1,2-Dihydroisoquinoline-N-Acetic Acid Derivatives as New Carriers for Specific Brain Delivery I: Synthesis and Estimation of Oxidation Kinetics Using Multivariate Calibration Method

Article abstract of DOI:10.1002/ardp.200300808

In order to overcome the slow oxidation of 1,2-dihydro-N-alkylisoquinoline, 1,2-dihydroisoquinoline-N-acetic acid derivatives (3 b-d) were designed and synthesized as new chemical delivery systems (CDS) for the brain. Molecular orbital calculations for th

Full text of DOI:10.1002/ardp.200300808

Arch. Pharm. Pharm. Med. Chem. 2003, 336, 573–584
1,2-Dihydroisoquinolines as Carriers for Specific Brain Delivery 573
Sahar Mahmouda,
1,2-Dihydroisoquinoline-N-Acetic Acid Derivatives
as New Carriers for Specific Brain Delivery I:
Synthesis and Estimation of Oxidation Kinetics
Using Multivariate Calibration Method.
Tarek Aboul-Fadla,
Mahmoud Shehaa,
Hassan Faraga,
Abdel-Maaboud I.Mouhamedb
a
Dept. Pharm. Med. Chem.
Dept. Anal. Pharm. Chem.,
In order to overcome the slow oxidation of 1,2-dihydro-N-alkylisoquinoline, 1,2-dihy-
droisoquinoline-N-acetic acid derivatives (3b–d) were designed and synthesized
as new chemical delivery systems (CDS) for the brain.Molecular orbital calculations
for the suggested derivatives revealed that these carriers are stable against oxida-
tion.However, hydrolysis to their corresponding anions will accelerate the rate of oxi-
dation, and accordingly increase the efficiency of brain specific delivery.A multivari-
ate calibration method for in vitro determination of the oxidation rate for the suggest-
ed brain specific chemical delivery system is described. The method is based on
measurement of individual rates of oxidation of prepared 1,2-dihydroisoquinolines,
using silver ions to provide the corresponding quaternary forms.Components of bi-
nary mixtures formed through the oxidation step (composed of the dihydro-com-
pound and its quaternary form), showed a considerable degree of spectral overlap-
ping – more than 90% in all cases.Resolution of these binary mixtures under inves-
tigation has been accomplished mainly using classical least squares analysis. Ki-
netic oxidation data of the tested compounds revealed that these new CDSs have a
reasonable oxidation rate for efficient brain delivery.
b
Faculty of Pharmacy,
Assiut University,
Assiut 71526, Egypt
Keywords: 1,2-Dihydroisoquinoline; Brain chemical delivery system; Spectro-
photometry; multivariate analysis; MO calculations; Oxidation kinetics
Received: April 12, 2003; Accepted: August 6, 2003 [FP808]
DOI 10.1002/ardp.200300808
compounds protects positions at which hydration takes
place [9–12].
Introduction
Drug targeting to the defective site is at present one of
the most important goals of pharmaceutical research.
Delivery to the brain is a particularly important part of
this goal. The impermeability of the blood brain barrier
(BBB) to many drugs due to its hydrophilic characteris-
tics prevents treatment of many central nervous system
(CNS) diseases.The 1,4-dihydropyridine chemical deliv-
ery system (PCDS) was reported for specific delivery of
drugs to the brain and was found to be very successful
for this purpose [1–3]. The main disadvantages hinder-
ingPCDSdevelopment forclinicaluse aretheirsuscepti-
bility to hydration at the C5–C6 double bond and prompt
air oxidation.Consequently, they are unsuitable for phar-
maceutical formulation and have a short shelf-life time
[4–8].
The present investigation is concerned with design, syn-
thesis and study of a novel brain specific chemical deliv-
ery system, displaying both stability and efficiency. The
suggestedcarriersystem(Scheme1)isbasedon1,2-di-
hydro-N-carboxymethyl-isoquinoline derivatives, 3a–d,
as they are able to overcoming the very slow oxidation
rate of 1,2-dihydro-N-alkylisoquinoline [13].
Scheme 1. a: R = H; b: R = CH3; c: R = C2H5;
d: R = iso-C3H7
To overcome these problems, 1,2-dihydroisoquinoline
and/or 1,4-dihydroquinoline were tested as brain specif-
ic carriers. The presence of the benzene ring in these
Delivery to the brain by this system is based on penetra-
tion of reduced forms of the compound (1,2-dihydroiso-
quinolines), with their considerable lipophilic character-
istics, through the BBB. Oxidation of the reduced forms
Correspondence: Tarek Aboul-Fadl, Department of
Pharmceutical and Medicinal Chemistry, Faculty of Pharmacy,
Assiut University, Assiut 71526, Egypt.Phone:+20 88 411-321,
Fax: +20 88 332-776, e-mail: Fadl@acc.aun.edu.eg
© 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

574 Mahmoud et al.
Arch. Pharm. Pharm. Med. Chem. 2003, 336, 573–584
Scheme 2. Proposed mechanism for brain delivery of the designed carrier system.
to corresponding quaternary forms results in their reten-
tion in the brain. Therefore, oxidation rate of the carrier
system is the limiting step for brain delivery. Scheme 2,
illustrates the proposed mechanism of brain delivery for
the designed carrier system.
easily accessible statistical software. It is preferable to
maintain a simple variable selection system when using
this method, i.e., restricted variables for compounds or
characteristics under study, and low or insignificant
baseline effects and noise.
