Synthesis of a Gemcitabine Prodrug for Remote Loading into Liposomes and Improved Therapeutic Effect

Article abstract of DOI:10.1021/acs.bioconjchem.5b00619

The chemotherapeutic gemcitabine was actively and stably loaded into lipid nanoparticles through the formation of a prodrug. Gemcitabine was chemically modified to increase the lipophilicity and introduce a weak base moiety for remote loading. Several derivatives were synthesized and screened for their potential to be good liposomal drug candidates for remote loading by studying their solubility, stability, cytotoxicity, and loading efficiency. Two morpholino derivatives of GEM (22 and 23) were chosen as the preferred prodrugs for this purpose as they possessed the best loading efficiencies (100% for drug-to-lipid ratio of 0.36 w/w). This is a considerable improvement over a passive loading strategy where typical loading efficiencies are on the order of ～10-20% for a drug-to-lipid ratio of ～0.01. Liposomes loaded with these two prodrugs were studied in an s.c. tumor model in vivo and showed improved therapeutic effect over free GEM (～2-fold) and saline control (8- to 10-fold). This work demonstrates how chemical modification of a known hydrophilic drug can lead to improved loading, stability, and drug delivery in vivo.

Full text of DOI:10.1021/acs.bioconjchem.5b00619

Subscriber access provided by CMU Libraries - http://library.cmich.edu
Article
Synthesis of a gemcitabine pro-drug for remote
loading into liposomes and improved therapeutic effect
Jonathan May, Elijus Undzys, Aniruddha Roy, and Shyh-Dar Li
Bioconjugate Chem., Just Accepted Manuscript • DOI: 10.1021/acs.bioconjchem.5b00619 • Publication Date (Web): 16 Dec 2015
Downloaded from http://pubs.acs.org on December 23, 2015
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.
Bioconjugate 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 39
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
9
Synthesis of a gemcitabine prodrug for remote
loading into liposomes and improved therapeutic
effect
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
,
†,‡
‡
†,‡
†,‡
Jonathan P. May , Elijus Undzys , Aniruddha Roy and Shyh-Dar Li *
†
Faculty of Pharmaceutical Sciences, University of British Columbia, Vancouver, BC Canada
‡ Drug Discovery Program, Ontario Institute for Cancer Research, Toronto, ON Canada
KEYWORDS: Gemcitabine, prodrug, liposomal drug delivery, remote loading
ABSTRACT: The chemotherapeutic, gemcitabine, was actively and stably loaded into lipid
nanoparticles through the formation of a prodrug. Gemcitabine was chemically modified to
increase the lipophilicity and introduce a weak base moiety for remote loading. Several
derivatives were synthesized and screened for their potential to be good liposomal drug
candidates for remote loading by studying their solubility, stability, cytotoxicity and loading
efficiency. Two morpholino derivatives of GEM (22 and 23) were chosen as the preferred
prodrugs for this purpose as they possessed the best loading efficiencies (100% for drugꢀtoꢀlipid
ratio of 0.36 w/w). This is a considerable improvement over a passive loading strategy where
typical loading efficiencies are in the order of ~10ꢀ20% for a drugꢀtoꢀlipid ratio of ~0.01.
ACS Paragon Plus Environment
1

Bioconjugate Chemistry
Page 2 of 39
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
Liposomes loaded with these two prodrugs were studied in an s.c. tumour model in vivo and
showed improved therapeutic effect over free GEM (~2ꢀfold) and saline control (8ꢀ to 10ꢀfold).
This work demonstrates how chemical modification of a known hydrophilic drug can lead to
improved loading, stability and drug delivery 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
INTRODUCTION
Gemcitabine (GEM), 2’,2’’ꢀdifluoroꢀ2’ꢀdeoxycytidine, is
a
wellꢀtolerated, potent
1
chemotherapeutic. However, part of the reason for its low systemic toxicity is the fact that it is
readily metabolised to the less effective difluoroꢀuridine derivative and undergoes rapid renal
2
ꢀ3
elimination. This has led us and others to investigate whether liposomal delivery of GEM
could lead to a formulation which was wellꢀtolerated, but more stable to degradation. Liposomes
are nanoꢀsized lipid vesicles consisting of a lipid bilayer surrounding an aqueous core. Their use
in drug delivery stems from the improved pharmacokinetics and biodistribution, partly due to the
4
EPRꢀeffect, which has been described for tumours, whereby the rapidly growing tumour forms
leaky vasculature, allowing increased uptake for nanoparticles in the 50ꢀ200 nm range.
5
ꢀ6
Liposomal encapsulation of GEM has often been performed via a passive loading strategy,
taking advantage of the drug’s high aqueous solubility (22.3 mg/mL). Passive loading performed
by ourselves and others for liposomes of a size suitable for EPRꢀeffect (~100 nm) has delivered
encapsulation efficiencies of around 10% for solutions of 10ꢀ20 mg GEM/mL, leading to final
liposomal GEM concentrations of 1ꢀ2 mg/mL and a corresponding drugꢀtoꢀlipid ratio of 0.01ꢀ
0.02. However, loading efficiency is generally much lower for passive loading than for a remote
7
loading strategy, a method which could also increase the stability of the drug in the liposome.
Remote loading, refers to where the drug is “actively loaded” into the liposome core via a pH or
ACS Paragon Plus Environment
2

Page 3 of 39
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
9
ion gradient (e.g. ammonium sulfate) across the liposome membrane, which is best represented
®
8ꢀ9
in the case of Doxorubicin (DOX) with the DOXIL liposome formulation (Figure 1).
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
Figure 1: Schematic for remote loading with a weak base drug (e.g. DOX).
Crucially DOX contains a weak base amine which becomes protonated once inside the
ammonium sulfate containing liposome core, forming a sparingly soluble DOX sulfate salt that
crystallizes out while liberating ammonia gas that can escape through the liposome membrane.
GEM however, is lacking some of the key physical properties of DOX that make it suitable for
this remote loading strategy. For example, the pKa is higher for DOX (primary amine pKa 8.9),
relative to GEM (pyrimidine pKa 3.6). This means that the pH inside the liposomes needs to be
strongly acidic (pH <3) to protonate GEM for drug entrapment, and this low pH could increase
lipid hydrolysis and reduce the stability of the liposomes. A study of the logP also reveals that
GEM is considerably less lipophilic (logP = ꢀ1.5) than DOX (logP = 1.4) which suggests GEM
has relatively poor membrane permeability potentially hindering its permeation through the
liposomal bilayer into the inner core of the liposomes for drug loading (logP values were
predicted by ChemAxon). These differences in physical properties suggest GEM would be a
ACS Paragon Plus Environment
3

