells). Compounds 21 and 23 demonstrated similar internalization, 1.77 ± 1.19 %IA/10 LNCaP
6
cells and 2.99 ± 1.65 %IA/10 LNCaP cells, respectively.
Finally, the last set of compounds exhibited very low internalization into LNCaP cells (1.24 ±
6
6
0
.24 %IA/10 LNCaP cells for 25, 1.43 ± 0.21 %IA/10 LNCaP cells for 26, and 0.03 ± 0.02
6
%
IA/10 LNCaP cells for 27). The last compound of this set, 28, with an additional aromatic
6
moiety, showed higher internalization than the other members of this set, 2.70 ± 2.06 %IA/10
LNCaP cells. The results from the binding and internalization assays are summarized in Table 2.
ACS Paragon Plus Environment

Page 11 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
For the overall cell surface binding and internalization (without subtracting the nonꢀspecific cell
surface binding and the nonꢀspecific internalized fraction), see Supporting Information.
68
The time–activity curves obtained from ꢀPET imaging of Gaꢀlabeled compounds in nude mice
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
bearing LNCaP tumor xenografts resulted in SUV values of accumulated radioactivity in the
tumor, muscle, and kidneys as well as for tumorꢀtoꢀmuscle (tumorꢀtoꢀbackground, TꢀtoꢀB) ratios
at 1 h postꢀinjection (Table 3). The highest TꢀtoꢀB ratio (8.5) was demonstrated by 17 when
SUVmean for the tumor was 0.59 ± 0.32 and SUVmean for the muscle was 0.07 ± 0.02. The uptake
in the kidneys was low compared to other compounds (SUVmean for the kidneys = 2.87 ± 0.66).
The lowest TꢀtoꢀB ratio (0.9) was demonstrated by 27 (SUVmean for the tumor = 0.18 ± 0.07 and
SUVmean for the muscle = 0.21 ± 0.05). In general, the PET imaging properties were in an
excellent agreement with the obtained internalization results in vitro. In the first set of
compounds, the most promising one with the highest binding affinity and internalization, 14,
showed also one of the highest uptakes in the tumor and TꢀtoꢀB ratio (SUVmean for the tumor =
0
.58 ± 0.20 and TꢀtoꢀB ratio = 5.4). On the other hand, the less favorable compound on these
parameters, 11, demonstrated a TꢀtoꢀB ratio of 3.8 while the SUVmean for the tumor was 0.30 ±
.09. In the second set of compounds, high tumor uptake was only found for 17 (SUVmean for the
0
tumor 0.59 ± 0.32) and 18 (SUVmean for the tumor 0.69 ± 0.15), the former of which also showed
the most favorable internalization in this set compared to 19 and 20 (SUVmean for the tumor 0.29
±
0.07 and 0.39 ± 0.07, respectively). In the third set of compounds, 22 exhibited the highest
internalization and also the highest tumor uptake (SUVmean for the tumor 0.52 ± 0.26) and the Tꢀ
toꢀB ratio (5.2) in this set. The TꢀtoꢀB ratio for 21 was 1.6 (SUVmean for the tumor 0.31 ± 0.06)
and for 23 2.0 (SUVmean for the tumor 0.14 ± 0.10). The last set of compounds (25–27) showed
almost no internalization and also demonstrated low SUVmean in the tumor as well as low TꢀtoꢀB
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ratios (SUVmean for the tumor 0.23 ± 0.04, 0.30 ± 0.10, and 0.18 ± 0.07, respectively; TꢀtoꢀB
ratio 1.5, 2.3, and 0.9, respectively). However, 28 with higher internalization showed also higher
tumor uptake (SUVmean for the tumor 0.54 ± 0.16) and higher TꢀtoꢀB ratio (3.9) compared to
other compounds of this set. The graphic illustrations of the time–activity curves for each
compound are stated in the Supplemental Information.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The ꢀPET images 2 hours postꢀinjection revealed variable tumor visualization as well as
clearance of the radiolabeled compounds from the kidneys and the whole body. The first set of
compounds with generally high inhibition potency and internalization into LNCaP cells
demonstrated good visualization of the tumor except for 11. The ꢀPET imaging of the second set
of compounds correlates highly with the in vitro properties. The best imaging potential was
demonstrated by 17 and 18 which have also shown the highest internalization in this set.
Compounds 19 and 20 exhibited very low tumor accumulation as well as almost no
internalization. The in vitro assays also pointed to the most promising compound of the third set,
22, which showed high in vitro binding affinity and internalization as well as high in vivo
accumulation in the tumor. The last set of compounds (25–28) with the overall worst in vitro
properties exhibited also the worst in vivo imaging properties. The ꢀPET coronal slices are
68
depicted in Figure 3 for the first set of Gaꢀlabeled compounds (11–16), in Figure 4 for the
second set (17–20), in Figure 5 for the third set (21–23), and in Figure 6 for the fourth set (25–
28), all 2 hours postꢀinjection. The in vivo distribution is additionally demonstrated by the
respective maximum intensity projections (MIP). All ꢀPET images for each time point (20, 40,
60 and 120 min p.i.) are presented in the Supplemental Information.
