Synthesis and evaluation of photolabile insulin prodrugs

Article abstract of DOI:10.1016/j.bmcl.2005.05.112

We have developed two photolabile insulin prodrugs, insulin-2P and insulin-3P. These prodrugs were synthesized by protecting GlyA1 (N ^αA1), and one or both of the PheB1 (N^αB1) and LysB29 (N^εB29) amino groups in insulin usi

Full text of DOI:10.1016/j.bmcl.2005.05.112

Bioorganic & Medicinal Chemistry Letters 15 (2005) 3917–3920
Synthesis and evaluation of photolabile insulin prodrugs
Lian-Sheng Li,a, Jennie L. Babendure,b, Subhash C. Sinha,a,*
Jerrold M. Olefskyb,* and Richard A. Lernera,*
aThe Skaggs Institute for Chemical Biology and Department of Molecular Biology, The Scripps Research Institute,
10550 North Torrey Pines Road, La Jolla, CA 92037, USA
bDepartment of Medicine, Division of Endocrinology and Metabolism, University of California,
San Diego, La Jolla, CA 92037, USA
Received 24 March 2005; revised 20 May 2005; accepted 25 May 2005
Available online 1 July 2005
Abstract—We have developed two photolabile insulin prodrugs, insulin-2P and insulin-3P. These prodrugs were synthesized by pro-
tecting GlyA1 (NaA1), and one or both of the PheB1 (NaB1) and LysB29 (NeB29) amino groups in insulin using 50-(a-methyl-nitro-
piperonyl)oxy-carbonyl as the protecting group. These insulin prodrugs were efficiently activated by exposure to longwave UV light
to produce insulin quantitatively. Using 2-deoxyglucose uptake assays, both di- and tri-protected compounds were less active than
native insulin in the protected state, and showed comparable activity to native insulin upon photoactivation.
ꢀ 2005 Elsevier Ltd. All rights reserved.
1. Introduction
advantage of using photolabile compounds in a closed-
loop system is that they will facilitate the development
of smaller, simpler, and less invasive devices than insulin
pump-based methods could provide. As the photolabile
insulin is released by light rather than a pump, there
would be no moving parts to replace, and accidental re-
lease of the prodrug would be less damaging to the body
than accidental release of fully active insulin. In the pres-
ent study, we describe the synthesis of these insulin pro-
drugs and evaluate their ability to stimulate glucose
uptake in an in vitro system.
Hyperglycemia is the etiologic cause of most of the com-
plications of diabetes mellitus. To prevent or reduce
these complications, people with diabetes must maintain
glycemic control as close to normal as possible. This of-
ten requires multiple daily injections of insulin as well as
frequent self-glucose monitoring to maintain euglyce-
mia. Patients often have great difficulty adhering to this
rigorous schedule, and therefore it has long been a goal
in diabetic treatment to establish a closed-loop system of
regulated insulin delivery in response to elevated blood
glucose levels. In hopes of making steps toward this
goal, we have developed two insulin prodrug com-
pounds by protecting the primary amines of human
insulin with photolabile groups. These compounds are
minimally active in the protected state, but become
active after photolysis with 365 nm light. These com-
pounds have the potential for use in an implantable
closed-loop device, coupling a glucose sensor to a small
UV lamp to photoactivate and release the correct dose
of insulin in response to elevated glucose levels. The
The insulin molecule consists of two polypeptide chains,
A and B, joined together by two disulfide bonds. In
addition to two N-terminal primary amines, A1 and
B1, the insulin molecule also contains one e amine on
the lysine residue of chain B. To date, most insulin pro-
drugs or derivatives have been synthesized by protecting
the primary amines at A1, B1, or B29.1
Keywords: Diabetes; Insulin; Photolysis; Prodrug; Ultraviolet.
*
Corresponding authors. Tel.: +1 858 7848512; fax: +1 858 7848735
(R.A.L); Tel.: +1 858 5346651; fax: +1 858 5346653 (S.C.S.); Tel.: +1
858 7848265; fax: +1 858 7849899 (J.M.O.); e-mail addresses:
subhash@scripps.edu; jolesfsky@ucsd.edu
These authors contributed equally to this work.
0960-894X/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.05.112