Bioconjugate Chemistry
Page 4 of 39
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
poor choice for remote loading and may explain why attempts to remotely load GEM into
liposomes have not produced significant improvements over that found for passive loading; in
1
0ꢀ12
many cases the loading efficiency has remained ~20%
and in a few cases where the
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
encapsulation of GEM has been quoted as much higher (~90%), the drugꢀtoꢀlipid ratios (w:w)
1
3
were extremely low: 0.013ꢀ0.015 (see review and articles cited therein). In summary, liposomal
formulations of GEM seem to be limited by low drug loading and/or low drugꢀtoꢀlipid ratios.
Most recently, a method has been described which attempts to improve both the loading
1
0
efficiency and drugꢀtoꢀlipid ratio by using passive small volume incubation, leading to
encapsulation efficiencies of ~33% and a drugꢀtoꢀlipid ratio of 0.04 (w:w). This is an
improvement, but it should be noted that this still falls short of the near quantitative remote
loading at high drugꢀtoꢀlipid ratio (>0.05) observed for liposomal formulations of drugs such as
DOX.
Another technique used to improve the pharmacokinetics of GEM has been the formation of
GEM prodrugs, usually with lipophilic chains added to the N4 position, such as stearoyl and
squalene, which in some cases are used for insertion into hydrophobic nanoparticles or liposome
1
4ꢀ22
membranes.
There is also a case where short isoprene polymers have been grown from a
2
3
GEM anchor. This strategy also provides chemical protection of the N4 position of the cytosine
base, which is susceptible for the rapid hydrolysis to uracil via cytidine deaminase present in
body fluids. Other prodrugs have been investigated including a recent example of a theranostic
24
GEM derivative with a redox sensitive disulfide linker and RGD targeting moiety, a PEGylated
2
5
26
folateꢀtargeted GEM derivative and an antibody conjugated derivative.
Here we investigated whether prodrugs of hydrophilic drugs, such as GEM, can enable their
remote loading with a high efficiency and stable encapsulation. We hypothesized that for a
ACS Paragon Plus Environment
4

Page 5 of 39
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
9
hydrophilic drug, modifications may be necessary to temporarily increase the logP to a value
more suitable for passage across a lipid membrane (logP >0). To achieve this, we decided to
modify the N4 position with a small alkyl group to increase the prodrug lipophilicity. A weak
base appendage was then conjugated to a 3’ or 5’ hydroxyl via a cleavable ester linkage to form a
difunctionalized GEM derivative with suitable physical properties for remote loading.
Modifications in these positions should be labile to enzymes – for the N4 modification a
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
2
7
carboxylesteraseꢀ2 has been described, and for the 5’ꢀester there are myriad examples of
2
8
primary ester hydrolysis in many cells including those of tumour tissue. We then sought to
investigate the solubility, stability, cytotoxicity, liposomal loading, and preliminary efficacy of
the liposomal drug to elucidate the validity of this technique for improving the delivery of GEM.
ACS Paragon Plus Environment
5

Bioconjugate Chemistry
Page 6 of 39
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
Synthesis of gemcitabine derivatives
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
Six derivatives of GEM were synthesized incorporating both a lipophilic tail attached to the
cytosine base to increase lipophilicity and an ionizable weak base appendage attached to the 5’ꢀ
hydroxyl of the furanose ring to enable remote loading into liposomes. These derivatives differed
by the identity of the weak base (Nꢀmethylꢀpiperazine, or morpholine) and the length of the
carbon chain linker between the 5’ꢀhydroxyl point of conjugation and the weak base moiety.
Scheme 1. Synthetic scheme for the preparation of weak base appendages 1ꢀ6.
Initially the 6 different weak base appendages were prepared with minor modifications to the
2
9
preparations described in the literature (Scheme 1). This afforded three morpholino and three
Nꢀmethylꢀpiperazino derivatives, each with either 2, 4 or 5 carbon chain length linkers. These
amino acids were isolated as their HCl salts in the form of white crystalline compounds (overall
yields for the two steps ranged from 70ꢀ85%).
ACS Paragon Plus Environment
6

Page 7 of 39
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
9
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
Scheme 2. Preparation of 4ꢀNꢀoctanoylꢀ2’ꢀdifluoroꢀdeoxycytidine, 8.
Addition of the lipophilic tail to the 4ꢀamino group of the nucleoside base of GEM was achieved
using a transient protection of the 3’ and 5’ꢀhydroxyls with trimethyl silane (Scheme 2). The
amide was formed through reaction with the acid chloride derivative of octanoic acid. Following
removal of the transient protection and a number of aqueous washing steps, including a wash
with a dilute solution of phosphoric acid, the derivative was acquired in good yield and relatively
high purity even without a purification step (90%). However, purification with a short silica gel
column improved the purity and afforded the desired GEM derivative 8 in good yield (80%). The
nucleoside base modified derivative 8 was then orthogonally protected (Scheme 3) in the 5’ꢀ
position with TBDMS (97%) and at the 3’ꢀposition with THP (73%). This allowed deprotection
of the silyl group with fluoride (either as the pyridine or trimethylamine complex) to yield
3
0
compound 11 (87%). The 5’ꢀhydroxyl that was exposed during this last deprotection was then
coupled with the six different weak base appendages (1-6), and after isolation of the crude
products followed by a series of short deprotections of the acid labile THP groups with TFA,
products 18-23 were afforded. All final compounds were purified by automated silica gel
chromatography (overall yield for 5 steps 38ꢀ50%) and characterized by NMR and LCꢀMS to
confirm the desired identity and purity.
ACS Paragon Plus Environment
7

Bioconjugate Chemistry
Page 8 of 39
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
O
R1
NH
O
R1
NH
O
R1
NH
N
i) TBDMS-Cl, pyr
N
HF.NEt3
N
N
O
N
O
N
O
HO
ii) Tos-pyr, CH Cl2
Si O
HO
2
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
O
O
O
O
F
F
F
OH
F
O
O
F
F
8
10
11
O
O
OH.HCl
N
n
O
R2
1
-6
EDC, NEt , CH Cl
3
2
2
O
R1
NH
O
R1
NH
#
1
1
R1
R2
NMe
O
NMe
O
NMe
O
n
3
3
1
1
4
4
N
N
8:
9:
20:
21:
2:
3:
C H
7
15
C H
7
15
15
15
15
15
N
O
N
O
O
TFA
O
C H
N
7
N
n
n
O
O
O
C H
7
O
O
CH Cl
R2
R2
2
2
F
F
2
2
C H
7
OH
F
O
F
C H
7
18-23
12-17
Scheme 3. Synthetic route to the 5’ꢀweak base modifiedꢀ4ꢀNꢀoctanoylꢀ2’ꢀdifluoroꢀdeoxycytidine
derivatives, 18-23.
Evaluation of GEM prodrug solubility
Once the compounds were in hand, they were tested for aqueous solubility over a range of pH
(
pH 3, pH 4, pH 5, pH 6, pH 7). Compounds were readily soluble at lower pHs (pH 3ꢀ5) at >20
mg/mL (the maximum concentration tested), but a solubility limit of ~10 mg/mL was reached at
pH 6ꢀ7.
ACS Paragon Plus Environment
8