ACS Paragon Plus Environment

Page 13 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
DISCUSSION AND CONCLUSIONS
In this study, we evaluated the effect of modifying the hydrophobic linker moiety between the
glutamate–urea–lysine binding motif and the DOTA chelator. DOTA is a chelating agent
applicable predominantly for the complexation of trivalent metal cations, which includes the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
4
4
68
86
67
111
90
clinically relevant radiometals Sc, Ga, and Y for PET; Ga and In for SPECT; Y and
1
77
213
225
Lu for beta therapy; and Bi and Ac for alpha therapy. The intentionally small library of
PSMA inhibitors prepared here provides a promising tool for the targeting and optimal detection,
staging, and treatment of prostate cancer, resulting in the improved management of patients
suffering from prostate cancer.
1
8
The most commonly used radiopharmaceutical for PET by far is 2ꢀdeoxyꢀ2ꢀ[ F]fluoroꢀDꢀ
1
8
glucose ([ F]FDG). FDGꢀPET can be used for the diagnosis and therapy response monitoring of
many types of cancers, but it is not effective for most primary prostate tumors and metastases
4
5–47
due to the low glycolytic rate of prostate cancer.
Both androgenꢀsensitive and androgenꢀ
1
1
18
18
independent prostate tumors have shown higher Cꢀ or Fꢀcholine uptake than that with Fꢀ
FDG. However, the detection rate with choline, especially at a PSA below 1 ng/m , is still
As a result, there has been a sustained effort for the development of radiolabeled
L
4
8–50
low.
13,15,16
51–53
peptides/peptidomimetics
and antiꢀPSMA monoclonal antibodies (mAb).
In contrast to previous approaches, several groups have designed lowꢀmolecular weight
inhibitors of PSMA with binding affinities similar to that of huJ591, but possessing faster
accumulation uptake rates by tumor tissue and with faster background clearance. The DOTAꢀ
conjugated PSMA inhibitors 11 to 28 described here can potentially be used as PET tracers after
68
radiolabeling with the readily available generatorꢀproduced positron emitter Ga, and, after
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
177
225
radiolabeling with particle emitters such as
Lu and
Ac, also as therapeutic
radiopharmaceuticals for endoradiotherapy of prostate cancer. This aimedꢀat theranostic
approach represents the main rationale for the work described here.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Thus, structural features of the enzymatic binding pocket of PSMA were used for the design of a
rational inhibitor. The pocket has a strong preference for P1 acidic residues, e.g., Asp and Glu,
and therefore the pharmacophore was proposed to present three invariant carboxylic groups able
to interact with the respective S1 and S1' pockets. In addition, the carbonyl oxygen of the urea
2
+
function coordinates to the Zn active site. Besides the electrostatic interactions of urea and
2+
carboxylic groups at the Zn active site, there are also lipophilic interactions with an accessible
3
9,41
hydrophobic pocket adjacent to the S1 subpocket.
In general, mechanistic explanations of
dissimilarities in inhibition constants for putative PSMA inhibitors highlight the importance of
the combination of P1 carboxylates, P1 linker length, and the properties of the P1 distal segment.
Differences in linker or distal groups lead to different inhibition potencies. Hydrophobic distal
functionalities and relatively short linkers engage nonꢀpolar residues in the entrance funnel
40,41
leading to higher affinities.
The recently discovered areneꢀbinding site (ABS) belongs to a
part of the entrance lid of PSMA preferring aryl functionalities, which significantly increases the
binding affinity of the respective PSMA inhibitor and thus can be exploited for structureꢀaided
5
4
inhibitor design.
5
5
PSMAꢀtargeting GluꢀNHꢀCOꢀNHꢀLys(Ahx)ꢀHBEDꢀCC (PSMAꢀ11 ) has become one of the
56
most successful radiotracers with respect to clinical application over the last few years and has
served as the benchmark for our development of modified glutamate–urea inhibitors. Due to the
fact that the chelator HBEDꢀCC is not suitable for the stable complexation of relevant
therapeutic radionuclides, we now redesigned the linker in order to match the capability of
ACS Paragon Plus Environment

Page 15 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
HBEDꢀCC to bind to the aforementioned PSMA exoꢀsites, putatively including the hydrophobic
pocket of the PSMA S1 site. It has been reported that replacement of the HBEDꢀCC chelator in
29 (PSMAꢀ11) simply by the DOTA chelator, introduced as carboxamide into the molecule,
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
57
resulted in significant decrease in the tumorꢀtargeting properties in the new compound. In
regard to this observation, the favorable tumorꢀtargeting as well as pharmacokinetic properties of
prospective DOTAꢀbased compounds can be achieved by corresponding linker alteration.