3918
L.-S. Li et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3917–3920
Studies have revealed that the PheB1 (NaB1) and LysB29
(NeB29) amino groups of insulin do not directly partici-
pate in receptor binding.2 Thus, protection of these
two amino groups has minimal effect on insulin activity.
The GlyA1 (NaA1) amino group of insulin, however,
plays a critical role in receptor binding and its protection
leads to diminished binding affinity and biological activ-
ity. Utilizing this property of insulin, various derivatives
have been synthesized. We previously investigated the
functional properties of native insulin in comparison
to an aldol-derivatized organoinsulin prohormone in
the absence and presence of the catalytic aldolase anti-
body 38C2.3 In the aldol-derivatized organoinsulin, all
three amine functions of insulin were protected with al-
dol linkers. This prodrug compound exhibited markedly
diminished receptor binding and biological activity both
in vitro and in vivo. When the pro-insulin compound
was incubated with 38C2, native insulin was regenerat-
ed, displaying restored receptor-binding affinity as well
as normal in vitro and in vivo biological activity. Taking
a cue from this study, we sought to design photolabile
insulin prodrugs that could be activated by photolysis
with near ultra violet (UV) light. Such photolabile insu-
lin prodrugs could be stored in a receiver and photoac-
tivated for administration. As a proof of concept, we
have synthesized two photolabile insulin derivatives
and studied their properties, including activation with
UV light and the effect of the photolabile insulins on
glucose uptake, in vitro.
Scheme 1. Protection of the primary amines in insulin using the
photolabile groups to produce insulin prodrugs, insulins-2P and 3P.
lyzed by analytical HPLC assay using a reverse phase
C18 column and a UV detector at 280 nm. Four major
peaks at RT 24.3, 27.7, 27.9, and 30.2 min were observed
in the chromatogram (Fig. 1A). It was found that both
insulin and 4-nitro-phenol generated from the reaction
had the same retention time corresponding to 24.3 min
(data not shown). Therefore the rest of the major peaks
were expected to be modified insulins. The low molecu-
lar weight compounds (4-nitro-phenol, etc.) were re-
moved completely after dialysis (Fig. 1B). Thus, the
reaction mixture was diluted with H2O (3.0 ml) and then
transferred into a 3500 MWcutoff, 12 ml dialysis mem-
brane slide (Pierce) and dialyzed with either 0.1 N
ammonium carbonate buffer or H2O. The residue inside
the membrane was chromatographed using preparative
HPLC and two fractions were collected. The first frac-
tion contained two inseparable compounds that corre-
sponded to RT 27.7 and 27.9 min, while the second
fraction was a single compound corresponding to RT
2. Synthesis, isolation, and activation of the photolabile
insulins
For the synthesis of the photolabile insulins, we used the
50-(a-methyl-nitro-piperonyl)oxy-carbonyl (MeNPOC)
group to mask the free primary amines. It has been
reported that the MeNPOC protecting group used in
many applications, including the light-directed synthesis
of DNA arrays on glass substrates, could be rapidly
deprotected using UV light of 365 nm. As shown in
Eq. 1, a compound with the general structure 1 is depro-
tected to the corresponding amine (RNH2) and nitroso
derivative 2 in less than a minute.4 It is noteworthy that
the 365 nm emission is almost exclusively responsible for
the photochemistry, as the absorbance of the MeNPOC
chromophore (kmax = 345 nm, e = 5 · 103 Mꢀ1 cmꢀ1
becomes negligible at wavelengths above 400 nm.
)
ð1Þ
Synthesis of the N-protected insulins was achieved by
treatment of parent insulin with carbonate 3 in DMSO
followed by dialysis and Prep-HPLC purification
(Scheme 1). Briefly, compound 35 (3 equiv with respect
to insulin) was added to a suspension of the recombi-
nant human insulin (purchased from Sigma–Aldrich,
70 mg, 12 lmol) in reagent grade DMSO (3.0 ml). The
resulting mixture was stirred in the dark at room tem-
perature overnight. Then the crude mixture was ana-
Figure 1. Sections of the HPLC traces showing: (A) the crude reaction
mixture of insulin with compound 3, (B) the reaction mixture shown in
(A) after dialysis, and (C) the diprotected (insulin-2P) and (D)
triprotected (insulin-3P) insulin derivatives separated by preparative
HPLC. Conditions for the analytical HPLC: reverse phase C18
analytical HPLC column (VydacTM, Cat# 201SP54). Water–acetoni-
trile (100:0–20:80) at a rate of 0.8 ml/min using gradient system,
starting with 100% water at 0 min and 80% acetonitrile in water at
40 min.