Page 9 of 39
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
9
Testing the stability of gemcitabine derivatives
All final compounds were tested for their stability in the neutral and weakly acidic buffers used
for drug loading. To test the stability of each of the GEM derivatives (18-23) a study was
performed at 37°C with buffers at pH 5 and pH 7; two relevant pHs for liposome loading and
stability. Compounds were incubated for up to 24 h in these buffers and samples were checked
for the degree of decomposition by LCꢀMS (Figure S1).
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
All samples studied were significantly more stable at pH 5 than at neutral pH. In fact, most
compounds showed excellent stability at pH 5, with the majority still remaining greater than 90%
intact by the end of the 24 h study period. The 2 carbon linker with a morpholine base attached
(21) appeared to be the most labile at pH 5 (~12 % decomposition), while the 5 carbon linkers in
compounds 20 and 23 and the 2 carbon linker with Nꢀmethylꢀpiperazine (18) were essentially
stable during the study period (Figure 2). Interestingly, it appears the ester bond at the 5’ꢀ
position of the GEM derivative was particularly labile when the 4 carbon linker was used, no
matter which basic moiety was tethered to the other end (19 and 22). The ester cleavage was
most accelerated at neutral pH, and following 24 h incubation only 17% or 29% was still intact
for compounds 19 and 22, respectively.
ACS Paragon Plus Environment
9

Bioconjugate Chemistry
Page 10 of 39
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
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
Figure 2. Summary of GEM derivative (18-23) stability based on the percent remaining after
incubation at 37°C for 24 h at either (a) pH 5 (NaOAc buffer) or (b) pH 7 (HBS).
In vitro cytotoxicity (IC50) study
In order to assess the relative cytotoxicity of the GEM derivatives against tumour cells, multiple
cell viability studies were performed. Initially, the murine mammary cancer cell line EMTꢀ6 was
studied with GEM, the baseꢀmodified GEM (8) and the six weakꢀbase GEM derivatives (18-23),
with all results being repeated in at least 3 independent experiments (Table 1).
Table 1. Average IC50 values for the GEM derivatives against the EMTꢀ6 murine breast cancer
cell line from >3 separate experiments. Mean ± S.D. For the raw data please check the
supplementary information Figure S2.
ACS Paragon Plus Environment
10

Page 11 of 39
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
9
Sample
Average IC50 (nM)
GEM
30.5 ± 2.7
43.3 ± 2.7
GEM-oct (8)
p2Go (18)
p4Go (19)
p5Go (20)
m2Go (21)
m4Go (22)
m5Go (23)
166.6 ± 20.1
132.6 ± 22.7
201.4 ± 51.8
219.9 ± 76.2
146.9 ± 27.7
152.8 ± 26.0
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 IC50 values of the GEM derivatives (18-23) are higher than that found for native GEM (31
nM), correlating to a 4ꢀ to 7ꢀfold reduction in potency. However, it should be noted that the
potency for the octanoyl derivative 8 (43 nM) is almost equal to that of the native drug.
Two of the more potent GEM derivatives (22 and 23) were also studied with the multidrug
resistant cell line, EMT6ꢀAR1 (Pꢀglycoprotein overexpression) and similar IC50 values were
obtained (Table 2).
Table 2. Average IC50 values for the GEM derivatives against the EMT6ꢀAR1 multiꢀdrug
resistant murine breast cancer cell line. See supplementary information for curves, Figure S3.
Sample
Average IC50 (nM)
14.1 ± 0.5
GEM
GEM-oct (8)
m4Go (22)
m5Go (23)
39.0 ± 5.1
150.5 ± 12.7
78.5 ± 3.7
ACS Paragon Plus Environment
11

Bioconjugate Chemistry
Page 12 of 39
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
Loading of gemcitabine derivatives into liposomes
Preliminary screening of drug loading was performed with a nonꢀPEGylated liposome
(
DSPC/Chol, 55:45) prepared with an ammonium sulfate gradient and acetate external buffer
pH5) as the derivatives were found to show greater stability and solubility at this pH. Different
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
(
drugꢀtoꢀlipid ratios were screened as well as incubation time to assist in optimization of the
method. All samples were studied for drug loading using ultrafiltration to separate encapsulated
drug from nonꢀencapsulated drug, followed by LCꢀMS detection (Figure S4). This proved to be
a robust method and was able to indicate quantities of drug decomposition products if that
occurred at a significant rate during the loading procedure.
Figure 3. Loading of GEM derivatives on a small scale (100 µL) with DSPC/Chol liposomes at
different drugꢀtoꢀlipid ratios (w:w) for a) m4Go (22); and b) m5Go (23).
The initial loading data represented in Figure S4 suggested that the morpholine derivatives
m4Go (22) and m5Go (23) were loaded rapidly at high drugꢀtoꢀlipid ratios. To further investigate
ACS Paragon Plus Environment
12

Page 13 of 39
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
9
the limitations of loading for these derivatives the drugꢀtoꢀlipid ratios were increased and
experiments were performed multiple times with incubation at 60°C for 60 mins (Figure 3).
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
Figure 4. Loading of GEM derivatives on a small scale (100 µL) at different drugꢀtoꢀlipid ratios
(w:w) with a PEGylated liposome for a) p4Go (19); b) p5Go (20); c) m4Go (22); and d) m5Go
(23).
Once optimal conditions were identified, the GEM prodrugs were tested for loading efficiency
with a PEGylated liposome (DSPC/Chol/DSPEꢀPEG, 56:39:5) also prepared with an ammonium
sulfate gradient and acetate external buffer (pH 5). This is similar to the liposome used in the
®
DOXIL formulation and is known to be a long circulating liposome. As before, several drugꢀtoꢀ
lipid ratios (3:1, 4:1, 5:1; w:w) were studied in order to establish the optimal system (Figure 4).
Liposomes were incubated with drug derivative in acetate buffer at pH 5 at 60°C for at least 30
min, and all samples were then processed with ultrafiltration and followed by quantification with
LCꢀMS to calculate encapsulated drug concentrations (Figure 4). These figures clearly show that
the loading of the piperazine derivatives (19 and 20) was less effective (~40ꢀ60% loading
efficiency) than for the morpholine derivatives (22 and 23), even at the lower drugꢀtoꢀlipid ratios
(0.18 and 0.24). The morpholine derivative 22 showed particularly efficient loading (>80%) even
ACS Paragon Plus Environment
13