We, therefore, constructed lowꢀmolecular weight PSMA inhibitors by varying systematically the
structure of the linker. The different modifications were divided into four sets each based on the
chemical similarities in the structural changes of the linker. The first set of compounds included
different numbers and arrangement of aromatic rings in the linker area. These compounds, i.e. 11
to 16, clearly indicate the requirement of aromatic moieties between the pharmacophore and the
DOTA chelator. The optimal number of benzene rings is indicated by 14 with three aromatic
moieties in the linker and this compound served as an initial template for further optimizations.
As a consequence, the compounds of the second set, i.e. 17 to 20, including the lead candidate 17
were synthesized. Compound 17, which contains 2ꢀnaphthylꢀLꢀalanine as the first building block
of the linker, proved that at least one aromatic moiety with a rigid conformation in the linker
seems to be requisite. However, slight changes within the linker building blocks, cf. 18 with 2ꢀ
naphthylꢀDꢀalanine and 19 with 1ꢀnaphthylꢀLꢀalanine, can result in significant alteration of the in
vitro properties of the PSMA inhibitors. Despite, 20 demonstrated indeed high tumor uptake
which, however, was counteracted by high putative background activity. Moreover, the washout
of the radioactivity from the kidneys was significantly slower compared to 17. The preference
for 2ꢀnaphthylꢀLꢀalanine (17) compared to 1ꢀnaphthylꢀLꢀalanine (19) originates from the
perpendicular positioning of the ring system of the 1ꢀnaphthyl residue. This formation causes the
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
steric clash within the PSMA binding pocket which influences the overall binding affinity of the
prospective compound. Moreover, the simple sequence interchange of the preferred building
blocks 2ꢀnaphthylꢀLꢀalanine and transꢀ(aminomethyl)cyclohexanecarboxylic acid in 20 also
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
resulted in a strong deterioration of the compound’s desired pharmacological properties.
The third set of DOTAꢀconjugated PSMA inhibitors, i.e. 21 to 24, evaluates the influence of the
building block while maintaining the 2ꢀnaphthylꢀ
Lꢀalanine building block as a constant in the
linker. The results provide evidence that 2ꢀnaphthylꢀ
Lꢀalanine alone cannot afford favorable
properties as, e.g., 21 showed significantly reduced binding affinity (p = 0.0276) and
internalization (p = 0.0040). In general, an additional aromatic moiety, e.g., 22, improves the
characteristics of the inhibitor compared to the insertion of acyclic, nonꢀaromatic building
blocks, as shown by 23 with an aminohexanoic acid linker. However, 22 demonstrated
significantly lower washout of the radioactivity from the kidneys compared to 17, most probably
due to its higher lipophilicity.
The last set of compounds, i.e. 25 to 28, examines the importance of the cyclohexane moiety in
the linker and proves again the need for additional aromatic moieties between the
pharmacophore and the DOTA chelator. Without any aromatic moieties, significant reduction of
both PSMA inhibition potency and internalization into LNCaP cells was observed. Thus,
aromatic moieties can be considered as crucial for efficient tumor targeting of PSMA inhibitors.
This assertion was amply demonstrated by compounds 25, 26, and 27. Additionally, the µPET
imaging for 27 revealed high background activity, especially in the lungs and brain, which refers
to the aggregation of the compound in vivo. With the insertion of a benzene moiety, e.g.
(aminomethyl)benzoic acid, in addition to a cyclohexane moiety, the internalization at least is
regained in this set of compounds (28). This compound demonstrated also high tumor uptake,
ACS Paragon Plus Environment

Page 17 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
however, the clearance from the tumor as well as other organs (e.g. kidneys) was very fast within
a short time period.
In general, we observed significant differences between individual compounds which are
dependent on the nature of the linkers. Especially the PSMA inhibition potency, the cellular
internalization, and also the in vivo imaging quality, can be beneficially influenced by careful
selection of the linker moiety. Coronal scans obtained from dynamic PET measurements showed
differential tumor uptake, background activity, and clearance from the kidneys. Moreover, there
are no significant interactions of the PSMA inhibitors with serum proteins which may affect
disadvantageously their biodistribution and rate of excretion from the body. Significantly, the in
vitro results, especially the internalization of the ligands in LNCaP cells, correlate highly with
the in vivo tumor imaging properties. Thus, cellular internalization in vitro seems to be an
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
important indicator of successful tumor uptake
.