L.-S. Li et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3917–3920
3919
30.2 min in the analytical HPLC (Figs. 1C and D,
respectively). Both fractions were separately lyophilized,
to give 13 and 10 mg of the protected insulins.
The protected insulins were analyzed by liquid chroma-
tography mass spectrometry (LCMS) electrospray ioni-
zation (ESI) experiments, which gave a molecular ion
peak (M+) at 6281 for the first fraction and at 6518
for the second fraction. As a control, recombinant hu-
man insulin showed a consistent molecular weight of
5807 under the same conditions. The data confirmed
that the first fraction contained both diprotected insu-
lins (insulins-2P), whereas the second fraction contained
the triprotected insulin derivative (insulin-3P). Based on
our previous study, it was evident that only the three pri-
mary amine functions at GlyA1 (NaA1), PheB1 (NaB1),
and LysB29 (NeB29) could be protected in insulins-2P
as well as 3P under the reaction conditions, however,
it was not clear which two amines were protected in
insulins-2P. Fortunately, the biological data suggest that
the GlyA1 (NaA1) amino group, which directly partici-
pates in receptor binding, was protected in both dipro-
tected compounds. Alternative protection of PheB1
(NaB1) and LysB29 (NeB29) amino groups made two
diprotected insulins, which were inseparable by HPLC.
The attempts to convert all the insulin to insulin-3P by
increasing the amount of compound 3 and prolonging
the reaction time were unsuccessful.
Figure 2. Part of the HPLC traces showing the regeneration of insulin
and production of compound 2 from insulin-2P (A, C, and E) at the
end of 1, 2, and 5 min and from insulin-3P (B, D, and F) at the end of
1, 2, and 10 min, respectively. (HPLC conditions: see Fig. 1).
Next, we examined the activation of the insulin pro-
drugs using a high intensity longwave UV lamp (Ul-
tra-Violet Products, model # B series 100AP). Both
insulins-2P and 3P underwent rapid deprotection to re-
lease free insulin and the nitroso derivative 2 (Scheme
2) as evidenced by the HPLC (Fig. 2) and (LCMS) anal-
yses of the crude products, using authentic insulin as
comparison. As shown in Figures 2A and B, respective-
ly, insulins-2P and 3P were converted to the free insulin
(RT 24 min) as well as the corresponding intermediates,
which are most likely the mono-protected insulins
(RT ꢁ 26 min) from insulins-2P, and the mixtures of
mono- and diprotected insulins (RT ꢁ 26–28 min) from
insulin-3P. Thus, at the end of the first minute of UV
treatment, more than 50% and 30% insulin were gener-
ated from insulin-2P and 3P, respectively. The insulin
concentration from insulins-2P and 3P increased to
75% and 50% at the end of second minute (Figs. 2C
and D), and finally rose to almost complete deprotection
at 5 and 10 min, respectively (Figs. 2E and F).
3. Prodrugs insulins-2P and 3P stimulate glucose uptake
when photoactivated
To demonstrate the activity of insulin regenerated from
insulins-2P and 3P, we performed 2-deoxyglucose up-
take assays before and after photoactivation. Glucose
uptake was determined by assessing the amount of
[3H]2-deoxyglucose taken up by cells stimulated with na-
tive insulin, protected, or photoactivated compounds.
For this study, 3T3-L1 adipocytes were cultured and dif-
ferentiated as previously described.6 Cells were seeded
into 12 well plates and the assay was conducted on days
12–14 of differentiation. Insulins-2P and 3P were pro-
tected from light and dissolved in 10 mM HCl before
use. Dissolved compounds were placed in Acryl-cuvettes
(Sarstedt), and illuminated for 15 min with a BLAK-
RAY Longwave Ultraviolet Lamp (Model B-100A,
San Gabriel, CA) to induce photolysis. Native human
insulin, insulins-2P and 3P, were used at two doses,
0.5 and 17 nM. Glucose uptake was measured as de-
scribed elsewhere,6 with some adjustments. 3T3-L1 adi-
pocytes were serum starved for 2 h and stimulated with
insulin, the protected insulins-2P or -3P, or the photoac-
tivated compounds for 20 min. Glucose uptake was
Scheme 2. Activation of insulins-2P and 3P by photolysis under UV
light (hm = 365 nm) to regenerate insulin.