It is important to determine the oxidation rate of the de-
signed carrier system for sensible stability during formu-
lation and storage. High-performance liquid chromatog-
raphy (HPLC) techniques are not suitable for such deter-
minations since the columns used, become blocked by
silver ions.Derivative spectrophotometric techniques al-
so cannot precisely analyze these mixtures due to the
high degree of spectral overlapping (more than 90%).
Full-spectrum methods usually provide significant im-
provement in precision over methods restricted to a
small number of wavelengths. A convenient method for
resolving the mixtures, which can in principle be applied
to the present case, is the least squares analysis. Multi-
variate calibration methods applied to spectral data are
increasingly used for pharmaceutical analysis [14–19].
Classical least squares (CLS) analysis is one of the sim-
plest multivariate methods that can be performed with
The present work discusses the possibility of simultane-
ous analysis of combined mixtures of prepared dihydro-
derivatives with their corresponding quaternary (Q)
forms, produced by oxidation with silver ions.The meth-
od is based on spectrophotometric measurements in the
270–370 nm range, together with multivariate calibration
analysis. Results obtained showed that CLS regression
allows the accomplishment of this goal, and the tech-
nique gives satisfactory results compared to other meth-
ods.
Theoretical molecular orbital (MO) calculations have al-
so been carried out on the suggested derivatives in order
to predict if the carrier will be stable upon oxidation, and if
hydrolysis to the anionic form will accelerate the rate of
oxidation, henceincreasingthe efficiency ofbrain specif-
ic delivery.
© 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Arch. Pharm. Pharm. Med. Chem. 2003, 336, 573–584
1,2-Dihydroisoquinolines as Carriers for Specific Brain Delivery 575
however, is shifted to 3116 cm–1 for 2a due to the elec-
tron withdrawing effect of the carboxylate group.Carbon-
yl groups appeared at 1732 cm–1 for compounds 2b–d,
whereas it appeared at 1720 cm–1 for compound 2a be-
side the characteristic band of hydroxyl group at
3460 cm–1.
Results and discussion
Chemistry
1,2-Dihydroisoquinolines were designed, and synthe-
sized by a two step reaction. The first step involves
quaternization of isoquinoline with bromoacetic acid or
the corresponding ester in boiling acetone as illustrated
by Scheme 3. The structures of quaternized products
2a–d were verified with elemental analysis and spectro-
scopic methods.
1H-NMR data of quaternized derivatives, 2a–d, revealed
asignificantdownfieldshiftof1-Hfortheisoquinolinemoi-
ety as it resonates around 10.2 ~ 10.6 ppm, whereas in
isoquinoline it resonates at around 8.6 ppm. N-alkyl pro-
tons resonate at around 5.9 ~ 6.2 ppm for 2b–d, however,
it appeared at 5.3 ~ 5.5 ppm for 2a.The upfield shift of N+-
CH2 protons of the carboxylic acid derivative, 2a, com-
pared to corresponding esters, 2b–d, may be attributed to
the dissociation of the former as carboxylate ion in DMSO.
With the exception of the ester groups, resonance of the
other sets of protons in synthesized derivatives are almost
superpossible Please rephrase & replace last word.
UV Spectral data was similar to isoquinoline since the
quaternized derivatives, 2a–d, showed characteristic
maxima at approximately 325, 260, 220 nm. However, a
significant red shift was observed for the quaternized de-
rivatives, ranging from 15 to 20 nm.
IR spectral data of quaternized isoquinolines revealed
characteristic bands for N+-R at 3020 cm–1. This band,
Scheme 3. 2a: R = H; 2b: R = CH3; 2c: R = C2H5; 2d: R = iso-C3H7
© 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

576 Mahmoud et al.
Arch. Pharm. Pharm. Med. Chem. 2003, 336, 573–584
The second step involves reduction of quaternized prod-
ucts to corresponding 1,2-dihydroisoquinolines, 3a–d.A
variety of methods have been reported for preparation of
1,2-dihydroisoquinoline from their corresponding qua-
ternaries, e.g., reduction with sodium borohydride [20,
21], lithium aluminium hydride [20, 22], sodium dithionite
in alkaline medium [23], 1-benzyl 1,2-dihydroisonico-
tinamide [24], organometallic reagents [25–27], and po-
tassium cyanide in alkaline medium [28, 29]. In the
presentinvestigation sodiumdithionite and1-benzyl1,2-
dihydroisonicotinamide were used for preparation of 1,2-
dihydroisoquinolines, 3a–d.
i. Adiabatic (sequential mechanism)
Electron-proton-electron is the mechanism of oxidation
with ferricyanide, and will not be involved in the current
study.
ii. Hydride transfers mechanism
The process occurs by loss of one hydride ion from C1 of
isoquinolinetogive the quaternaryform(IV), as shownin
Scheme 5.
The dithionite method is achieved by reaction of the iso-
quinoline quaternary salt and dithionite in alkaline medi-
um at 0°C in a biphasic reaction under nitrogen. The
method was found to give good yields for corresponding
dihydroisoquinolines. The main disadvantage of this
method is susceptibility of esters to hydrolysis in the
aqueous alkali reaction medium, and difficulties encoun-
tered in separating the product in pure form. However,
1-benzyl-1,2-dihydroisonicotinamide method is based
on mixing reagent (4) with the isoquinoline quaternary
salt in methanol.The disproportion reaction results in re-
duction of the isoquinolinum salt to the corresponding di-
hydroisoquinoline while the reagent is converted to its
quaternary form (5) as a byproduct.The latter is insolu-
ble in methanol and can be easily removed by filtration
[30]. Scheme 4 summarizes the reduction of isoquino-
line quarterneries, 2a–d, to their corresponding dihydro-
derivatives, 3a–d.