The aqueous solubility of small molecules depends on their hydrophobicity. Introduced
hydrophobic group(s) in the linker may interfere with the PSMA targetꢀinhibitor interaction as
the determined logD values of these PSMA inhibitors clearly show that the lipophilicity
influences both clearance by the kidneys and the background activity. In order to select the most
promising compound, SUVmean values in LNCaP tumor, muscle (representing background
activity), and kidneys were compared by analysis of the ꢀPET scans. The best combination of all
desirable properties was possessed by 17. While compounds 14, 16, and 18 showed similar, or
even higher, tumor uptake than 17, they also exhibited higher background activity as well as
increased kidney uptake. Thus, 17 displayed the highest TꢀtoꢀB ratio (8.5) 1 hour postꢀinjection.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Its fast renal clearance of radioactivity is important to reduce potential radiation toxicity to the
kidneys as a possible chronic longꢀterm side effect of PSMAꢀspecific endoradiotherapy.
By modifying the linker between the pharmacophore and the DOTA chelator, a series of 18 new
glutamate–ureaꢀbased PSMA inhibitors (11 to 28) were synthesized. The pharmacokinetic
properties, including PSMA inhibition potencies, cellular internalization, and biodistribution
behavior of the DOTAꢀconjugates can be significantly influenced by modification of the linker.
Thus, the chemical constitution of the linker moiety has a significant impact on the in vivo
tumorꢀtargeting and pharmacokinetic properties of PSMAꢀtargeting radioligands. The approach
presented here could potentially lead to a more accurate rational and structureꢀaided design of
new PSMAꢀtargeting radioligands based on the glutamate–urea motif. The ultimate aim is the
introduction of the best of these DOTAꢀconjugated PSMA inhibitors, viz. 17, as a theranostic
radioligand, i.e. a molecular vector which is applicable to both diagnostics and the
endoradiotherapy of patients suffering from prostate cancer.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment

Page 19 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
EXPERIMENTAL SECTION
Chemicals. For synthesis and analysis, all chemicals (>95% purity; for radiolabeling, highly
pure chemicals) and solvents (HPLCꢀgrade purity; for radiolabeling, metalꢀfree ultraꢀpure water)
were purchased commercially from CheMatech (Dijon, France), Iris Biotech (Marktredwitz,
Germany), Merck (Darmstadt, Germany), Sigma–Aldrich (Taufkirchen, Germany), and VWR
International (Bruchsal, Germany) and used as received without any further purification or
drying.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
General Synthesis Outline. The solidꢀphase synthesis of the peptidomimetic glutamate–urea–
lysine binding motif is summarized in Scheme 1. The ensuing coupling of the linker was
performed according to standard fluorenylmethoxycarbonyl (Fmoc) protocol (Scheme 2).
Finally, conjugation of the DOTA chelator (Scheme 2) was realized using HBTUꢀactivated
5
7,58
DOTAꢀtris(tꢀBu)ester or DOTAꢀNHS ester.
2
ꢀChloroꢀtritylresin (0.3 mmol, substitution capacity 1.22 mmol/g, 100–200 MESH) was first
agitated in dry dichloromethane (DCM) for 45 min and then washed with dry DCM followed by
reaction with 1.2 equiv of Alloc (εꢀallyloxycarbonyl), Fmoc protected ꢀlysine [Fmocꢀ
Lys(Alloc)ꢀOH, 1], and 4.8 equiv of Nꢀethyldiisopropylamine (DIPEA) in 3 m of dry DCM.
L
L
The coupling of the first protected amino acid on the resin proceeded over the course of 16 h
with gentle agitation.
The lysineꢀimmobilized resin 2 was then washed with DCM with unreacted chlorotrityl groups
remaining in the resin blocked with a mixture of DCM, methanol (MeOH), and DIPEA in a ratio
of 17:2:1, respectively. Selective removal of the Fmocꢀprotecting group was realized by washing
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
with a mixture of dimethylformamide (DMF) and piperidine in a ratio of 1:1 once for 2 min and
then once again for 5 min in order to get product 3.
In the next step, 10 equiv of bis(tꢀBu)ꢀ
were used for generation of the isocyanate of the glutamyl moiety ii. An appropriate amount of tꢀ
Buꢀprotected glutamate was dissolved in 200 m of dry DCM followed by, shortly afterwards,
the addition of 3 m of DIPEA. This solution was added dropwise over 4 h to a flask with 1
mmol of iceꢀcooled bis(trichloromethyl)carbonate (Caution: BTC, also named triphosgene, is
toxic and fatal if inhaled) in 5 m of dry DCM. The lysineꢀimmobilized resin 3 was added
Lꢀglutamate hydrochloride [HꢀGlu(OtꢀBu)ꢀOtꢀBuꢀHCl, i]
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
L
L
L
afterwards in one portion to the solution of the isocyanate of the glutamyl moiety and stirred for
16 h.
The product 4 was filtered off and washed with DCM. The cleavage of the Allocꢀprotecting
group was realized by reaction with 0.3 equiv of tetrakisꢀ(triphenyl)palladium (TPP palladium)
in the presence of 30 equiv of morpholine in dry DCM for 2 h. The reaction was performed in
the dark using aluminum foil. To remove residuals of the palladium catalysator, the resin was
washed with 1% DIPEA in DMF and subsequently with a solution of sodium
diethyldithiocarbamate trihydrate (15 mg/mL) in DMF (10 times for 5 min). The resinꢀ
immobilized and tꢀBuꢀprotected binding motif 5 was dried under vacuum and split into three
portions.