3920
L.-S. Li et al. / Bioorg. Med. Chem. Lett. 15 (2005) 3917–3920
measured by incorporation of [3H]2-deoxyglucose after
a 15 min incubation.
explanation for the increased activity seen with both
protected and deprotected insulin-3P is that a small
portion of the compound may have been deprotected
prior to measurement and use.
Figure 3 compares glucose uptake using insulins-2P and
3P. Both compounds were moderately active in the
protected state, perhaps indicating some degree of
degradation. Upon photoactivation, the compounds
demonstrated variable degrees of activity, with the
diprotected insulin-2P activating 70% of the maximal
response (Fig. 3A). The triprotected insulin-3P appeared
to activate 116% of the maximal response; however, this
difference from native insulin was not statistically signif-
icant as shown by one-way ANOVA followed by
DunnettÕs multiple comparison post-test (Fig. 3B). One
In conclusion, photolabile insulin prodrugs, insulins-2P
and insulin-3P, were developed by protecting GlyA1
(NaA1) and one or both of the PheB1 (NaB1) and LysB29
(NeB29) amino groups using MeNPOC as the protecting
group. These insulin prodrugs were efficiently activated
by exposure to longwave UV light to produce insulin
quantitatively. Using 2-deoxyglucose uptake assays,
both di- and tri-protected compounds were significantly
less active than native insulin, and activity was restored
upon photoactivation. These compounds provide a
possible further step toward the development of a
closed-loop system of concomitant glucose detection
and insulin delivery.
Acknowledgments
Financial supports from the Skaggs Institute for Chem-
ical Biology (S.C.S. and R.A.L.) are gratefully acknowl-
edged. This work was also supported in part by NIH
Grant DK-33651 (J.L.B. and J.M.O.). This is manu-
script number 17354-MB from The Scripps Research
Institute.
References and notes
1. (a) Gershonov, E.; Goldwaser, I.; Fridkin, M.; Shechter, Y.
J. Med. Chem. 2000, 43, 2530; (b) Hinds, K.; Koh, J. J.;
Joss, L.; Liu, F.; Baudys, M.; Kim, S. W. Bioconjugate
Chem. 2000, 11, 195.
2. (a) Gliemann, J.; Gammeltoft, S. Diabetologia 1974, 10,
105; (b) Hashimoto, M.; Takada, K.; Kiso, Y.; Muranishi,
S. Pharm. Res. 1989, 6, 171; (c) Murray-Rust, J.; McLeod,
A. N.; Blundell, T. L.; Wood, S. P. BioEssays 1992, 14, 325.
3. Worrall, D. S.; McDunn, J. E.; List, B.; Reichart, D.;
Hevener, A.; Gustafson, T.; Barbas, C. F., III; Lerner, R.
A.; Olefsky, J. M. Proc. Natl. Acad. Sci. U.S.A. 2001, 98,
13514.
4. McGall, G. H.; Barone, A. D.; Diggelmann, M.; Fodor, S.
P. A.; Gentalen, E.; Ngo, N. J. Am. Chem. Soc. 1997, 119,
5081.
Figure 3. Glucose uptake stimulated by native insulin, protected and
photoactivated insulin compounds. 3T3-L1 adipocytes were stimulated
with the indicated amounts for 20 min and glucose uptake was
quantitated by incorporation of [3H]2-deoxyglucose. (A) Activity of
insulin 2P in comparison to native insulin. (B) Activity of insulin 3P in
comparison to native insulin. Figures are an average of three
experiments.
5. Compound 3 was synthesized by a reaction of 50-(a-methyl-
nitro-piperonyl)ol with 4-nitrophenyl chloro formate in the
presence of pyridine at 0 ꢁC.
6. Imamura, T.; Vollenweider, P.; Egawa, K.; Clodi, M.;
Ishibashi, K.; Nakashima, N.; Ugi, S.; Adams, J. W.;
Brown, J. H.; Olefsky, J. M. Mol. Cell. Biol. 1999, 19, 6765.