Scheme 5. a: R = CH2CO2H; b: R = CH2CO2CH3;
c: R = CH2CO2C2H5; d: R = CH2CO2CH(CH3)2
Structures of dihydroderivatives, 3a–d, were confirmed
by spectroscopic analysis methods. Insignificant blue
shift was observed in the UV spectra of dihydroderiva-
tives compared to their corresponding quaternary forms.
IR spectra of the dihydro(DH) group, however, are char-
acterized by disappearance of the N+-R band at
3030 cm–1. A significant upfield shift of 1-H of isoquino-
line was observed in the 1H-NMR spectra of dihydrode-
rivatives, 3a–d, which ranged from 5.7–6.2 ppm and in-
tegrated as 2H.Another upfield shift was observed in the
chemical shifts of N-CH2 ranging from 1.2–2.4 ppm
which is further evidence for reduction.
This pathway can take place when a good hydride ac-
ceptor as oxidant is used [33], such as 1-methyl acridi-
num ions. The heat of formation in this pathway can be
calculated by equation (1).
∆∆Hf = ∆Hf of quaternary form
– ∆Hf of the DH form
(1)
These two pathways resemble enzymatic conversions in
vivo.
In the present investigation it was essential to perform
the theoretical calculation to predict if the putative CDS
will be“shelf-stable”, and if after ester hydrolysis acceler-
ationofoxidationrateswilloccur;i.e., ifbothobjectivesof
the study, shelf stability and efficiency of brain-specific
delivery, can be fulfilled.
Theoretical MO calculations and C log P
Molecular orbital calculations were done using AM1
semi-empirical calculation [31], in the MOPAC package.
MOPAC is a general-purpose semi-empirical molecular
orbital package for the study of chemical structures and
reactions. The oxidation of dihydroisoquinolines to their
corresponding quaternary forms can occur by two
mechanisms [32], which explain the results of experi-
mental and mathematical calculations.
The Austin Model 1 (AM1) calculations were used to de-
termine the heats of formation of the suggested deriva-
tives of isoquinoline, (Scheme 1).The ∆∆Hf for the reac-
tion is calculated by equation (2):
© 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Arch. Pharm. Pharm. Med. Chem. 2003, 336, 573–584
1,2-Dihydroisoquinolines as Carriers for Specific Brain Delivery 577
∆∆Hf = Σ∆Hf(p) – Σ∆Hf(R)
(2)
the esters will be stable against oxidation in air.But once
administrated, distributed, and hydrolyzed to the corre-
sponding anions, fast oxidation will occur, locking the
compound in the brain, resulting in highly efficient brain-
specific delivery.The acetic acid ester derivatives do not
differ largely in their heats of formation, but vary in their
lipophilicity, expressed by C log P, (Table 2).
where Σ∆Hf(p) is the sum of the calculated heat of forma-
tion of products, and Σ∆Hf(R) is the sum of heat of forma-
tion of reactants (for each step or for all reaction) [34].
Since the main aim of the new carrier-ester is its in vivo
hydrolysis to the corresponding anion (3Јa) after admin-
istration and distribution, these anions were included in
the theoretical calculation, to compare the results with
the original carriers (acid and esters).Results of calcula-
tions are shown in Table 1.
Multivariate calibration of tested compounds
Absorption spectra for 3c and its corresponding quater-
nary form 2c, as representative example of the studied
compounds are shown in Figure 1. As can be seen, a
considerable degree of spectral overlapping occurs in
the region between 270 to 370 nm.
Table 1. Summarized calculated results for ∆Hf of sug-
gested compounds.
Compound
⌬HKcal.
⌬⌬Hf(e)
⌬⌬Hf(H:)
The degree of spectral overlapping can be calculated by
(Di)0.5, where Di is the magnitude of dependency that can
becalculatedfora twocomponentmixturefromequation
(3) [35]:
3Ј a (I)
3Ј a (IV)
3Ј a (II)
3a (I)
3a (IV)
3a (II)
–68.72
6.87
–3.31
–36.29
116.38
134.62
65.41
75.52
(Σk1 kt2)2
Σk1 kt1 · Σk2 kt2
Di =
(3)
170.91
177.04
152.67
160.35
where k1 and k2 are the l × n matrices of regression coef-
ficients for compounds 3c and 2c, respectively.With re-
spect to the compounds studied, the spectra shown in
Figure 2 for DH-derivative and its Q-form, lead to Di =
0.831 implying 91.1% spectral overlapping.
3b (I)
3b (IV)
3b (II)
3c (I)
3c (IV)
3d (I)
–38.56
121.79
138.48
–42.98
114.65
–45.71
109.7
NA
NA
157.63
155.41
A linear relationship between absorbance and compo-
nent concentrations at each wavelength is assumed,
such a relationship is described by equation (4):
3d (IV)
3Ј a is the anionic form of 3a.
NA – not available.