Relative to the resin (0.1 mmol), 4 equiv of the first Fmocꢀprotected building block were
activated with 3.92 equiv of HBTU (Oꢀ(benzotriazolꢀ1ꢀyl)ꢀN,N,N',N'ꢀtetramethyluronium
hexafluorophosphate) in the presence of 4 equiv of DIPEA in dry DMF. Two min after the
addition of DIPEA, the solution was added to the DMF preꢀswollen immobilized pharmacophore
ACS Paragon Plus Environment

Page 21 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
5
and agitated for 1 h. Selective removal of the Fmocꢀprotecting group from the product 6 was
realized by washing with a mixture of DMF and piperidine in a ratio of 1:1 once for 2 min and
then once again for 5 min in order to obtain the product 7. Depending on an individual
compound, 4 equiv of a second, third, and finally fourth Fmocꢀprotected building block was
reacted with 3.92 equiv of HBTU and 4 equiv of DIPEA in dry DMF followed by the
deprotection of the Fmocꢀprotecting group as per above.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The conjugation of the chelator to the immobilized aminoꢀfunction of the linker in 7 was
performed by either of two different methods. The first strategy used 3.95 equiv HBTUꢀactivated
4
equiv
tetraazacyclododecanꢀ1ꢀyl)acetic acid) and 4 equiv of DIPEA in DMF in order to obtain product
. The final product, i.e. the series of 18 PSMA inhibitors 11–28, was obtained by agitation and
subsequent cleavage from the resin within 2 h with a mixture consisting of trifluoroacetic acid
TFA), triisopropylsilane, and water in a ratio of 95:2.5:2.5, respectively. The second strategy
of
DOTAꢀtris(tꢀBu)ester
(2ꢀ(4,7,10ꢀtris(2ꢀ(tꢀbutoxy)ꢀ2ꢀoxoethyl)ꢀ1,4,7,10ꢀ
8
(
involved preferential cleavage of the tꢀBuꢀprotected 7 from the resin using 30% 1,1,1,3,3,3ꢀ
hexafluoroisopropanol (HFIP) in DCM. After reversedꢀphase highꢀperformance liquid
chromatography (RPꢀHPLC) purification of the product 9, 1.5 equiv of DOTAꢀNHS ester
(2,2',2''ꢀ(10ꢀ(2ꢀ((2,5ꢀdioxopyrrolidinꢀ1ꢀyl)oxy)ꢀ2ꢀoxoethyl)ꢀ1,4,7,10ꢀtetraazacyclododecaneꢀ
1,4,7ꢀtriyl)triacetic acid) and 1.5 equiv of N,N,Nꢀtriethylamine (TEA) were added and reacted for
16 h in dry DMF. The mixture of solvents from the product was evaporated and the crude
compound was dissolved in acetonitrile (ACN) and water in a ratio of 1:1 for subsequent HPLC
purification. After HPLC purification, the remaining tꢀBu groups in compound 10 were cleaved
within 2 h using TFA at room temperature. TFA was evaporated and the crude compound 11–28
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
was dissolved in acetonitrile (ACN) and water in a ratio of 1:1 followed with subsequent
lyophilization.
General Analytical and Purification Methods. All synthesized compounds (Figure 2) were
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
first purified by RPꢀHPLC and then analyzed by matrixꢀassisted laser desorption ionization mass
1
13
spectrometry (MALDIꢀMS), 1D H and C nuclear magnetic resonance (NMR), and 2D NMR
techniques, viz. correlation spectroscopy (COSY), heteronuclear single quantum correlation
(HSQC), and heteronuclear multiple bond correlation (HMBC).
Both analytical (100 × 4.6 mm) and semiꢀpreparative (100 × 10.0 mm) RPꢀHPLC columns
utilized Chromolith RPꢀ18e stationary phase (Merck, Darmstadt, Germany) with mobile phases
consisting of 0.1% TFA in water (A) and 0.1% TFA in ACN (B). For analytical runs, a linear
gradient of solvent A (100–0% in 6 min) in solvent B at a flow rate of 4 mL/min was used. For
semiꢀpreparative runs, a suitable gradient was determined individually for each compound. All
compounds were obtained in >98% purity as determined by HPLC and NMR.
MALDIꢀMS was performed on a Daltonics Microflex system (Bruker Daltonics, Bremen,
Germany). Compounds were mixed with either sinapinic, gentisic, or αꢀcyanoꢀ4ꢀ
hydroxycinnamic acid (CCA) and applied to a MALDI target. MS were acquired over the range
+
+
+
4
00–2000 Da with pseudomolecular [M + H] , [M + Na] , or [M + K] ions detected.