Bioconjugate Chemistry
Page 14 of 39
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
at a D:L of 0.3. The morpholine derivative 23 also had a good loading efficiency at lower drugꢀ
toꢀlipid ratios, but dropped to ~50% at a drugꢀtoꢀlipid ratio of 0.3.
Table 3. Physical parameters of liposomes (L22 and L23), following loading with GEM
derivatives 22 and 23, purification by dialysis and sterile filtration for use 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
Final
Liposome
Z-ave
(nm)
Polydispersity
index (PDI)
Sample
% Loading
D:L
composition
(w/w)
Liposome
DSPC/Chol/DSPE-PEG
(56:39:5)
1
00.1 ± 0.9
0.04 ± 0.00
0.06 ± 0.01
0.04 ± 0.01
-
-
pre-loading
DSPC/Chol/DSPE-PEG
L22
L23
98.4 ± 0.3
100
100
0.15
0.15
(
56:39:5)
DSPC/Chol/DSPE-PEG
56:39:5)
101.2 ± 2.0
(
Once an optimal drugꢀtoꢀlipid ratio had been found, the drug loading was scaled up (from 100
µL to 1 mL) to generate dose for in vivo studies (Table 3). A drugꢀtoꢀlipid ratio of 0.16 was
selected to achieve >90% drug encapsulation, followed by dialysis (HBS pH 7.4) to remove any
unencapsulated drug giving a final D:L of 0.15. Liposomes were routinely prepared with
diameters of 95ꢀ105 nm, low polydispersity (<0.06), and ultrafiltration/LCꢀMS was used to
confirm all drug was contained within the liposomes. The identity of the drug (22 or 23) did not
seem to have any effect on these physical parameters. Liposomal stability was studied over a
period of 1 month at 4°C and no changes were observed (supplementary info, Figure S5).
ACS Paragon Plus Environment
14

Page 15 of 39
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
9
Study of gemcitabine derivative loaded liposomes in vivo
The drug loaded liposomes (L22 and L23) were tested in vivo in a preliminary efficacy study
using the murine multiꢀdrug resistant breast cancer cell line EMT6ꢀAR1. Cells were inoculated
subcutaneously in the right flank and allowed to grow for one week before treatment was started
with one of the four formulations: Saline, free GEM (40 mg/kg), L22 (40 mg drug/kg), L23 (40
mg drug/kg). All formulations were dosed twice per week for 2 doses (day 0, 3). Tumour
volumes were calculated and plotted as average values per group (Figure 5).
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
Figure 5. Efficacy plot of GEM formulations versus the multiꢀdrug resistant tumour model
5
EMT6ꢀAR1. Tumours were inoculated with 2 x 10 cells/100 µL s.c. in the flank of each animal.
Animals were then dosed with either saline control, free GEM (40 mg/kg, dosed day 0, 3), L22
or L23 (40 mg/kg, dosed day 0, 3). Data points are the average of at least 5 mice per group ±
SEM. Values annotated with * are significant relative to Control and GEM with p < 0.01.
ACS Paragon Plus Environment
15

Bioconjugate Chemistry
Page 16 of 39
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
This cell line is particularly aggressive and saline control tumours had reached endpoint within
just one week – at this point the study was stopped and tumour volumes were compared for
statistical significance. The tumour sizes for the mice treated with the liposomal formulations
were significantly smaller at endpoint (p <0.01) relative to either saline control or free GEM
treatments.
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
In addition to tumour size, the animal body weights were recorded and plotted for each treatment
group (Figure 6). All formulations seemed to be well tolerated although there was a slight drop
in body weight for the free drug relative to other treatments. It is worth highlighting that there
was no sign of toxicity for the liposomal formulations of GEM derivatives and body weights
actually showed an increase throughout the week of study.
Figure 6. Body weight plot for the 4 formulations studied. Animals were dosed with either saline
control, free GEM (40 mg/kg, dosed day 0, 3), L22 or L23 (40 mg drug/kg, dosed day 0, 3). Data
points are the average of 5 or more animals ± SEM.
ACS Paragon Plus Environment
16

Page 17 of 39
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
9
DISCUSSION
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 development of a liposomal GEM formulation with increased stability in vivo is warranted.
Current formulations of GEM are injected at high dose and/or with long infusions, as a
significant amount can be converted to the less active fluorinated uridine analog. Liposomal
formulations of GEM have been prepared via numerous methods, although all have loading
efficiencies similar to passive loading (normally ~10ꢀ20% at a drugꢀtoꢀlipid ratio of 0.01ꢀ0.02) as
1
0
GEM does not possess suitable chemical characteristics for remote loading and therefore a
significant issue is wastage of drug (80% of the drug may be discarded during the final
purification step). Improvements to liposome drug loading efficiency and retention may be
3
1
observed with the use of a remote loading strategy, rather than a passive one. However, certain
drug characteristics are necessary for this to be a success, including appropriate logP (0ꢀ2) and
pKa (either a weak base with a pKa of ~8, or a weak acid with a pKa of ~5). Native GEM does
not fit into this category as its logP is relatively low (ꢀ1.5) and it does not have any ionizable
functional groups to allow the active loading and stable encapsulation. However, through the use
of some minor chemical modifications we sought to create a GEM prodrug with suitable
properties for remote loading.
Chemical modification of an insoluble drug with a pH dependent ionizable functional group for
the remote loading into liposomes was first reported by the Cullis group to prepare liposomal
32
docetaxel. Introduction of a weak base group to docetaxel increased water solubility at a low
pH and promoted remote loading via the ammonium gradient. The current work presented here
further explored the development of a new chemical modification strategy to remote load another
ACS Paragon Plus Environment
17