A = C K + E
(4)
where A is the m × n matrix of calibration spectra, C is the
m × l matrix of component concentrations, K is the l × n
matrix of regression coefficients and E is the m × nmatrix
of spectral errors or residuals not fit by the model. Using
the calibration sample set, estimation of absorptivities is
possible by solving matrix K according to the general
least-squares solution indicated by equation (5):
From results given inTable 1 it was found that:there is no
large difference between the ∆∆Hf(e) for adiabatic or
∆∆Hf(H:) in the hydride transfer mechanisms, which
means that either one can be used for calculation of
∆∆Hf.On the other hand, the smallest ∆∆Hf value is that
of the anion (3Јa), which indicates it is the fastest com-
pound in oxidation. Accordingly, it is expected that
K = (Ct · C)–1 · Ct · A
(5)
Table 2. MO parameters and C log P of compounds 3a–d.
Core
repulsion
Gradient
norm
Ionization
potential
Compound
Hf
E energy
Dipole
Clog P
3a
3Ј a
3b
3c
3d
–36.291199 –12886.41
–68.723713 –12594.12
–38.562594 –14761.14
10479.46
10199.35
12198.56
13396.41
14740.
7.730028
2.27849
11.53787
9.847591
3.87798
4.40899
13.47553
1.50328
1.8953
8.3214
2.741
–0.538
3.21
3.739
4.048
4.495027
8.149747
8.080092
8.027066
–42.98608
–16114.78
–45.717819 –17613.96
2.77087
© 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

578 Mahmoud et al.
Arch. Pharm. Pharm. Med. Chem. 2003, 336, 573–584
Figure 1. Degree of overlap in the studied binary mixtures, as indicated by adsorption spectra of, 3c, and its corre-
sponding quaternary structure, 2c, at 40 µg/mL of each.
Figure 2. Predicted concentrations of 3c errors, as calculated by MLR method.
Analysis is then based on the spectrum (Aun) of unknown
samples by:
proportions, were subjected to CLS analysis. Recover-
ies in all cases were satisfactory and relative deviations
between estimated and true concentrations, expressed
by the relative root mean squared error (RRMSE), were
found to be between 0.5 and 2.5%. Results may reflect
the precision of the method for prediction of concentra-
tions of both mixture components following the oxidation
of DH-compounds, 3b–d, to corresponding Q-deriva-
tives, 2b–d, and consequently the oxidation rate of the
former or the formation rate of the latter.Figure 4 andTa-
ble 3 illustrate the results obtained for compounds 3b–d,
at different time intervals over60.0minaftersilverion ad-
dition.
Cun = Aun Kt (K · Kt)–1
(6)
where Cun is the vector of concentrations sought.
Figures 2 and 3 illustrate the actual and predicted
amounts of the studied compound, 3c, and its quater-
nary form, 2c, given by the least squares regression
analysis of spectral data obtained experimentally in the
range 270–370 nm.
Several synthetic mixtures of each compound and its
corresponding Q-form, prepared in continuously varied
© 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Arch. Pharm. Pharm. Med. Chem. 2003, 336, 573–584
1,2-Dihydroisoquinolines as Carriers for Specific Brain Delivery 579
Figure 3. Predicted concentrations of 2c errors, as calculated by MLR method.
Figure 4. First order plots of oxidation of 3b–d and formation of 2b–d at room temperature ᭡ 3b, ᭡ 2b, ᭺ 3c, ᭺ 2c,
᭹ 3d, ᭹ 2d.
Table 3. Results obtained by applying CLS analysis to DH-derivatives (3b–d), after oxidation with silver nitrate at vari-
ous times between 0 to 60 min.
Time
(min)
Compd.
(DH)
Found*
(µg/mL)
Found
(%)
Product
(Q)
Found*
(µg/mL)
Found
(%)
Starting
Conc.
3b
3c
3d
60.00
153.00
87.00
100.00
100.00
100.00
2b
2c
2d
0.00
0.00
0.00
0.00
0.00
0.00
0
3b
3c
3d
57.22
143.39
81.90
95.36
93.72
94.14
2b
2c
2d
1.16
5.29
2.75
1.76
2.77
3.82
© 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

580 Mahmoud et al.
Arch. Pharm. Pharm. Med. Chem. 2003, 336, 573–584
Table 3. (continued).
Time
(min)
Compd.
(DH)
Found*
(µg/mL)
Found
(%)
Product
(Q)
Found*
(µg/mL)
Found
(%)
5
3b
3c
3d
55.51
138.66
79.93
92.51
90.63
91.87
2b
2c
2d
3.91
10.71
4.31
5.92
5.60
5.90
10
15
30
40
60
3b
3c
3d
53.42
133.88
78.11
89.03
87.5
89.78
2b
2c
2d
5.36
14.23
5.94
8.12
7.44
8.14
3b
3c
3d
51.86
129.16
76.51
86.43
84.42
87.94
2b
2c
2d
7.84
19.58
7.53
11.88
10.24
10.32
3b
3c
3d
44.51
nd
74.18
nd
2b
2c
2d
15.54
nd
23.55
nd
nd
nd
nd
nd
3b
3c
3d
41.20
119.37
73.18
68.67
78.02
84.11
2b
2c
2d
19.94
33.53
10.31
30.21
17.53
14.13
3b
3c
3d
37.23
105.16
72.11
62.05
68.73
82.88
2b
2c
2d
24.70
50.30
11.20
37.42
26.30
15.34
* Calculated according to calibration data obtained for each compound separately, in the pure form.
nd not determined
Table 4. Pseudo first order rates of tested compounds
(3b–d), estimated by multivariate analysis method.
Oxidation kinetic study with silver nitrate
Oxidation with silver nitrate was used as a qualitative test
to determine whether the DH-compounds are suscepti-
ble to oxidation to from their corresponding quaternar-
ies.The test can also be used for quantitative analyses.