NMR (nuclear magnetic resonance) spectra were acquired using a Bruker Avance II NMR
spectrometer equipped with a 5ꢀmm inverseꢀconfiguration probe with tripleꢀaxisꢀgradient
1
13
capability at a field strength of 14.1 T operating at 600.1 and 150.9 MHz for H and C nuclei,
respectively. Spectra were also acquired using a Bruker Avance III NMR spectrometer equipped
with a 5ꢀmm, normalꢀconfiguration probe with zꢀaxis gradient capability at a field strength of 9.4
ACS Paragon Plus Environment

Page 23 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
13
T operating at 400.1 and 100.6 MHz for H and C nuclei, respectively, using the lower field to
combat dynamic effects. Spectra were acquired at 30, 70, or 90 °C where indicated to further
combat dynamic effects with samples contained in d ꢀDMSO. Pulse widths were calibrated
6
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
59
1
13
following the described protocol. The chemical shifts of H and C nuclei are reported relative
1
13
to TMS (
δ = 0 ppm for both H and C) using the solvent signals as secondary internal
1
13
1
references (
δ
= 2.50 ppm for H and δ = 39.52 ppm for C). The chemical shifts of H nuclei are
reported to three decimal places when the spin system was amenable to firstꢀorder analysis or to
two decimal places when the multiplet was beyond such interpretation, the overlap of signals
precluded such, or chemical shifts were taken from 2D NMR spectra–particularly in the case of
1
signals obscured by the water signal. The signal intensity of H NMR signals were set to the
nearest integer value of measured integral values; for overlapped signals and signals obscured by
the water signal, intensity was apportioned according to HSQC/DEPT results. The chemical
1
3
shifts of C nuclei are reported to two decimal places when sharp signals were observable
directly or one decimal place when the peak was broad or the signal was observed indirectly.
1
13
General NMR experimental and acquisition details for 1D H, C, and DEPT observation and
standard gradientꢀselected 2D COSY, HSQC, and HMBC spectra and routine chemical shift
6
0–63
64,65
assignment using 2D NMR have been previously described.
UDEFT
spectra were
6
4
acquired and processed as described with Acq and PAD times of 0.2 and 4 s, respectively, and
1
3
with linear prediction of the FIDs; otherwise parameters were the same as normal
C
acquisition. Partial and tentative assignments of compounds, for which 1D NMR spectra (singleꢀ
1
13
13
pulse acquisition for H and C nuclei and DEPT acquisition for C nuclei) only were acquired,
were accomplished by comparative analysis to fully assigned structures (12, 14, 17, and 19) and
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
by iterative calculation of chemical shifts using the chemical shift prediction routine within the
ChemDraw 12.0 program.
6
8
68
Radiolabeling. Positronꢀemitting Ga [E
= 1.9 MeV (88%)] was eluated from a Ge (t =
β+,max
½
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
68
66
67
68
2
71 d)/ Ga (t = 68 min) generator based on pyrogallol resin support as [ Ga]GaCl in 0.1
½
3
1
77
–
–
M HCl. The radioisotope Lu (t = 6.71 days) β ꢀemitting radionuclide [E
= 497 keV
β ,max
½
66
(
79%)] with few γꢀemissions [Eγ/x = 113 keV (6%), 208 keV (11%)] was purchased from
177
PerkinElmer (Rodgau, Germany) as [ Lu]LuCl in 0.05 M HCl.
3
6
8
68
3+
Radiolabeling with Ga was accomplished using 30 ꢀ
L of [ Ga]Ga eluate (~100 MBq) with
7
0 ꢀ of 2.4 M HEPES (4ꢀ(2ꢀhydroxyethyl)ꢀ1ꢀpiperazineethanesulfonic acid, pH 7.5). This
L
solution was adjusted with 10 or 30% NaOH to pH 4.2. Subsequently, 0.5–1 nmol of compound
diluted in 0.1 M HEPES was added and the reaction mixture was heated at 95–98 °C for 15 min.
The labeling efficiency was determined by RPꢀHPLC as per above or by RPꢀTLC (RPꢀthinꢀlayer
chromatography) over silica gel plates (60 RPꢀ18 F254S) using 0.1 M sodium citrate as the mobile
68
phase. Gaꢀlabeling resulted in a labeling efficiency >95% for all tested PSMAꢀinhibitors. The
6
8
specific activities for Gaꢀlabeled compounds were in the range of 100–200 GBq/ꢀmol.
1
77
177
LuꢀLabeling was performed with 5 ꢀ
L
of [ Lu]LuCl (~25 MBq), 112 ꢀ
L
of 0.4 M sodium
3
acetate (pH 5.0), and 0.5–1 nmol of compound in 0.1 M HEPES at 95 °C for 20 min. The
1
77
specific activities for Luꢀlabeled compounds were in the range of 25–50 GBq/ꢀmol and the
labeling efficiency was >98% for all compounds.