Bioconjugate Chemistry
Page 18 of 39
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
class of drug (highly water soluble, poorly membrane permeable and nonꢀionizable, i.e. GEM)
into liposomes. In order to transform GEM into a suitable candidate for remote loading, we
chemically modified the structure to increase the membrane permeability and to introduce an
ionizable functional group. Modifications to the N4 position of GEM with acyl chains have
already been discussed in the literature for preparation of GEM derivatives suitable for oral
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
1
8
dosing and we employed a similar strategy to increase the membrane permeability. Herein we
describe a method for the modification of a poorly membrane permeable and nonꢀionizable drug
to form a weakꢀbase prodrug suitable for remote loading into liposomes. It is envisaged that this
liposome encapsulated prodrug, once injected systemically into the blood stream, would rely on
the EPRꢀeffect to preferentially extravasate into tumour tissue, where the liposome may be taken
up by tumour cells through an endocytosis mechanism as observed for similar nanoparticle
9
species. Whether in the endosome/lysosome of a tumour cell or even in the tumour interstitial
2
8
space, the 5’ꢀester is likely to undergo cleavage due to the presence of endogenous esterases or
the lability of this linkage in the physiological environment (neutral pH).
Initially the synthesis of several derivatives was studied in order to ascertain the optimal weak
base and lipophilic modifications. The 4ꢀNꢀoctanoyl derivative 8 was prepared and 6 different
weak base appendages were studied with variations in chain length and base moiety (1ꢀ6), which
led to differences in solubility, stability, cytotoxicity and liposomal loading ability. Each of these
products was synthesized in good yield through a 6 step route utilizing an orthogonal protection
30
strategy. Following purification the physical and cell killing properties of these derivatives
were investigated in order to select the preferred candidates for further study in vivo.
Testing of GEM derivative solubility was performed and analyzed by HPLC, indicating that all
derivatives were found to have adequate solubility at 10 mg/mL between pH 3ꢀ7, although
ACS Paragon Plus Environment
18

Page 19 of 39
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
9
greater solubility was observed for all derivatives at lower pH (limit was reached at pH 6ꢀ7: ~10
mg/mL).
The stability at 37°C for each derivative was studied at pH 5 and pH 7 to select the optimal
loading conditions for the derivatives. For all derivatives the only major decomposition product
was that resulting from the cleavage of the ester to yield the 4ꢀNꢀoctanoyl derivative 8. The
results demonstrated that the 4 carbon linker (ie. p4Go (19) and m4Go (22)) was considerably
more labile than all of the other linkers at pH 7 (17% and 29% remaining intact after 24 h,
respectively). At pH 5 all derivatives showed considerably improved stability relative to pH 7,
but this was most pronounced for the 4 carbon linkers (~10ꢀfold improved stability for p4Go (19)
and m4Go (22) at pH 5 relative to pH 7). At pH 5 all derivatives showed less than 20%
decomposition over the 24 h period, but at pH 7 only the derivatives with 5 carbon linkers (20
and 23) were similarly stable (<20% decomposition). The derivatives with 2 carbon linkers
showed intermediate values of decomposition at pH 7 of between 40ꢀ60%. A plausible reason for
the increased lability of the 4 carbon linker is one of anchimeric assistance of the base moiety in
the transition state of the ester cleavage – possibly through the coꢀordination of a proton in the
vicinity of the δꢀnegative carbonyl oxygen. Knowledge of the derivatives’ stability could be
particularly important for optimizing the loading conditions.
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 cytotoxicity of each derivative was studied against the murine breast cancer cell line, EMTꢀ
6. Free GEM and 4ꢀNꢀoctanoyl derivative 8 were also studied for reference. While compound 8
was almost as potent as the free drug (43 and 30 nM, respectively), modification at the 5’ꢀ
position with the weak base appendage caused a 4ꢀ to 7ꢀfold drop in potency to the 120ꢀ210 nM
range. This is to be expected as these drugs will most likely not be recognized by the usual
transporters which are responsible for the uptake of free GEM which could hinder drug uptake.
ACS Paragon Plus Environment
19

Bioconjugate Chemistry
Page 20 of 39
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
Furthermore, even if the GEM derivatives were able to get into the cell intact, the modification at
the 5’ꢀposition would likely hinder the phosphorylation step in the known mechanism of action
of this drug. These multiple reasons provide good support for the increased IC50s observed,
however once the drug is encapsulated inside a liposome then the uptake could be significantly
altered, perhaps via a liposome endocytosis mechanism, which could have a significant effect on
cytotoxicity. Further studies were performed on the multiꢀdrug resistant form of this cell line,
known as EMT6ꢀAR1 and similar levels of IC50 were found for the derivatives studied.
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
Screening of optimal loading conditions for the derivatives with a nonꢀPEGylated liposome
demonstrated that incubation at 60°C for 30ꢀ60 min at drugꢀtoꢀlipid ratios of up to 0.21 (w:w)
were suitable for all 6 GEM derivatives (18ꢀ23). Encapsulation efficiency of the Nꢀmethylꢀ
piperazine functionalized derivatives began to be affected at higher drugꢀtoꢀlipid ratios (0.36,
~80%), but the morpholine functionalized derivatives 22 and 23 still showed near quantitative
loading. In fact, 22 and 23 still demonstrated excellent loading efficiency (~100%) even at ratios
as high as 0.36 and dropped to about 80% loading efficiency at the substantial drugꢀtoꢀlipid ratio
of 0.5. These results also translated to a similar trend with the PEGylated liposome, where the
morpholine derivatives 22 and 23 showed significantly better loading efficiency than the Nꢀ
methylꢀpiperazine derivatives 19 and 20. The reason for this was not clear, but could be due to
the differences in polarity of the morpholine moiety relative to the Nꢀmethylꢀpiperazine allowing
the derivative to pass through the hydrophobic membrane more readily. It is also clear that the
loading is not quite as effective with a PEGylated liposome as without this PEG layer, which is
expected because PEG is relatively polar, highly hydrophilic, and may interact with the
compounds as they pass across the membrane. In many ways it is these effects which we require
for the stealthꢀeffect provided by the PEGylated lipid in vivo, but the downside is that drug
ACS Paragon Plus Environment
20