Oxidation rates of compounds, 3b–d, were determined
at room temperature using the multivariate method dis-
cussed above. The pseudo-first order rate constants
(kobs.) for oxidation of the DH-forms, 3b–d, were ob-
tained from slopes of semilogarithm (kobs. = slope X
2.303) plots of dihydroderivative concentrations versus
time, using linear regression equation. The rate of the
free acid oxidation, 3a, was too fast to be determined
which is contrary to that reported for N-methyl isoquino-
line derivatives [13]. The prompted oxidation rate of 3a
that hindered its practical determination, may be due to
the existence of its ionized form (3Јa). Oxidation rates of
dihydroesters, 3b–d, are given in Table 4. It should be
noted that the oxidation rate is reversibly proportional to
the bulkiness of the ester, which indicates prior occur-
rence of hydrolysis.
Compound Kobs (min–1)
t (min)
½
r
3b
3c
3d
0.0143
0.0086
0.0042
48.5
80.5
164
0.999
0.996
0.959
Significant correlation was observed between the rates
of oxidation of tested compounds, 3b–d, (Table 4), and
the parameters obtained from MO calculation and their
C log P (Table 2). Correlation matrix is shown in Table 5.
A strong correlation between these parameters is clear.
However, best correlations were observed between the
rate of oxidation with ionization potential (r = 0.99999)
and heat of formation (r = 0.9981). Accordingly, theoreti-
calratesofoxidationofthesederivativescanbeestimated
by means of their heat of formation using equation (7):
© 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Arch. Pharm. Pharm. Med. Chem. 2003, 336, 573–584
1,2-Dihydroisoquinolines as Carriers for Specific Brain Delivery 581
Table 5. Correlation matrix between observed rates of oxidation of tested compounds (3b–d), and their calc. MO
parameter and C log P.
Rate of
Oxidation
E energy
Core
repulsion
Gradient
norm
Dipole
moment
Ionization Clog P
potential
Hf
1
Rate of Oxid.
Heat of
1
0.99811
Formation
E energy
0.994636 0.9863997
1
Core repulsion –0.994252 –0.9857936 –0.9999933
1
Gradient norm 0.926352 0.9014542 0.9603432 –0.96135
Dipole moment –0.957969 –0.9385300 –0.9825039 0.98317
1
–0.9955
1
Ionization
Potential
Clog P
0.999992 0.99834338 0.994223 –0.993824
0.924867 –0.956837
1
–0.997094 –0.99989108 –0.983867 0.983207 –0.894967
0.933333 –0.997385
1
Table 6. Observed and calculated rates of oxidation of synthesized compounds, 3a–d.
Compd.
Obs.
RO
Calcd.
RO (I)
Calcd.
RO (Hf)
Obs.
t (min)
½
Calc. (I)
t (min)
½
Calc.(Hf)
t (min)
½
3a
3Јa
3b
3c
3d
Nd
Nd
0.0143
0.0086
0.0042
0.028437
–0.2864592
0.01431
0.008577
0.004213
0.0176386
–0.0278025
0.0144609
0.00827247
0.00445077
nd
nd
48.46
80.85
165
24.38
2.42
48.43
80.88
164.49
52.34
24.92
47.92
3.77
155.70
nd – not determined.
Calcd. RO (Hf) = 0.06841 + 0.001399 (Hf)
r = 0.9981, n = 3
(7)
on oxidation. The calculated oxidation rate of the acid
with respect to its ionization potential reveals that oxida-
tion occurs mainly in the ionized form (t of 3Јa, is about
½
one tenth that of 3a).This observation confirms the sug-
gested sequential pathway of tested compounds, i.e.,
oxidation occurs after hydrolysis of the ester moiety.
It canalso be estimatedaccording toionization potential,
using equation (8).
Calcd. RO (I) = –0.6564 + 0.0823
(Ionization potential)
r = 0.99999, n = 3
(8)
Conclusions
Individual components of several synthetic mixtures and
the rate of oxidation of DH-derivatives were simultane-
ously determined using spectrophotometric measure-
ments together with CLS multivariate calibration analy-
sis.The high recoveries obtained in all cases, as well as
the reliable rates of oxidation which agree well with other
reports and analysis methods, proved that the proposed
technique could be efficiently applied for determination
of oxidation rates and other kinetic parameters in cases
of severe spectral overlapping, and/or presence of seri-
Consequently, oxidation rate of DH-acid 3a in its both
ionized (3Јa) and non ionized forms can also be estimat-
ed according to equation 7 or equation 8.The calculated
rates of oxidation of tested compounds, 3a–d, are
shown inTable 6.It was observed that calculated rates of
oxidation according to the ionization potential are almost
the same as the experimental values. However, those
obtained in relation to heat of formation deviate slightly,
indicating a pronounced influence of ionization potential
© 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

582 Mahmoud et al.
Arch. Pharm. Pharm. Med. Chem. 2003, 336, 573–584
ous interferences with other methods such as HPLC.
Kinetic oxidation data for the tested compound revealed
that this new CDS has a reasonable oxidation rate for ef-
ficient brain delivery.
IR (cm–1): 3020, 1732. 1H-NMR (CDCl3): 3.5 (s, 3 H, CH3), 6.1
(s, 2 H, N+-CH2), 7.3–8.3 (m, 6 H, isoquinoline), 10.6 (s, 1 H,
isoquinoline C1(H)).