Assays. For in vivo assays, radiolabeled compounds were additionally purified via SepꢀPak C18
Plus Light cartridges (Waters, Milford, USA). After dilution with 500 ꢀL of phosphate buffered
saline (PBS), the reaction mixture was applied directly to and trapped on a preconditioned Sepꢀ
ACS Paragon Plus Environment

Page 25 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Pak cartridge equilibrated with PBS. The cartridge was subsequently flushed with 1 mL of PBS
followed by 10 mL of air. The pure compound was then eluted with 70% ethanol and evaporated
to dryness. The radiochemical yield (RCY) for all labeled compounds then exceeded >99%.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The potential binding to the serum proteins was performed with seven representative
TM
compounds, viz. 11, 14, 17, 22, 26, and 28, and analyzed on a Superdex 75 5/150 GL column
(GE Healthcare, Munich, Germany) after 1 h incubation at 37 °C. The mobile phase consisted of
PBS (pH 7.4) and the runs were performed at a flow rate of 0.4 m /min. The lipophilicities of all
L
compounds were determined by their equilibrium distribution after thorough shaking in a twoꢀ
phase system consisting of nꢀoctanol and 0.1 M HEPES (pH 7.4). Small aliquots from both
phases were collected and analyzed in a gamma counter in order to calculate distribution
coefficients (logD, Equation 1).
+
In vitro experiments were performed using the PSMA LNCaP cell line (European Collection of
Cell Cultures, Salisbury, UK) derived from an androgenꢀsensitive human lymph node metastatic
lesion of prostatic adenocarcinoma (ATCC CRLꢀ1740). The cells were grown in RPMI 1640
medium (PANꢀBiotech, Aidenbach, Germany) supplemented with 10% fetal calf serum (FCS)
and 1% Lꢀglutamine. Cell cultures were maintained in an atmosphere of 5% CO at 37 °C in a
2
humidified incubator. For in vivo experiments (ꢀPET imaging), sixꢀ to eightꢀweekꢀold male
BALB/c nu/nu mice (The Jackson Laboratory, Maine, USA) were implanted subcutaneously in
6
the right trunk with LNCaP cells (6 × 10 ) in 100 ꢀ
L
of 50% Matrigel (Discovery Labware, Inc.,
Bedford, MA, USA) in OptiꢀMEM (Life Technologies GmbH, Darmstadt, Germany). Mice were
imaged when the tumor xenografts reached approximately 1 cm in diameter whereupon the
radiolabeled compound was injected via the tail vein (~30 MBq, 0.5 nmol, specific activity ~60
GBq/ꢀmol).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
68
The binding affinity was determined using a cellꢀbased competitive assay with Gaꢀlabeled
5
8
[
GluꢀNHꢀCOꢀNHꢀLys(Ahx)] ꢀHBEDꢀCC (PSMAꢀ10 ) and expressed as K values. Compounds
2 i
at twelve different concentrations (0, 0.5, 1, 2.5, 5, 10, 25, 50, 100, 500, 1000 and 5000 nM)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
68
5
were incubated with 0.75 nM of Gaꢀ30 (PSMAꢀ10) together with LNCaP cells (10 LNCaP
cells/well). After 45 min incubation at 37 °C, the cells were washed three times with of iceꢀcold
PBS. Radioactivity accumulated in the cells was measured by a gamma counter and the data
fitted using a nonlinear regression algorithm (GraphPad Software) to calculate 50% inhibitory
concentrations (IC values).
5
0
5
For the determination of internalization, LNCaP cells (10 LNCaP cells/well) were seeded in
polyꢀ ꢀlysine coated plates 24 h before the experiment. The cells were incubated for 45 min at 37
C with 32 nM of Luꢀlabeled compound (25 GBq/ꢀmol) in 250 ꢀ
L
1
77
°
L OptiꢀMEM. Additionally,
for confirmation of specific cellular uptake, the second set of compounds was also treated with
500 ꢀM/well of PSMAꢀinhibitor 2ꢀPMPA (2ꢀ(phosphonomethyl)pentaneꢀ1,5ꢀdioic acid). The
cells were washed four times with iceꢀcold PBS and the surfaceꢀbound radioactivity was
removed by washing twice with 50 mM glycine (pH 2.8). After washing the cells with PBS, the
internalized fraction was determined by lysis of the LNCaP cells using 0.3 M NaOH. The
radioactivity collected from the glycine and hydroxide fractions was measured in a gamma
6
counter and calculated as %IA/10 LNCaP cells. Additionally, the specific cell surface binding
and specific internalization were calculated by subtracting the nonꢀspecific cell surface binding
6
and the nonꢀspecific internalized fraction and expressed again as %IA/10 LNCaP cells.