Page 21 of 39
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
9
loading may be slightly affected. Investigations into a postꢀinsertion of this PEGylated lipid
could be one way to circumvent this.
At this stage, the morpholine GEM prodrugs 22 and 23 were chosen for further investigation in
vivo, based on their enhanced potency compared to other derivatives, and favourable liposome
loading capabilities. Both derivatives were used in order to investigate any differences inferred
from the major difference between the two: i.e. the stability of the ester linkage at pH 7, where
the 4 carbon linker had been demonstrated to be significantly less resistant to ester cleavage.
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
Liposomal drug loaded forms (L22 and L23) of derivatives 22 and 23 were prepared on a larger
scale with the PEGylated liposome (DSPC/Chol/DSPEꢀPEG, 56/39/5 mol%) at a drugꢀtoꢀlipid
ratio of 0.15, and similar loading efficiencies were observed as on the smaller scale, i.e. >90%.
Liposomes were all ~100 nm in size and polydispersity indexes were measured to be less than
0.06. The stability of these formulations was studied for at least 1 month at 4°C and no
significant changes were observed.
To demonstrate the effectiveness of these formulations in vivo, they were tested in a mouse s.c.
flank tumour model using the murine breast cancer cell line, EMT6ꢀAR1. Tumours were
3
inoculated a week before treatment, at which point the tumours had grown to about 125 mm ,
and then intravenous treatments were administered via the tail vein, twice weekly for all
treatments: Saline, GEM (40 mg/kg), L22 (40 mg drug/kg) and L23 (40 mg drug/kg). The
EMT6ꢀAR1 cancer cell line is known to be an aggressive tumour model and this was
demonstrated by the control animals developing large tumours within 7 days. Both free GEM
and the liposomal treatments (L22 and L23) demonstrated a significant inhibition of tumour
growth (4.8ꢀ, 8.8ꢀ and 11ꢀfold, respectively); indeed, the treatments with L22 and L23 showed
ACS Paragon Plus Environment
21

Bioconjugate Chemistry
Page 22 of 39
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
no significant difference between each other, but were also significantly better at inhibiting
tumour growth (1.8ꢀfold and 2.3ꢀfold, respectively), than the free GEM treatment (p <0.01 and
0.005, respectively). Furthermore, all liposomal formulations were shown to have low toxicity as
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
judged by changes in animal body weight, although a small but significant drop in body weight
was observed for the free GEM treated mice. This is particularly significant as previous
nanoparticle formulations of GEM have had the tendency to show increased toxicity, although
5
ꢀ6, 33
the mechanism of this has not been described.
There are still many challenges ahead for
more complicated chemotherapeutic formulations such as these, not least of which is the added
cost for synthesis scale up and production. However, as has been observed for cases such as
DOXIL, the improved efficacy, and perhaps more importantly the safety profile, can make such
formulations worthy of further study and development.
CONCLUSIONS
This work demonstrates how a prodrug strategy can be used to improve the liposomal loading of
a poorly membrane permeable drug, by chemical modification with groups to increase the
lipophilicity and a weak base moiety to enable the use of a remote loading strategy. Several
derivatives of the chemotherapeutic, GEM, were studied and a thorough investigation into their
stability, cytotoxicity and loading efficiency has been undertaken. Finally the preferred
formulations of L22 and L23 were prepared to analyze their relative efficacy in vivo. Both
derivatives demonstrated a significant inhibition of tumour growth with increased safety
compared to free GEM.
EXPERIMENTAL
ACS Paragon Plus Environment
22

Page 23 of 39
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
9
Materials
Gemcitabine was purchased from Beta Pharma (Branford, CT). 1,2ꢀDistearoylꢀsnꢀglyceroꢀ3ꢀ
phosphatidylcholine (DSPC) and 1,2ꢀDistearoylꢀsnꢀglyceroꢀ3ꢀphosphoethanolamineꢀNꢀ
methoxy(polyethyleneglycol)ꢀ2000] (DSPEꢀPEG) were purchased from Avanti Polar Lipids
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
[
(Alabaster, AL). All other reagents were purchased from Sigma Aldrich (Oakville, ON). All
synthesized compounds were checked for purity by LCꢀMS and 1H NMR where an acceptable
level was considered to be >95%.
29
General method for the synthesis of weak base appendages (1-6).
The bromoꢀalkyl ester (1 eq.) was dissolved in acetonitrile (~50 mol eq.) and the secondary
amine (1.05 eq., either methyl piperazine or morpholine) was added with stirring. To this was
added solid potassium carbonate (1 eq.) and the reaction temperature was increased to a gentle
reflux for 3 h. The reaction vessel was then allowed to cool to room temperature and the solvent
was evaporated. The residue was reꢀdissolved in dichloromethane (100 mL) and this organic
layer was washed with water (100 mL). The aqueous layer was then extracted twice with
dichloromethane (2 x 75 mL). The organic layers were combined and washed twice with
saturated sodium bicarbonate solution (200 mL) and saturated brine (100 mL). The organic layer
was then dried over anhydrous sodium sulfate and filtered to give a clear solution. The solvent
was evaporated to leave a pale yellow/orange solid (crude ester intermediate).
To prepare the HCl salt of the parent acid, the crude ester was suspended in a 1:1 mixture of
concentrated hydrochloric acid and water (20 mL). This mixture was refluxed for 1ꢀ3 h then
ACS Paragon Plus Environment
23

Bioconjugate Chemistry
Page 24 of 39
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
evaporated to dryness. The white residue was reꢀdissolved in water and evaporated to remove
residual acid and this process was repeated. Finally the solid was reꢀdissolved in a minimum
amount of ethanol and refluxed with a small amount of water added (~5% v:v). The refluxing
mixture was allowed to cool slowly and refrigerated overnight to aid crystallization. If the
crystallization was unsuccessful the second step was repeated and following ester hydrolysis the
residue was triturated with ethanol (this was found to be more successful for morpholino
derivatives). Pure compounds 1-6 were afforded in yields ranging from 70ꢀ85% over the two
steps. NMR characterisation of compounds 1-6 is reported in the supplementary information and
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
29
the patent literature.
Preparation of 4-N-octanoyl-2’-difluoro-deoxycytidine (8)
Reaction (a): 2’ꢀDifluorocytidine (1.0 g, 3.8 mmol) was dissolved in dry pyridine (40 mL) and
the mixture was cooled to 0°C. After 30 min trimethylsilylchloride (2.1 g, 2.4 mL, 19.0 mmol, 5
eq.) was added dropwise with rapid stirring, ensuring that the temperature remained in the range
of 0ꢀ5°C. Following this addition the mixture was stirred at low temperature for 2 h.
Reaction (b): Meanwhile, the octanoic acid or its sodium salt (0.6 g, 4.0 mmol, 1.05 eq.) was
dissolved/suspended in thionyl chloride (4 mL) and stirred for ~1.5 h as the temperature was
increased to 50°C. The reaction was allowed to cool slightly, diluted with dichloromethane (1:1)
and evaporated to dryness to give a yellow tinged white solid. This was redissolved in
dichloromethane (5ꢀ10 mL) and evaporated and this was repeated five times to remove residual
reagent.
ACS Paragon Plus Environment
24

Page 25 of 39
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
9
At this point the progress of reaction (a) was checked (TLC or LCꢀMS) and provided the desired
intermediate had been formed, then the residue from reaction (b) was dissolved/suspended in
dichloromethane (5 mL) and added to reaction (a) with stirring at 0°C. The reaction was then
allowed to warm to room temperature as it was stirred for a further 2ꢀ3 h. TLC and LCꢀMS were
used to follow the progress of this reaction. On completion, the reaction was evaporated to
dryness then this residue was partitioned between ethyl acetate and water (200 mL, 1:1) in a
separatory funnel. The mixture was acidified to pH 2 with phosphoric acid and the organic layer
was removed. The aqueous layer was extracted with ethyl acetate (2 x 150 mL) and the organic
layers were combined. The combined organic layers were washed with saturated sodium
bicarbonate (200 mL) and saturated sodium chloride (100 mL), then dried over anhydrous
sodium sulfate. Solid drying agent was removed by filtration and the solvent was evaporated to
yield crude product. If required, a short silica column was used to further purify the product
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
(dichloromethane/methanol 5ꢀ20%) to yield a foamy white solid (1.2 g, 80%).
1
H NMR (DMSOꢀd6) δ: 10.81 (s, 1H, NH), 8.20 (d, J = 7.6 Hz, 1H, CH), 7.27 (d, J = 7.6 Hz,
1H, CH), 6.64 (s, 1H, CH), 6.12 (t, J = 7.6 Hz, 1H, CH), 5.59 (t, J = 7.6 Hz, 1H, CH), 4.00ꢀ3.55
(m, 2H, CH), 2.39 (t, J = 7.1 Hz, 2H, CH), 1.50 (t, J = 7.3 Hz, 2H, CH), 1.27ꢀ1.12 (m, 8H, CH2),
13
0
.81 (t, J = 6.4 Hz, 3H, CH ); C NMR (DMSOꢀd6) δ: 175.1 (CONH), 163.3 (CN Ar4), 155.2
3
(CO Ar2), 145.2 (CH Ar6), 123.2 (br t CF 2’), 96.8 (CH Ar5), 84.6 (t CH 1’), 81.3 (CH 4’), 68.8
(t CH 3’), 59.1 (CH 5’), 36.9 (CH ), 31.4 (CH ), 28.6 (2CH ), 24.7 (CH ), 22.4 (CH ), 14.3
2
2
2
2
2
+
(
CH ); LC-MS (ES+): 1.48 min, 390.1 (M+H) .
3
Synthetic route to 5’-weak base modified 4-N-octanoyl-2’-difluoro-deoxycytidine
compounds
ACS Paragon Plus Environment
25

Bioconjugate Chemistry
Page 26 of 39
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
Preparation of 4-N-octanoyl-5’-TBDMS-2’-difluoro-deoxycytidine (9)
The baseꢀmodified gemcitabine 8 (500 mg, 1.3 mmol) was dissolved in dry pyridine (10 mL) in a
dry rb flask under a nitrogen atmosphere. TBDMSꢀCl (300 mg, 2.0 mmol, 1.5 eq.) was added
and the mixture was stirred at room temperature overnight. Progress of reaction was studied by
LCꢀMS and a further 0.5 eq of TBDMSꢀCl was added to consume the remaining starting
material. Solvent was evaporated under reduced pressure and the residue was reꢀdissolved in
dichloromethane (100 mL) and washed with saturated sodium bicarbonate (200 mL) and
saturated sodium chloride (200 mL). The organic layer was dried over magnesium sulfate
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
(anhydrous), filtered to remove solid drying agent, and then evaporated to dryness. This was
purified with silica gel column chromatography using hexanes/ethyl acetate (30ꢀ40%) to afford a
colourless viscous oil that formed a white foam on drying under reduced pressure (484 mg,
97%).
1
H NMR (CDCl3) δ: 9.50 (br s, 1H, NH), 8.20 (s, 1H, CH), 7.32 (d, J = 7.5 Hz, 1H, CH), 6.22
(t, J = 6.5 Hz, 1H, CH), 4.25 (dd, J = 11.2, 8.3 Hz, 1H, CH), 3.98ꢀ3.88 (m, 2H, CH), 3.79 (d, J =
1
1.2 Hz, 1H, CH), 2.41 (s, 2H, CH ), 1.55 (p(5), J = 7.4 Hz, 2H, CH), 1.25ꢀ1.07 (m, 8H, CH ),
2
2
t
0.85ꢀ0.79 (m, 9H, Bu), 0.74 (t, J = 7.0 Hz, 3H, CH ), 0.00 (s, 6H, CH ); LC-MS (ES+): 2.34
3
3
+
min, 50.4.4 (M+H) .
Preparation of 4-N-octanoyl-5’-TBDMS-3’-O-tetrahydropyranyl-2’-difluoro-deoxycytidine
30
(10)
The 5’ꢀTBDMSꢀbaseꢀmodified gemcitabine 9 (650 mg, 1.3 mmol) was dissolved in chloroform
10 mL) and stirred. To this was added, pyridinium tosylate (65 mg, 0.3 mmol, 0.2 eq.) and 3,4ꢀ
(
dihydroꢀ2Hꢀpyran (220 mg, 2.6 mmol, 2 eq.). The mixture was then stirred at 45°C for 2 h. TLC
ACS Paragon Plus Environment
26

Page 27 of 39
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
9
(
Hexane/EtOAc 1:1) was used to assess the progress of the reaction, but this showed only about
2
0ꢀ30% product. A further mol. equivalent of both pyridinium tosylate and 3,4ꢀdihydroꢀ2Hꢀ
pyran were added and the reaction was stirred overnight. TLC confirmed reaction completion the
next morning. The mixture was diluted with dichloromethane (~100 mL) and washed with an
aqueous solution of citric acid (5%, 100 mL), saturated sodium bicarbonate solution (200 mL)
and saturated brine (100 mL). The organic layer was then dried over sodium sulfate (anhydrous),
filtered and evaporated to dryness. Purification was with automated silica gel chromatography
using a gradient of ethyl acetate in hexane (20ꢀ80%). Fractions were combined to yield product
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
(550 mg, 73%) and a small amount of recovered residual starting material.
1
H NMR (CDCl3) δ: 10.19 (br s, 1H, NH), 7.98 (d, J = 7.6 Hz, 1H, CH), 7.34 (d, J = 7.6 Hz, 1H,
CH), 6.25 (t, J = 7.6 Hz, 1H, CH), 4.77ꢀ4.58 (m, 1H, CH), 4.43ꢀ4.18 (m, 1H, CH), 4.01ꢀ3.62 (m,
H, CH ), 3.44ꢀ3.37 (m, 1H, CH), 2.51ꢀ2.30 (m, 2H, CH ), 1.76ꢀ1.35 (m, 8H, CH ), 1.25ꢀ1.07
4
2
2
2
t
(m, 8H, CH ), 0.81 (s, 9H, Bu), 0.74 (t, J = 7.6 Hz, 3H, CH ), 0.00 6H (s, CH ); LC-MS (ES+):
2
3
3
+
2
.64 min, 588.2 (M+H) .
3
0
Preparation of 4-N-octanoyl-3’-O-tetrahydropyranyl-2’-difluoro-deoxycytidine (11)
Fully protected starting material 10 (550 mg, 0.9 mmol) was dissolved in THF (10 mL) and
hydrogen fluorideꢀtriethylamine complex (165 mg, 1.0 mmol, 2 eq.) was added to the stirring
reaction flask. Excess HF reagent was found to be necessary to get the reaction to progress to
completion on a reasonable timescale of ~ 36 h. Consumption of starting material was confirmed
by LCꢀMS and the solvent was evaporated to dryness. The residue was reꢀdissolved in
dichloromethane (50 mL) and the organic layer was washed with citric acid (5%, 100 mL), sat.
sodium bicarbonate (200 mL) and sat. sodium chloride (100 mL). The organic layer was dried
ACS Paragon Plus Environment
27