2-(Ethoxycarbonylmethyl)isoquinolinum bromide (2 c)
White residue washed with CHCl3 and dried to give 4.5 g
(77.5 %) of a white solid product, m.p. 206–209 °C as reported
[35].
Experimental
IR (cm–1): 3021, 1730. 1H-NMR (DMSO-d6): 1.1–1.5 (t, 3 H, J =
2.4, CH3CH2), 4.1–4.5 (q, 2 H, CH3CH2), 6.1 (s, 2 H, N+-CH2),
8–9 (m, 6 H, isoquinoline ), 10.5 (s, 1 H, isoquinoline C1(H)).
General
Melting points were determined with an electrothermal appara-
tus (Stuart Scientific, Staffordshire, England) and were uncor-
rected.
2- (Isopropyloxycarbonylmethyl)isoquinolinum bromide (2 d)
An oily residue, triturated with ether, filtered and recrystallized
from ether to give 2 g (71 %) of a white solid, m.p. 136–139 °C.
IR spectra were recorded as KBr disks on a Shimadzu IR 200–
91527 spectrophotometer (Shimadzu, Tokyo, Japan).
IR (cm–1): 3020 1729. 1H-NMR (DMSO-d6): 1.3 (d, J = 3, 6 H,
CH(CH3)2), 4.8–5.3 (m, 1 H, CH(CH3)2), 6.2 (s, 2 H, NH+-CH2),
7.8–8.8 (m, 6 H, isoquinoline ) 10.6 (s, 1 H, isoquinoline C1(H)).
Elemental analysis for C14H16BrNO2 · 1/2 H2O: calcd: C, 52.68;
H, 5.368; N, 4.388. found: C, 52.68; H, 5.313; N, 4.371.
1H-NMR spectra were obtained on Varian EM-360L (60 MHz),
TMS was used as internal standard, CDCl3 or DMSO-d6 were
used as solvents, and the chemical shifts are expressed in δ
(ppm).
UV measurements were carried out on a computerized UV-
1601, UV-visible Shimadzu spectrophotometer (Shimadzu, To-
kyo, Japan), using 1.00 cm quartz cells. The spectral data ob-
tained were saved in PC Shimadzu program and the subse-
quent statistical manipulation was performed by transfer of
spectral data to Microsoft Excel 2000 and processed with the
standard curve fit package and matrix calculations.
Synthesis of 1,2-Dihydroisoquinoline-N-methylcarboxylate de-
rivatives (3 a–d)
Method A
A solution of isoquinolinum bromide derivative (0.001 mol) in
deareated water (50 mL) was cooled in an ice-bath. Dichlo-
romethane (100 mL) was added while stirring followed by addi-
tion of sodium bicarbonate (0.5 g, 0.006 mol) and sodium
dithionite (0.7 g, 0.004 mol). The biphasic mixture was stirred
under nitrogen in an ice bath for 3 h.The solution was then acid-
ified with dil.HCl in the case of compound 3 a.The organic layer
was removed, washed with water, dried over anhydrous sodi-
um sulfate and evaporated to give the DH-products.
Elemental microanalyses were performed at the central labora-
tory of Assiut University.
Reactions were monitored by TLC using D Alufolien, Kieselgel
60 F 254 precoated plates (Merck, Darmstadt, Germany) and
were visualized by UV lamp.
Purified DH-derivatives and their corresponding quaternary
forms were used as standard compounds for UV-measure-
ments and the calibration phase obtained.The latter was sub-
sequently used for determining oxidation rates after confirma-
Method B
Freshly prepared 1-benzyl-1,2-dihydroisonicotinamide rea-
gent [37], (0.310 g, 0.0011 mol) was added to an ice-cooled so-
lution of isoquinolinum bromide derivative (0.297 g, 0.0011
mol) in methanol (50 mL) while stirring.The mixture was kept at
0 °C under nitrogen for 3 hr, the solution was then concentrated
followed by addition of ether.The precipitate formed was sepa-
rated by filtration to separate the quaternary reagent and the
filtrate evaporated under reduced pressure to give the dihydro
products.
1
tion of compound structures by CHN analysis, H-NMR and
melting point measurement.
Oxidizing reagent was 10 % methanolic silver nitrate solution.
Chemicals and solvents used throughout this work were of ei-
ther reagent or analytical grade. Solvents were distilled and
dried before use.
Synthesis of isoquinolinum bromide derivatives (2 a–d)
1,2-Dihydro-N-carboxymethylisoquinoline (3 a)
Bromoacetic acid or its esters (0.02 mol) were added dropwise
to a stirring solution of isoquinoline (1.65 g, 0.012 mol) in ace-
tone (20 mL).The reaction mixture was then refluxed while stir-
ring, up to 24 hrs. The solvent was then evaporated under re-
duced pressure to give quaternary products.
An oily product, produced in 62 % yield by method A.
1H-NMR (CDCl3): 3.6 (s, 2 H, N-CH2), 4.3 (s, 2 H, isoquinoline
C1H2), 7.1–7.8 (m, 6 H, isoquinoline).
1,2-Dihydro-2-( methoxycarbonylmethyl)isoquinoline (3 b)
N-Carboxymethylisoquinolinum bromide (2 a)
A brown oily product result in 60 % yield by method A and 72 %
by method B.
A pale rose colored solid, recrystallized from ether to yield 2.68
g (78 %), with m.p. 213–217 °C.