6
8
Anaesthetized mice (2% sevoflurane, Abbott) were injected with Gaꢀlabeled compound and
placed in a small animal PET scanner (Inveon PET, Siemens). Each examination was performed
ACS Paragon Plus Environment

Page 27 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
68
up to 2 h postꢀinjection and included a 15 min transmission scan before the injection of Gaꢀ
labeled compound, a 60 min emission scan, a second transmission scan, and a final emission
scan 100 to 120 min postꢀinjection. For testing each compound, one exemplary animal was used.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In vitro experiments (inhibition potency, surfaceꢀbinding, and internalization) were performed in
68
triplicate. Calculated IC values were converted into K values by the Cheng–Prusoff method
5
0
i
(Equation 2) which express the inhibition potency of the respective compound, using the
68
experimentally determined K value of Gaꢀ30 (3.8 ± 1.8 nM). Data were expressed as means ±
d
SD (standard deviation for confidence interval) where p values <0.05 were considered
statistically significant. p Values were calculated as unpaired twoꢀsided tꢀtest assuming the
gaussian distribution. Scans from ꢀPET imaging were reconstructed iteratively using the space
alternating generalized expectation maximization method (SAGE, 16 subsets, 4 iterations)
applying median root prior correction and were converted to standardized uptake value (SUV)
images. Quantitation was accomplished using an ROI technique and expressed as SUVmean.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 56
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ANCILLARY INFORMATION
Supporting Information
NMR data and spectra for all synthesized DOTAꢀconjugated PSMAꢀinhibitors. This material is
available free of charge via the Internet at http://pubs.acs.org.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Corresponding Author
*M.E.: phone/fax, +49 6221 56 7127; Eꢀmail, m.eder@dkfzꢀheidelberg.de.
Author Contributions
⊥
K.K. and M.E. contributed equally to this work and share senior authorship.
Notes
The authors declare no competing financial interest.
Acknowledgements
Financial support for a Helmholtz International Graduate School for Cancer Research Ph.D
stipend for M.B., a DFG grant for M.E. (ED234/2ꢀ1), and a Klaus Tschira Foundation grant for
U.H. are all gratefully acknowledged. All animal experiments complied with the current laws of
the Federal Republic of Germany and were conducted according to German Animal Welfare
guidelines. Biodistribution and ꢀPET imaging studies were kindly performed by Ursula
Schierbaum and Karin Leotta from the German Cancer Research Center, Heidelberg, Germany.
The authors would like to also thank to Lukas Welti for the initial realization of the topic.
ACS Paragon Plus Environment

Page 29 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Abbreviations used
PSMA, prostateꢀspecific membrane antigen; HPLC, highꢀperformance liquid chromatography;
TLC, thinꢀlayer chromatography; Glu, glutamate; Lys, lysine; TFA, trifluoroacetic acid; DOTA,
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
,4,7,10ꢀtetraazacyclododecaneꢀ1,4,7,10ꢀtetraacetic acid; HBEDꢀCC, N,N'ꢀbisꢀ[2ꢀhydroxyꢀ5ꢀ
(carboxyethyl)benzyl]ethylenediamineꢀN,N'ꢀdiacetic acid; 2D, two dimensional; 1D, one
dimensional; ꢀ, micro; °C, degrees Celsius; aa, amino acid; FDG, 2ꢀfluoroꢀ2ꢀdeoxyꢀ ꢀglucose;
D
Ala, alanine; COSY, correlation spectroscopy; DCM, dichloromethane; DMSO,
dimethylsulfoxide; Fmoc, 9ꢀfluorenylmethylmethoxycarbonyl; FT, Fourier transform; HSQC,
heteronuclear single quantum correlation; HMBC, heteronuclear multiple bond correlation;
UDEFT, uniform driven equilibrium Fourier transform; IC , halfꢀmaximum inhibitory
50
concentration; J, coupling constant; K , inhibition constant; logD, logarithm of distribution
i
coefficient; m/z, massꢀtoꢀcharge ratio; MALDIꢀMS, matrixꢀassisted laser desorption ionization
mass spectrometry; NMR, nuclear magnetic resonance; R , retention factor; rt; room temperature;
f
tꢀBu, tertꢀbutyl; PET, positron emission tomography; SPECT, single photon emission computed
tomography; SUV, standardized uptake value; DMF, dimethylformamide; MIP, maximum
intensity projections; SAR, structure–activity relationships; RCY, radiochemical yield; AMBA,
4ꢀ(Fmocꢀaminomethyl)benzoic acid; Ahx, 6ꢀ(Fmocꢀamino)hexanoic acid; AMCH, transꢀ4ꢀ
(Fmocꢀaminomethyl)cyclohexanecarboxylic acid; br, broad; d, doublet; ho, higher order; m,
multiplet; ol, overlapped; t, triplet; qt, quartet; s, singlet; v, very.
ACS Paragon Plus Environment