Bioconjugate Chemistry
Page 28 of 39
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
over sodium sulfate, filtered and the solvent evaporated to yield an offꢀwhite foamy solid.
Purification was with automated silica gel chromatography using a dichloromethane/methanol
gradient (0ꢀ10%, 10g column, 30 column volumes) to afford the title compound as a mixture of
diastereomers (380 mg, 87%).
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
1
H NMR (CDCl3) δ: 8.90 (br s, 1H, NH), 8.31ꢀ7.86 (m, 1H, CH), 7.35ꢀ7.21 (m, 1H, CH), 6.22ꢀ
6
.08 (m, 1H, CH), 4.83ꢀ4.69 (m, 1H, CH), 4.49ꢀ4.35 (m, 1H, CH), 4.03ꢀ3.89 (m, 2H, CH ), 3.86ꢀ
2
3
.75 (m, 1H, CH), 3.74ꢀ3.68 (m, 2H, CH ), 3.48ꢀ3.37 (m, 2H, CH), 2.41ꢀ2.25 (m, 2H, CH ),
2
2
1.76ꢀ1.35 (m, 8H, CH ), 1.23ꢀ0.98 (m, 8H, CH ), 0.75 (t, J = 7.6 Hz, 3H, CH ); LC-MS (ES+):
2
2
3
+
1
.8 min, 474.2 (M+H) .
General method for the preparation of 5’-weak base modified 4-N-octanoyl-3’-O-
tertrahydropyranyl-2’-difluoro-deoxycytidine (12-17)
The weak base acid derivative (1-6, 1.2 eq.) was dissolved in dry dichloromethane (to prepare a
solution of around 10 mM) and to this stirring mixture was added EDC.HCl (1.5 eq.), followed
by diisopropylethylamine (5 eq.) and the nucleoside (11, 1 eq.). The reaction was stirred at room
temperature overnight. LCꢀMS showed consumption of starting material, so the mixture was
diluted with dichloromethane (100 mL) and washed with sat. sodium bicarbonate (200 mL) and
sat. sodium chloride (100 mL). The organic layer was then dried over sodium sulfate, filtered and
the solvent was evaporated to dryness to yield a crude product (12-17) that was used in the next
step without further purification.
General method for the preparation of 5’-weak base modified 4-N-octanoyl-2’-difluoro-
deoxycytidine (18-23)
ACS Paragon Plus Environment
28

Page 29 of 39
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
9
Crude protected nucleoside (12-17) was dissolved in dichloromethane (10 mL) and
trifluoroacetic acid (5 mL) was added with stirring. This reaction was checked by LCꢀMS after 1
h, which showed a predominant product peak so the mixture was evaporated to dryness, reꢀ
dissolved in dichloromethane (20 mL) and evaporated again. This process of redissolving in
dichloromethane and evaporating was repeated 3 times. The crude product was then purified
directly by automated silica gel column chromatography eluting with dichloromethane/methanol
spiked with ammonium hydroxide (5%). A gradient was run from 5ꢀ20% methanol over 25
column volumes. The bulk of the material to elute was the desired product (yield over 2 steps:
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
65ꢀ90%) followed by a small amount of mixed fractions (~5ꢀ10%).
5’-O-(2-[4-methylpiperazin-1-yl]acetoyl)-4-N-octanoyl-2’-difluoro-deoxycytidine (18).
1
H NMR (MeOD) δ: 8.02 (d, J = 7.6 Hz, 1H, CH), 7.51 (d, J = 7.6 Hz, 1H, CH), 6.27 (t, J = 7.6
Hz, 1H, CH), 4.56 (dd, J =12.5, 2.2 Hz, 2H, CH), 4.51 (dd, J = 12.5, 5.0 Hz, 2H, CH), 4.38ꢀ4.27
m, 1H, CH), 4.22ꢀ4.18 (m, 1H, CH), 3.55 (s, 2H, CH ), 3.50ꢀ2.90 (m, 9H, CH & CH ), 2.48 (t,
(
2
2
3
J = 7.4 Hz, 2H, CH ), 1.68 (t, J = 7.0 Hz, 2H, CH ), 1.41ꢀ1.29 (m, 8H, CH ), 0.93 (t, J = 6.5 Hz,
2
2
2
13
3
H, CH ); C NMR (DMSOꢀd6) δ: 174.6 (CONH), 169.8 (COO), 163.5 (CN Ar4), 156.1 (CO
3
Ar2), 144.9 (CH Ar6), 122.1 (CF2’), 97.17 (CH Ar5), 85.75 (CH1’), 78.8 (CH4’), 70.5 (t CH3’),
2.3 (CH5’), 56.7 (CH ), 53.4 (N CH ), 48.8 (CH ), 42.1 (CH ), 36.7 (CH ), 31.4 (CH ), 28.7 (3
6
2
3
2
2
2
2
CH ), 24.6 (CH ), 24.5 (CH ), 22.2 (CH ), 12.9 (CH ); LC-MS (ES+): 1.0 min, 1057.6
2
2
2
2
3
+
(2M+H) ; HRMS (ESI) m/z: [M+H]+ Calculated for C H F N O 530.27847; Found
2
4
37
2
5
6
530.27802.
5’-O-(4-[4-methylpiperazin-1-yl]butanoyl)-4-N-octanoyl-2’-difluoro-deoxycytidine (19).
ACS Paragon Plus Environment
29