1H-NMR (CDCl3): 1.2 (s, 3 H, CH3), 4.0 (s, 2 H, N-CH2), 4.6 (s,
2 H, isoquinoline C1H2), 7.1–7.7 (m, 6 H, isoquinoline).
IR (cm–1): 3460, 3116, 1720. 1H-NMR (DMSO-d6): 5.3–5.5 (s,
2 H, N+-CH2), 7.8–8.8 (m, 5 H, isoquinoline), 10.6 (s, 1 H, iso-
quinoline C1(H)).Elemental analysis for C11H10BrNO2:calcd:C,
49.28; H, 3.76; N, 5.22. found: C, 48.96; H, 4.04; N, 5.26.
1,2-Dihydro-2-(ethoxycarbonylmethyl)isoquinoline (3 c)
A brown oily product result in 64 % yield by method A and 75 %
by method B.
2-(Methoxycarbonylmethyl)isoquinolinium bromide (2 b)
An oily residue triturated with ether in order to obtain a yellow
solid, 5.8 g (95 %) of the title compound, m.p 162–164 °C as re-
ported [36].
1H-NMR (CDCl3): 1.3 (t, J = 2.6, 3 H, CH3CH2), 3.7 (q,
2 H,CH3CH2), 4.0 (s, 2 H, N-CH2), 4.6 (s, 2 H, isoquinoline
C1H2), 7–7.6 (m, 6 H, isoquinoline).
© 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Arch. Pharm. Pharm. Med. Chem. 2003, 336, 573–584
1,2-Dihydroisoquinolines as Carriers for Specific Brain Delivery 583
1,2-Dihydro-2- (isopropoxycarbonylmethyl)isoquinoline (3 d)
scanned in the 270–37 nm range. Data obtained was analyzed
as mentioned above to determine the concentration of the DH-
compound and its quaternary form at various intervals.The re-
maining DH concentration or formed quaternary salts were cal-
culated and the apparent pseudo first order rate constants of
oxidation (Kdisap., min.–1) of the tested compound calculated as
the average of 3 experiments. Results are given in Table 4.
A brown oily product result in 59 % yield by method A and 70 %
by method B.
1H-NMR (CDCl3):1–1.3 (d, J = 2.1, 6 H, CH(CH3)2), 3.9 (s, H, N-
CH2), 4.4 (s, 2 H, isoqunoline C1H2), 5–5.4 (m, 1 H CH(CH3)2),
7–7.5 (m, 6 H, isoquinoline).
Theoretical calculations
MO calculations
References
The MOPAC 6/PC version 6.12 package program was used.
The program was obtained from Quantum Chemistry Program
Exchange project (QCPE), Department of Chemistry, Indiana
University, USA. Structures studied were generated using
PCMODEL program which was obtained from Serena Soft-
ware, USA. MMX molecular mechanics was performed with
PCMODEL minimization of the energy of these compounds.
The resulting files obtained from PCMODEL were introduced to
MOPAC program for AM1 calculation of the heat of formation
(∆Hf, Kcal/mole) in gaseous phase. These calculations were
carried out on an IBM PIV 333 Processor, 124Ram computers.
[1] N. Bodor, Adv. Drug Deliv. Rev. 1994, 14, 157–166.
[2] N. Bodor, M. E., Brewster, Pharmacol. Ther. 1983, 19,
337–386.
[3] D. J. Begley, J. Pharm. Pharmacol. 1996, 48, 136–146.
[4] N. Bodor, H. Farag, J. Pharm. Sci. 1984, 73, 385–389.
[5] K. S. Choi, G. A. Alivisatos, Biochemistry, 1968, 7, 190–
197.
[6] E.Pop, M.E.Brewster, M.J.Hung, N.Bodor, J. Mol. Struct.
(Theochem) 1995. 337, 49–55.
Calculation of log P values
The log P values of synthesized derivatives were computed
with a routine method called “calculated log P” (C log P) con-
tained in a PC-software package (MacLogP 2.0, BioByte Corp.,
CA, USA). A representation of the molecular structure where
hydrogens are omitted, or “suppressed” (SMILES notation),
was entered into the program, which computed log P based on
the fragment method developed by Leo [38].
[7] S.L. Johnson, P.T.Tuazon, Biochemistry 1977, 16, 1175–
1183.
[8] D. M. Stout, A. I. Meyers, Chem. Rev. 1982, 82, 223.
[9] E.Pop, W.Wu, E.Shek, N.Bodor, J.Med.Chem. 1989, 32,
1774–1781.
[10] W.Wu, E.Pop, E.Shek, N.Bodor, J.Med.Chem.,1989, 32,
UV spectral measurements
1782–1788.
Standard solutions for multivariate calibration
[11] E. Pop, Z. Liu, J. Vlasak, W. Anderson, M. E. Brewster, N.
Bodor, Drug Del. 1993, 1, 143–149.
To obtain the calibration matrix for applying CLS analysis, six
solutions of each of the pure compounds (2 b–d and 3 b–d)
were prepared with concentrations in the range 10–60 µg/mL.
This range was previously verified to obey Beer’s law for each
of the compounds studied in the selected solvent (methanol).In
order to obtain the calibration K matrix, absorption data in the
range of 270–370 nm (digitized every 1.0 nm) were subjected
to least squares analysis (see below). Laboratory mixtures
were then prepared by mixing known amounts of DH-deriva-
tives, 3 b–d, with their corresponding Q-forms, 2 b–d, in varied
proportions, in order to verify the precision of the method for
analysis of such mixtures and determination of DH-compound
oxidation rates, as well as the rate of formation of correspond-
ing quaternary ammonium salts.
[12] A. G. Anderson, G. Berkehammer, J. Am. Chem. Soc.
1958, 80, 992–999.
[13] M. A. Abo-Elnile, Master Thesis, Assuit University, Egypt,
1997.
[14] K. Mahalanabis, D. Basu, Analyst 1989, 114, 1311–1314.
[15] H. C. Goicoechea, A. C. Olivieri, Analyst 1999, 124, 725–
731.
[16] C. Demir, R. G. Brereton, Analyst 1998, 123, 181–189.
[17] N. Capitan-Vallvey, M. Kalil, A. Deheidel, I. de Orbe, R.
Avidad, Analyst 1999, 124, 49–53.
Data processing
[18] M.J. Benito, M. C. Ortiz, M. S. Sanchez, L. A. Sarabia, M.
Iniguez, Analyst 1999, 124, 547–552.
Data was processed on an Intel Pentium III 750 MHz PC-com-
patible computer. The relative root mean squared error
(RRMSE) was chosen to compare the different models:
[19] J. A. Hernandez-Arteseros, R. Compano, R. Ferrer, M. D.
Prat, Analyst 2000, 125, 1155–1158.
[20] J. Keay in Advance in Heterocyclic Chemistry, (Ed.: A.
Katritzky), Academic Press Inc., NewYork, NY 1986, Vol.
39, pp. 33–35.
2
ˆ
Σ(Ci – Ci)
RRMSE (%) = 100 ×
s
ΣC2i
[21] Y. Kikugawa, M. Kuramoto, I. Satio, S. Yamado, Chem.
Pharm. Bull. 1973, 21, 1914–1926.
compound in standard or sample solutions.
[22] H.Schmid, P.Karrer, Helv.Chim.Acta, 1949, 32, 960–966.
Oxidation with silver nitrate
[23] P.Karrer, F.Kahnt, R.Epstein,W.Jaffe,T.Ishii, Helv.Chem.
Acta 1938, 21, 223–236.
To 3 sets of tubes (each containing 5 mL of AgNO3 reagent), 1
mL of stock solutions of tested DH-derivatives (3 b–d) was add-
ed at ambient temperature. At time intervals (5, 10, 15, 30, 40,
60 min), samples were removed, centrifuged, filtered and
[24] A.Nuvole, G.Paglietti, P.Sanna, R.M.Achesson, J.Chem.
Res. 1984, Supplement, 356–357.
© 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

584 Mahmoud et al.
Arch. Pharm. Pharm. Med. Chem. 2003, 336, 573–584
[25] C. Koelsch, N. Albertson, J. Am. Chem. Soc. 1953, 75,
[33] M. E. Brewester, J. Kaminski, M. Huang, N. Bodor, J. Org.
Chem. 1990, 55, 2361–2366.
2095–2097.
[26] W.Bradley, S.Jeffrey, J.Chem.Soc.1954, 76, 2770–2778.
[34] E. Pop, M. J. Huang, N. Bodor, Internat. J. Quant. Chem.
Quantum Biology Symposium 1992, 19, 77–85.
[27] R.Fryer, J.Earley, E.Evans,V.Schneider, L.Sternbach, J.
Org. Chem. 1970, 35, 2455–2459.
[35] H. C. Goicoechea, A. C. Olivieri, Talanta 1998, 47, 103–
108
[28] J. Padbury, H. Lindwall, J. Am. Chem. Soc. 1945, 67,
1268–1270.
[36] A. Wander, British Patent, 1963, No 916, 507.
[29] M. Johnson, J. Chem. Soc. 1964, 200–213.
[37] G.Paglietti, P.Sanna, A.Nuvole, F.Soccolini, R.M.Achhe-
son, J. Chem. Res. (m) 1983, 2326–2342.
[30] M. J. Phelan, N. Bodor, Pharm. Res. 1989, 6, 667–676.
[38] A. J. Leo, Chem. Rev. 1993, 93, 1281–1306
[31] M. J. S. Dewar, E. G. Zoebisch, E. F. Healy, J. P. Stewart, J.
Am. Chem. Soc. 1985, 107, 3902–3909.
[32] M. Powell, T. Bruice, J. Am. Chem. Soc. 1983, 105, 7139–
7149.
Key-Technologies of the Future
VINCENZO BALZANI, Italy,
ALBERTO CREDI, and
MARGHERITA VENTURI, all of
University of Bologna, Italy
nanotechnology is already being
pursued in information technology,
with many other branches about to
follow.
Written by Vincenzo Balzani and
his team of co-authors, this book
provides a comprehensive
overview of this fascinating and
multifarious topic, offering
newcomers and experts alike a
wealth of information.
Molecular Devices and
Machines
A Journey into the Nanoworld
2003. XVII, 494 pp Hbk
The miniaturization of bulky
devices and machines is a
process that confronts us on a
daily basis. However, nanoscale
machines with varied and novel
characteristics may also result
from the enlargement of
extremely small building blocks,
namely individual molecules.
This bottom-up approach to
•
•
Most important new
technologies for a broad
audience
ISBN 3-527-30506-8
€ 99,-* / sFr 146,-
Unifying, critical, and
stimulating overview of
red-hot topics
*The Euro price is only valid for Germany.
Wiley-VCH Verlag
Register now for the free
WILEY-VCH Newsletter!
www.wiley-vch.de/home/pas
Fax: +49 (0)6201 606 184
e-Mail: service@wiley-vch.de
www.wiley-vch.de
50303032_kn
© 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim