Identification of N-benzyltetrahydroisoquinolines as novel anti-neuroinflammatory agents

Article abstract of DOI:10.1016/j.bmc.2018.10.020

A series of simplified berberine analogs was designed, synthesized, and evaluated for anti-inflammatory activity. SAR studies identified N-benzyltetrahydroisoquinoline 7d as a potent berberine analog. 7d suppressed LPS-induced inflammatory cytokine levels

Full text of DOI:10.1016/j.bmc.2018.10.020

Accepted Manuscript
Identification of N-benzyl tetrahydroisoquinolines as novel anti-neuroinflam-
matory agents
Brian Gabet, Ping-Chang Kuo, Steven Fuentes, Yamini Patel, Ahmed Adow,
Mary Alsakka, Paula Avila, Teri Beam, Jui-Hung Yen, Dennis A. Brown
PII:
S0968-0896(18)31531-1
https://doi.org/10.1016/j.bmc.2018.10.020
BMC 14579
DOI:
Reference:
Bioorganic & Medicinal Chemistry
To appear in:
Received Date:
Revised Date:
Accepted Date:
29 August 2018
10 October 2018
20 October 2018
Please cite this article as: Gabet, B., Kuo, P-C., Fuentes, S., Patel, Y., Adow, A., Alsakka, M., Avila, P., Beam, T.,
Yen, J-H., Brown, D.A., Identification of N-benzyl tetrahydroisoquinolines as novel anti-neuroinflammatory agents,
Bioorganic & Medicinal Chemistry (2018), doi: https://doi.org/10.1016/j.bmc.2018.10.020
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Identification of N-benzyl tetrahydroisoquinolines as novel anti-neuroinflammatory agents
Brian Gabeta, Ping-Chang Kuob, Steven Fuentesa, Yamini Patela, Ahmed Adowa, Mary Alsakkaa,
Paula Avilaa, Teri Beama, Jui-Hung Yenb, and Dennis A. Browna*
aManchester University College of Pharmacy and Health Sciences, 10627 Diebold Rd, Fort
Wayne, IN 46845
bDepartment of Microbiology and Immunology, Indiana University School of Medicine, 2101 E.
Coliseum Blvd, Fort Wayne, IN
*Corresponding author.
Contact information:
Dennis Brown, PhD
Associate Professor – Medicinal Chemistry
Manchester University College of Pharmacy and Health Sciences
10627 Diebold Rd
Fort Wayne, IN 46845
dabrown@manchester.edu

Abstract
A series of simplified berberine analogs was designed, synthesized, and evaluated for anti-
inflammatory activity. SAR studies identified N-benzyl tetrahydroisoquinoline 7d as a potent
berberine analog. 7d suppressed LPS-induced inflammatory cytokine levels in both BV2 cells
and primary microglia. Taken together, our results suggest that simplified BB analogs have
therapeutic potential as a novel class of anti-neuroinflammatory agents.

1. Introduction
Multiple sclerosis (MS) is an autoimmune disorder that is characterized by the deterioration of
the insulating myelin sheath of neurons in the central nervous system1. The interactions of
antigen-presenting CNS microglia, CNS infiltrating myelin-specific T cells, and other cells result
in the release of pro-inflammatory mediators, ultimately causing brain inflammation leading to
myelin deterioration2. The symptoms of the disorder are related to the extent of neuronal
damage and the location of attack, but can include numbness/weakness in the limbs, vision
problems, tremor, slurred speech, electric-shock sensations, fatigue, and bowel and bladder
control problems3.
The importance of inflammation in MS is underscored by the immunomodulatory activity of the
current MS therapies. While there are over a dozen disease-modifying MS agents available for
the most common type of the disease (relapsing/remitting), there is currently no cure.
Additionally, nearly one third of patients develop severe disability. As such, the identification of
novel anti-neuroinflammatory agents is warranted.
Figure 1. Structure of berberine chloride.
The benzylisoquinolium alkaloid berberine (BB, Figure 1), found in several plant species, has
been shown to possess multiple interesting pharmacological properties, including anticancer,
antibacterial, antifungal, antidiabetic, and anti-inflammatory activities4. Excitingly, BB has been
demonstrated to ameliorate inflammatory processes in the experimental autoimmune
encephalomyelitis (EAE) model of neuroinflammation5, 6. This anti-inflammatory activity has
been attributed to modulation of numerous cellular pathways in immune cells, including AMPK,
MAPK, Nrf2, and NF-κB7.
While BB has promising anti-inflammatory activity, it has limited utility as a therapeutic agent.
BB suffers from poor solubility, limited absorption, and low bioavailability8, 9. Additionally, BB is
reported to undergo extensive metabolism with the production of multiple metabolites, some
of which are pharmacologically active10. The unfavorable pharmacokinetic properties of BB
likely are a consequence of several features of its chemical structure: the presence of a
contiguous ring system, a permanent positive charge and metabolically-sensitive isoquinolium
and alkoxy moieties.
The drug-like properties of BB can likely be improved via structural modification. Numerous
structure-activity relationships (SAR) studies have been reported for BB; however, most of

these studies evaluated analogs that are alkylated and/or phenolic derivatives of BB11-13. While
many of these compounds have displayed interesting activities, most are sufficiently similar to
BB to still suffer from poor physicochemical properties. Furthermore, few of these studies
explored anti-inflammatory activity. Herein, we report a structure-activity relationships study
of simplified BB analogs as anti-neuroinflammatory agents.
2. Chemistry
2.1 Design of initial series of molecules
To probe the structural features of BB that contribute to its anti-inflammatory activity, a series
of structurally-diverse molecules was designed (1-7a). The roles of the stationary positive
charge of the isoquinolium nitrogen, along with the role of planarity on ring D, were evaluated
by designing tetrahydroberberine 1 and benzalkonium 2. The functions of rings C and D were
assessed by designing molecules 5-7a.
Ph
O
O
O
O
BnBr, CH3CN
reflux, 93%
NaBH4
Br
N
N
BB
H
H
MeOH, 71%
O
O
O
2
1
O
O
O
O
O
O
MeI, CH3CN
reflux, 74%
N
1. POCl3, CH2Cl2
HN
O
NH
O
5
O
O
2. NaBH4, MeOH
67%
O
4
3
O
O
O
R1
R2
R1
R2
O
O
NH
O
I
N
NaBH(OAc)3, AcOH
CH2Cl2
MeI, CH3CN
reflux, 59%
N
4
7a-o
6
R3
O
+
R3
O
O
O
O
PyBOP, Et3N
DMF, 79%
O
N
O
O
+
OH
NH
O
O
10
8
O
O
9
Scheme 1. Synthesis of simplified berberine analogs.
2.2 Synthesis

Compound 1 was prepared as previously described by reducing berberine chloride with NaBH4
in MeOH14. Analog 2 was synthesized by treating 1 with benzyl bromide in refluxing CH3CN11.
Treatment of amide 3 with first POCl3 followed by NaBH4 provided secondary amine 415.
Methylation of 4 with MeI in refluxing CH3CN gave amines 516 and 6. N-benzyl analogs 7 were
prepared by reductive amination of appropriate benzaldehydes with corresponding
tetrahydroisoquinolines using NaBH(OAc)3 in dichloroethane in good yields. Amide 10 was
prepared by activating acid 9 with PyBOP with subsequent coupling to amine 8.
3. Results and discussion
3.1 Evaluation of BB analogs and structure-activity relationship analysis
Tumor necrosis factor (TNF) is one of several important pro-inflammatory cytokines produced
by immune cells, including microglia17. TNF has been found to be elevated at sites of active MS
lesions and contributes to processes that damage oligodendrocytes and myelin18-20. Given the
importance of TNF in neuroinflammation, synthesized BB analogs were evaluated for their
ability to suppress TNF levels induced by lipopolysaccharide (LPS) in the BV2 immortalized
mouse microglial cell line. BV2 cells were pretreated with test compounds at a concentration
of 50 μM for 1 h, then stimulated with LPS for 24 h, at which time the media was collected and
assayed for TNF levels via ELISA. BB was evaluated at a concentration of 1 μM as higher
concentrations did not provide additional TNF suppression (data not shown). The results are
expressed in Table 1 and are expressed as a percentage of LPS.
As shown in Table 1, treatment of BV2 cells with LPS caused a ten-fold increase of TNF levels.
Synthesized molecules demonstrated a range of activities, from significant TNF suppression to
minimal. Interestingly, the presence of a permanent cationic nitrogen was as a common
feature of the most potent molecules assayed (BB and 2). The nature of the cationic amine
appears to be less important, as both isoquinolium (BB) and benzalkonium cations (2) were
tolerated. However, enhanced activity does not seem to apply to cation 6 as minimal TNF
suppression was observed (97% of LPS).
Entry
DMSO
TNF levels (% LPS)
10 ± 2.3
LPS
100 ± 0
BB (1 uM)
25 ± 8.2
1
2
45.0 ± 12
22 ± 9.1
4
77 ± 2.4
5
77 ± 5.7
6
97 ± 7.0
63 ± 8.6
7a 1-9 (R1, R2 = -CH2OCH2-), R3 = 2, 3-OMe
Table 1. Suppression of TNF by synthesized compounds. BV2 cells were treated with DMSO or experimental
compounds (50 μM) for 1 hour, followed by LPS (100 ng/mL) activation for 24 hour. The culture media was
collected and TNF levels evaluated by ELISA. Data are expressed as mean ± SEM and are representative of three
different experiments.

One of our approaches to improve the physicochemical properties of BB was to eliminate its
permanent positive charge. We were pleased to see that activity can be retained by
substitution of the isoquinolium core of BB for tetrahydroisoquinoline (1). This molecule
suppressed TNF levels to 45% of LPS, suggesting that this portion of the BB molecule is
amendable to modification.
Ring connectivity also appears to influence TNF suppression. The activity of molecules with the
D ring of BB removed appears to depend on the site of C-D ring junction. Compound 7a,
connected to ring D via N9, was more active (63% of LPS) than the analogs 4-6 (77, 77, and 97 %
of LPS, respectively), all of which are connected via C18. Indeed, the introduction of an
asymmetric center into these compounds may also impact activity.
While 1 and 2 were the most active of the molecules synthesized, their chemical structures still
possessed features that were not ideal, namely positive charges and/or a contiguous ring
system. N-Benzyl tetrahydroisoquinoline 7a, with its modest activity and synthetic simplicity,
seemed a good candidate for optimization. Thus, we designed, synthesized, and evaluated a
second series of analogs in an effort to increase TNF-suppression. The results are shown in
Table 2.
Entry
DMSO
LPS
7b
7c
7d
7e
7f
7g
7h
7i
7j
7k
7l
7m
7n
7o
R1
R2
R3
TNF levels (% LPS)
10 ± 2.3
100 ± 0
95 ± 9.4
90 ± 13
H
H
-CH2OCH2-
-CH2OCH2-
-CH2OCH2-
-CH2OCH2-
-CH2OCH2-
-CH2OCH2-
-CH2OCH2-
-CH2OCH2-
H
2, 3-OMe
Ph
2, 6-OMe
3, 5-OMe
4-OMe
4-CF3
2, 4-OMe
2-OMe
39 ± 10
89 ± 5.2
88.± 6.8
102 ± 14
103 ± 13
78 ± 4.0
78 ± 9.6
78 ± 11
101 ± 1.3
107 ± 6.9
82 ± 5.0
64 ± 9
3-OMe
H
2, 6-OMe
2, 6-Me
2, 6-Cl
2, 6-OMe
2, 6-OMe
-CH2OCH2-
-CH2OCH2-
H
OMe
OMe
H
10
73 ± 11
Table 2. Suppression of TNF by synthesized compounds. BV2 cells were treated with DMSO or experimental
compounds (50 μM) for 1 hour, followed by LPS (100 ng/mL) activation for 24 hour. The culture media was
collected and TNF levels evaluated by ELISA. Data are expressed as mean ± SEM and are representative of three
different experiments.
Compound 7b lacks the methylenedioxy group of 7a. This moiety appears to be important for
activity as TNF suppression was greatly reduced by its removal (95% of LPS). The substitution of
the N-benzyl group was next explored by assaying compounds 7c-j. Excitingly, a substitution

pattern 2, 6-dimethoxy (7d) resulted in the most TNF suppression (39% of LPS, compared to
61% for 7a). In contrast, 2, 4- and 3, 5-dimethoxy substituted analogs 7h and 7e gave minimal
activity, as did monomethoxy congeners 7f, i and j, and trifluoromethyl analog 7g. N-Benzyl
analog 7c also showed minimal activity (89% of LPS).
The introduction of a 2, 6-dimethoxy moiety into compound 7a clearly augmented its activity;
however, it was not clear the extent to which the activity of 7d could be attributed to this
substitution. As such, we synthesized and assayed compounds 7k-o. Confirming what was
observed for 7c, removal the methylenedioxy of 7d (7k) weakened TNF suppression (79% of
LPS). Interestingly, this activity can be partially restored by incorporation of a 6-OMe (7o, 64%
of LPS), but to a lesser extent than a 7-OMe (7n, 82% of LPS) group. The electronic properties
of the 2, 6-dimethoxy moiety of 7d appear to be crucial, as 2, 6-dimethyl (7l) or 2, 6-dichloro
groups (7m) resulted in loss of TNF suppression (101 and 107% of LPS, respectively).
As a cationic nitrogen was seen to give the most activity in BB and 1, it appeared likely that an
ionizable nitrogen is needed for 7d’s activity. This was probed by synthesizing and assaying
amide 10. Expectedly, this molecule was quite weaker than parent 7d (73 vs. 39% of LPS,
respectively). While this decrease in activity is likely due to the loss of ionizability, introduction
of conformational restriction may also be involved.
The rationale for the improved activity of 7d over 7a is not clear. However, examining low
energy conformations of BB and 7d yields a possible explanation (Figure 2). The minimized
conformation of BB produces an essentially flat structure due to the high degree of sp2
hybridization in the molecule. The minimized conformation for 7d suggests that the
tetrahydroisoquinoline and N-benzyl ring systems are in a perpendicular relationship to each
other. Overlaying these two structures shows the methylenedioxy, amine, and 12-methoxy
groups of BB lining up with their counterparts in 7d. The other methoxy group of 7d does not
seem to overlap with a group in BB. It is possible that this methoxy group is accessing an
additional binding site in its target. As the 6 position of the N-benzyl ring of 7a is unsubstituted,
this interaction is not possible, and may contribute the weaker activity compared to 7d.
Figure 2. Low energy conformations of BB and 7d using Chem3D.
3.2 Suppression of pro-inflammatory cytokines by 7d

As described above, the production of pro-inflammatory cytokines and mediators are important
to the pathophysiology of MS. We next assayed the suppression of these molecules by various
concentrations of 7d in LPS-stimulated BV2 cells. Treatment with LPS alone caused a large
increase in TNF release (ten-fold compared to DMSO, Figure 3). Interestingly, treatment with
7d was able to dose-dependently decrease TNF protein levels was seen beginning at a
concentration of 12 μM, and continued to decrease up to 50 μM. IL-6, another pro-
inflammatory cytokine, has been demonstrated to be upregulated in both the plasma and CNS
of MS patients21, 22. As seen for TNF, 7d was able to suppress LPS-mediated IL-6 induction in a
dose-dependent manner, in BV2 cells, beginning at a concentration of at 25 μM.
Figure 3. Suppression of TNF, IL-6, and nitrate by 7d. BV2 cells were treated with 7d at various concentrations for
one hour, followed by LPS (100 ng/mL) for 24 hours. Culture media was then collected and assayed for respective
inflammatory products. Data are expressed as mean ± SEM and are representative of three different experiments.
Statistical significance determined as * p < 0.05, ** p < 0.01 and *** p < 0.001.
Nitric oxide (NO) is produced by microglia in response to inflammatory stimuli, and is known to
contribute to oligodendrocyte injury, demyelination, and axonal degeneration23. As NO is
converted rapidly to nitrite under physiological conditions, levels of NO2 were assayed in LPS-
stimulated BV2 cells. As seen in Figure 3, LPS treatment induced significant nitrite production.
However, treatment with 7d ameliorated this increase dose-dependently.
The interleukins IL-12, formed from p35 and p40 subunits, and IL-23, formed from p19 and p40
subunits, are important in the development of the pro-inflammatory TH1 and TH17 helper T cell
phenotypes, respectively24. Both of these cell types have are associated with EAE and MS. As
such, the mRNA expression of these molecules, along of IL-1β, GM-CSF, TNF, and IL-6 was
assessed in LPS-stimulated primary mouse microglia cells. As shown in Figure 4, expression of
these genes was increased upon LPS stimulation. Excitingly, however, treatment with 7d (100
μM) suppressed all but IL-12p40 and TNF, after both 2 and 4 hours of LPS exposure. The
suppression of inflammatory cytokines and mediators by 7d in LPS-stimulated BV2 cells and
primary microglia suggest that this compound has anti-neuroinflammatory activity.

Figure 4. Suppression of inflammatory cytokines by 7d. Primary microglia cells were treated with 7d (100 uM) for
one hour, followed by stimulation with LPS (100 ng/mL) for 2 and 4 hours. Cells were then harvested and
subjected to RNA extraction followed by QPCR analysis of IL-1β, IL-23p19, IL-12p35, IL-12p40GM-CSF, TNF, and IL-
6. Data are expressed as mean ± SEM and are representative of three different experiments. Statistical
significance determined as * p < 0.05, ** p < 0.01 and *** p < 0.001.
The current unmet medical in need in MS treatment warrants the development of additional
therapeutic options. The identification of novel small molecules that target additional
inflammatory targets may improve treatment. The anti-inflammatory activity of simplified
analog 7d represents an advancement in the exploration of BB as an immunomodulatory agent.
Future planned studies include the determination of mechanism of action of 69, additional
structure-activity relationship studies to improve potency, and evaluation the in vivo EAE model
of neuroinflammation.
4. Conclusion
The isoquinolium alkaloid BB is widely known for its multiple pharmacological activities.
However, the therapeutic potential of this interesting natural product is hampered by

unfavorable pharmacokinetic properties arising from aspects of its chemical structure. The
results of our studies have delineated the important structural features of BB that influence its
anti-inflammatory activity. These studies identified simplified lead compound 7d, which
suppressed LPS-induced inflammatory mediators in both the BV2 cell line and primary
microglia, suggesting a potential role for simplified BB analogs as anti-neuroinflammatory
agents.
5. Experimental
5.1 Chemistry
All solvents and reagents obtained from commercial sources were used without further
purification, unless otherwise noted. Berberine chloride was purchased from TCI (Tokyo,
Japan). All reactions were performed under an inert argon atmosphere, unless otherwise
indicated. The purity of all final molecules was judged by elemental analyses (C, H, N) and were
1
found to be within ± 0.4% of theoretical values. H and 13C NMR analyses were performed on a
Varian Mercury 300 spectrophotometer at 300 and 75 MHz, respectively. Chemical shifts are
given in ppm in reference to tetramethylsilane (TMS) as an internal standard. Multiplicities are
given as s (singlet), d (doublet), t (triplet), m (multiplet), and br s (broad signal). High-resolution
mass spectral data were obtained on a Waters Acquity TQD triple quadrupole mass
spectrometer. Melting points were taken on a Mel-Temp apparatus and are uncorrected. Thin
layer chromatography (TLC) was performed on silica gel 60 F254-coated glass plates purchased
from EMD Millipore, and visualized with UV light and/or basic KMnO4. Rf values correspond to
the free base, where applicable. Low energy conformations were generated using Chem3D.
General reductive amination procedure. To a solution of tetrahydroisoquinoline (1.41 mmol)
and AcOH (2.82 mmol) in (CH2Cl)2 (5mL) at room temperature was added benzaldehyde (1.41
mmol). The resulting solution was stirred for 15 min, at which time NaBH(OAc)3 (2.82 mmol)
was added. The reaction was allowed to proceed overnight, poured into 1 M NaOH (10 mL),
and extracted 3 times with CH2Cl2. The pooled organic extracts were washed with H2O, dried
over Na2SO4, filtered and concentrated. The crude residue was then purified by flash

chromatography (CH2Cl2/MeOH). Purified N-benzylamines were next converted to their
corresponding hydrochloride or oxalate salts.
5-(3,4-dimethoxybenzyl)-6,6-dimethyl-5,6,7,8-tetrahydro-[1,3]dioxolo[4,5-g]isoquinolin-6-ium
iodide (6). Tetrahydroisoquinoline 5 (31 mg, 0.095 mmol) and MeI (0.500 mmol) were refluxed
in acetonitrile for 4 hours. The reaction mixture was cooled and concentrated. The crude solid
1
was recrystallized from isopropyl alcohol to afford 6 as a white solid (56 mg, 59%). H NMR
(300 MHz, DMSO-d6) δ 6.91 – 6.83 (m, 2H), 6.69 (d, J = 2.1 Hz, 1H), 6.64 (dd, J = 8.2, 2.0 Hz, 1H),
5.97 (d, J = 0.9 Hz, 1H), 5.93 (d, J = 0.9 Hz, 1H), 5.88 (s, 1H), 4.78 (dd, J = 9.4, 4.0 Hz, 1H), 3.81 –
3.54 (m, 8H), 3.32 (s, 4H), 3.11 - 3.08 (m, 5H), 2.89 (dd, J = 13.5, 9.4 Hz, 1H). 13C NMR (75 MHz,
DMSO) δ 148.6, 147.8, 147.2, 145.2, 128.0, 123.6, 122.7, 122.0, 113.5, 111.7, 108.3, 107.9,
101.2, 71.3, 55.6, 55.5, 53.9, 51.4, 50.4, 36.9, 23.3. HRMS (ESI) m/z calcd for: C21H26NO4 M+ =
356.1862, found: 365.1850.
6-(2,3-dimethoxybenzyl)-5,6,7,8-tetrahydro-[1,3]dioxolo[4,5-g]isoquinoline oxalate (7a).
1
Yield: 77%. White solid (m.p. = 218 – 220 oC). H NMR (300 MHz, DMSO-d6) δ 7.18 – 6.97 (m,
3H), 6.72 (d, J = 4.7 Hz, 2H), 5.96 (s, 2H), 4.10 (s, 2H), 3.93 (s, 2H), 3.83 (s, 3H), 3.77 (s, 3H), 3.13
(t, J = 6.1 Hz, 2H), 2.87 (t, J = 6.1 Hz, 2H). 13C NMR (75 MHz, DMSO) δ 163.4, 152.4, 147.7, 146.4,
145.7, 126.3, 125.5, 124.0, 123.5, 122.9, 113.4, 108.1, 106.4, 100.8, 60.5, 55.7, 53.5, 53.1, 49.1,
26.2. Rf = 0.4 (EtOAc). HRMS (ESI) m/z calcd for: C19H21NO4 [M+H]+ = 328.1549, found:
328.1548.
2-(2,3-dimethoxybenzyl)-1,2,3,4-tetrahydroisoquinolineoxalate (7b). Yield: 67%. White solid
(m.p. = 167 -169 oC) 1H NMR (300 MHz, DMSO-d6) δ 9.88 (s, 2H), 7.27 – 7.03 (m, 7H), 4.12 (s,
2H), 4.05 (s, 2H), 3.83 (s, 3H), 3.78 (s, 3H), 3.17 (t, J = 6.1 Hz, 2H), 2.98 (t, J = 6.1 Hz, 2H). 13C
NMR (75 MHz, DMSO) δ 163.4, 152.4, 147.7, 132.4, 130.9, 128.4, 126.9, 126.5, 126.4, 126.1,
124.0, 122.9, 113.4, 60.5, 55.7, 53.7, 53.2, 49.1, 26.3. Rf = 0.5 (EtOAc). HRMS (ESI) m/z calcd
for: C18H22NO2 [M+H]+ = 284.1651, found: 284.1640.

6-benzyl-5,6,7,8-tetrahydro-[1,3]dioxolo[4,5-g]isoquinoline hydrochloride (7c). Yield: 91%.
White solid (m.p. = 235 -238 oC) 1H NMR (300 MHz, DMSO-d6) δ 11.66 (s, 1H), 7.77 – 7.55 (m,
2H), 7.47 (dp, J = 4.8, 1.9 Hz, 3H), 6.77 (d, J = 7.1 Hz, 2H), 5.98 (s, 2H), 4.40 (d, J = 6.3 Hz, 2H),
4.13 (t, J = 4.4 Hz, 2H), 3.55 (s, 1H), 3.20 (d, J = 10.8 Hz, 2H), 2.92 (q, J = 13.3, 11.4 Hz, 1H). 13C
NMR (75 MHz, DMSO) δ 146.8, 146.0, 131.3, 129.8, 129.4, 128.7, 124.8, 120.9, 108.1, 106.3,
101.0, 57.9, 51.0, 48.1, 24.7. Rf = 0.4 (EtOAc). HRMS (ESI) m/z calcd for: C17H18NO2 [M+H]+ =
268.1338, found: 268.1324.
6-(2,6-dimethoxybenzyl)-5,6,7,8-tetrahydro-[1,3]dioxolo[4,5-g]isoquinoline oxalate (7d).
Yield: 76%. White solid (m.p. = 218 -220 oC) 1H NMR (300 MHz, DMSO-d6) δ 7.41 (t, J = 8.4 Hz,
1H), 6.79 – 6.71 (m, 4H), 5.98 (s, 2H), 4.17 (s, 2H), 4.07 (s, 2H), 3.82 (s, 6H), 3.28 (d, J = 6.5 Hz,
2H), 2.92 (t, J = 6.3 Hz, 2H). 13C NMR (75 MHz, DMSO) δ 164.1, 159.2, 158.6, 133.1, 131.6,
128.0, 121.6, 113.1, 113.0, 107.0, 104.2, 56.0, 55.2, 52.2, 49.0, 47.4, 25.5. Rf = 0.4 (EtOAc).
HRMS (ESI) m/z calcd for: C19H22NO4 [M+H]+ = 328.1546, found: 328.1544.
6-(3,5-dimethoxybenzyl)-5,6,7,8-tetrahydro-[1,3]dioxolo[4,5-g]isoquinoline oxalate (7e).
Yield: 81%. White solid (m.p. = 219 -221oC) 1H NMR (300 MHz, DMSO-d6) δ 6.74 (d, J = 4.5 Hz,
2H), 6.67 (d, J = 2.3 Hz, 2H), 6.52 (t, J = 2.2 Hz, 1H), 5.97 (s, 2H), 4.07 (s, 2H), 3.93 (s, 2H), 3.75 (s,
6H), 3.14 (t, J = 6.2 Hz, 2H), 2.89 (t, J = 6.1 Hz, 2H). 13C NMR (75 MHz, DMSO) δ 164.1, 161.0,
146.9, 146.3, 126.0, 123.7, 108.6, 108.4, 107.0, 101.3, 100.8, 59.8, 55.7, 53.3, 49.4, 26.5. Rf =
0.3 (EtOAc). HRMS (ESI) m/z calcd for: C19H22NO4 [M+H]+ = 328.1546, found: 328.1545.
6-(4-methoxybenzyl)-5,6,7,8-tetrahydro-[1,3]dioxolo[4,5-g]isoquinoline hydrochloride (7f).
Yield: 94%. White solid (m.p. = 231 -233oC) 1H NMR (300 MHz, DMSO-d6) δ 11.48 (s, 1H), 7.63 –
7.55 (m, 2H), 7.07 – 6.95 (m, 2H), 6.77 (d, J = 5.8 Hz, 2H), 5.98 (s, 2H), 4.43 – 4.24 (m, 2H), 4.10
(d, J = 5.4 Hz, 2H), 3.79 (s, 3H), 3.54 (d, J = 4.3 Hz, 1H), 3.23 – 3.12 (m, 2H), 2.97 – 2.82 (m, 1H).
13C NMR (75 MHz, DMSO) δ 160.0, 146.8, 146.0, 132.8, 124.8, 121.5, 121.0, 114.1, 108.1, 106.3,
101.0, 57.5, 55.2, 50.8, 47.9, 24.7. Rf = 0.4 (EtOAc). HRMS (ESI) m/z calcd for: C18H20NO3
[M+H]+ = 298.1443, found: 298.1435.

6-(4-(trifluoromethyl)benzyl)-5,6,7,8-tetrahydro-[1,3]dioxolo[4,5-g]isoquinoline oxalate (7g).
Yield: 67%. White solid (m.p. = 219 -222 oC) 1H NMR (300 MHz, DMSO-d6) δ 7.83 – 7.64 (m, 4H),
6.72 (s, 1H), 6.68 (s, 1H), 5.95 (s, 2H), 4.12 (s, 2H), 3.82 (s, 2H), 3.04 (t, J = 6.0 Hz, 2H), 2.85 (t, J =
6.1 Hz, 2H). 13C NMR (75 MHz, DMSO) δ 163.0, 146.2, 145.6, 139.1, 130.6, 128.9, 128.5, 125.8,
125.4, 125.3, 124.2, 108.1, 106.3, 100.7, 59.1, 53.4, 49.4, 26.7. Rf = 0.4 (EtOAc). HRMS (ESI)
m/z calcd for: C18H17FNO2 [M+H]+ = 336.1200, found: 336.1203.
6-(2,4-dimethoxybenzyl)-5,6,7,8-tetrahydro-[1,3]dioxolo[4,5-g]isoquinoline oxalate (7h). Yield: 70%.
White solid (m.p. = 187 -189 oC) 1H NMR (300 MHz, DMSO-d6) δ 7.36 (d, J = 8.4 Hz, 1H), 6.75 (d, J =
4.8 Hz, 2H), 6.64 (d, J = 2.4 Hz, 1H), 6.59 (dd, J = 8.4, 2.4 Hz, 1H), 5.98 (s, 2H), 4.10 (s, 2H), 4.00
(s, 2H), 3.82 (s, 3H), 3.80 (s, 3H), 3.20 (t, J = 6.3 Hz, 2H), 2.90 (t, J = 6.3 Hz, 2H). 13C NMR (75
MHz, DMSO) δ 163.1, 161.4, 159.2, 146.5, 145.9, 132.7, 125.6, 123.4, 112.8, 108.1, 106.4,
105.4, 100.8, 98.7, 55.7, 55.4, 53.4, 52.8, 48.8, 26.1. Rf = 0.3 (EtOAc). HRMS (ESI) m/z calcd for:
C19H22NO4 [M+H]+ = 328.1540, found: 328.1542.
6-(2-methoxybenzyl)-5,6,7,8-tetrahydro-[1,3]dioxolo[4,5-g]isoquinoline oxalate (7i). Yield:
78%. White solid (m.p. = 208 - 210 oC1H NMR (300 MHz, DMSO-d6) δ 7.54 – 7.32 (m, 2H), 7.20 –
7.07 (m, 1H), 7.00 (t, J = 7.3 Hz, 1H), 6.74 (d, J = 7.0 Hz, 2H), 5.97 (s, 1H), 4.12 (s, 1H), 3.97 (s,
1H), 3.83 (s, 1H), 3.17 (s, 1H), 2.90 (d, J = 6.0 Hz, 1H). 13C NMR (75 MHz, DMSO) δ 163.9, 157.9,
146.5, 145.8, 132.1, 130.6, 125.2, 122.8, 120.4, 111.2, 108.1, 106.4, 100.9, 55.5, 53.1, 52.5,
48.8, 25.6. Rf = 0.5 (EtOAc). HRMS (ESI) m/z calcd for: C18H20NO3 [M+H]+ = 298.1443, found:
298.1436.
6-(3-methoxybenzyl)-5,6,7,8-tetrahydro-[1,3]dioxolo[4,5-g]isoquinoline oxalate (7j). Yield:
1
83%. White solid (m.p. = 235 -238 oC). H NMR (300 MHz, DMSO-d6) δ 7.33 (t, J = 7.8 Hz, 1H),
7.09 – 6.89 (m, 3H), 6.72 (d, J = 7.9 Hz, 2H), 5.95 (s, 2H), 4.04 (d, J = 10.6 Hz, 2H), 3.91 – 3.81 (m,
2H), 3.76 (s, 3H), 3.08 (d, J = 8.7 Hz, 2H), 2.87 (d, J = 6.5 Hz, 2H). 13C NMR (75 MHz, DMSO) δ
162.9, 159.7, 146.5, 145.9, 135.7, 129.7, 126.1, 124.4, 122.2, 115.5, 114.2, 108.2, 106.5, 100.8,

60.0, 55.3, 53.6, 49.7, 26.7. Rf = 0.4 (EtOAc). HRMS (ESI) m/z calcd for: C18H20NO3 [M+H]+ =
298.1443, found: 298.1440.
2-(2,6-dimethoxybenzyl)-1,2,3,4-tetrahydroisoquinoline oxalate (7k). Yield: 96%. White solid
1
(m.p. = 193 -195 oC). H NMR (300 MHz, DMSO-d6) δ 7.41 (t, J = 8.4 Hz, 1H), 7.31 – 7.11 (m, 4H),
6.75 (d, J = 8.4 Hz, 2H), 4.19 (s, 4H), 3.82 (s, 6H), 3.39 – 3.28 (m, 2H), 3.03 (t, J = 6.4 Hz, 2H). 13C
NMR (75 MHz, DMSO) δ 163.3, 159.2, 132.1, 131.1, 130.3, 128.3, 127.1, 126.6, 126.1, 108.1,
104.3, 56.0, 52.9, 49.2, 47.7, 39.8, 25.6. Rf = 0.5 (EtOAc). HRMS (ESI) m/z calcd for : C18H22NO2
[M+H]+ = 284.1651, found: 284.1642.
6-(2,6-dimethylbenzyl)-5,6,7,8-tetrahydro-[1,3]dioxolo[4,5-g]isoquinoline oxalate (7l). Yield:
63%. White solid (m.p. = 196 -198 oC1H NMR (300 MHz, DMSO-d6) δ 7.33 – 6.93 (m, 3H), 6.68
(s, 2H), 5.93 (s, 2H), 3.86 (s, 2H), 3.81 – 3.63 (m, 2H), 2.92 (s, 2H), 2.75 (t, J = 6.0 Hz, 2H), 2.37 (s,
6H). 13C NMR (75 MHz, DMSO) δ 161.9, 146.1, 145.7, 138.2, 132.8, 128.3, 127.6, 126.6, 126.2,
108.2, 106.4, 100.6, 54.7, 54.6, 49.8, 27.9, 19.9. Rf = 0.4 (EtOAc). HRMS (ESI) m/z calcd for:
C19H22NO2 [M+H]+ = 296.1651, found: 296.1645.
6-(2,6-dichlorobenzyl)-5,6,7,8-tetrahydro-[1,3]dioxolo[4,5-g]isoquinoline oxalate (7m). Yield:
67%. White solid (m.p. = 153 - 157 oC) 1H NMR (300 MHz, DMSO-d6) δ 7.55 – 7.46 (m, 2H), 7.38
(dd, J = 8.9, 7.1 Hz, 1H), 6.63 (d, J = 4.2 Hz, 2H), 5.92 (s, 2H), 3.94 (s, 2H), 3.68 (s, 2H), 2.85 (t, J =
5.9 Hz, 2H), 2.70 (t, J = 5.9 Hz, 2H). 13C NMR (75 MHz, DMSO) δ 161.3, 145.8, 145.4, 136.2,
130.3, 128.7, 126.4, 108.1, 106.3, 100.5, 55.3, 54.9, 50.1. Rf = 0.4 (EtOAc). HRMS (ESI) m/z
calcd for: C17H16NO2Cl2 [M+H]+ = 336.0559, found: 336.0558.
2-(2,6-dimethoxybenzyl)-7-methoxy-1,2,3,4-tetrahydroisoquinoline oxalate (7n). Yield: 69%.
White solid (m.p. = 180 -181 oC1H NMR (300 MHz, DMSO-d6) δ 7.41 (t, J = 8.4 Hz, 1H), 7.12 (d, J
= 8.4 Hz, 1H), 6.84 (dd, J = 8.4, 2.7 Hz, 1H), 6.80 – 6.68 (m, 3H), 4.17 (d, J = 15.0 Hz, 4H), 3.82 (s,
6H), 3.73 (s, 3H), 3.33 (s, 2H), 2.95 (t, J = 6.4 Hz, 2H). 13C NMR (75 MHz, DMSO) δ 164.1, 159.1,

157.7, 131.5, 130.6, 129.5, 123.5, 114.0, 111.2, 106.7, 104.0, 55.9, 55.2, 52.4, 49.2, 47.2, 24.3.
Rf = 0.5 (EtOAc). HRMS (ESI) m/z calcd for: C19H24NO3 [M+H]+ = 314.1756, found: 314.1764.
2-(2,6-dimethoxybenzyl)-6-methoxy-1,2,3,4-tetrahydroisoquinoline oxalate (7o). Yield: 55%.
White solid (m.p. = 186 – 188 oC) 1H NMR (300 MHz, DMSO-d6) δ 7.41 (t, J = 8.4 Hz, 1H), 7.10
(d, J = 8.4 Hz, 1H), 6.87 – 6.57 (m, 4H), 4.19 (s, 2H), 4.13 (s, 2H), 3.82 (s, 6H), 3.74 (s, 3H), 3.44 –
3.22 (m, 2H), 3.02 (t, J = 6.4 Hz, 2H). 13C NMR (75 MHz, DMSO) δ 164.1, 159.2, 158.6, 133.1,
131.6, 128.0, 121.6, 113.1, 113.0, 107.0, 104.2, 56.0, 55.2, 52.2, 49.0, 47.4, 25.5. Rf = 0.5
(EtOAc). HRMS (ESI) m/z calcd for: C19H24NO3 [M+H]+ = 314.1756, found: 314.1751.
(7,8-dihydro-[1,3]dioxolo[4,5-g]isoquinolin-6(5H)-yl)(2,6-dimethoxyphenyl)methanone (10).
To a solution of 2,6-dimethoxybenzoic acid (150 mg, 0.824 mmol) in DMF (3 mL) was added
sequentially trimethylamine (2.74 mmol) and PyBOP (514 mg, 1.00 mmol). The solution was
stirred at ambient temperature for 10 minutes. Next, 5,6,7,8-tetrahydro-[1,3]dioxolo[4,5-
g]isoquinoline (146 mg, 0.824 mmol) was added and the reaction was allowed to proceed for a
further 3 hours, at which time the solution was poured into water (50 mL) and extracted three
times with Et2O (50 mL). The pooled organic extracts were washed with H2O, dried over
Na2SO4, filtered and concentrated. The crude residue was then purified by flash
chromatography (EtOAc) to yield the titled compound (221 mg, 79%). White solid (m.p. = 161 –
164 oC). NMR analysis showed the presence of two rotamers in a ratio of 2:1, which was
confirmed by temperature-resolved NMR experiments (see Supporting Information). 1H NMR
Major rotamer: (300 MHz, DMSO-d6) δ 7.34 (t, J = 8.4 Hz, 1H), 6.82 (s, 1H), 6.78 – 6.64 (m, 3H),
1
5.96 (s, 2H), 4.65 (s, 2H), 3.73 (s, 6H), 3.30 (dd, J = 6.3, 5.4 Hz, 2H), 2.61 (t, J = 5.8 Hz, 2H). H
NMR Minor rotamer: (300 MHz, DMSO-d6) δ 7.33 (t, J = 8.4 Hz, 1H), 6.75 - 6.69 (m, 3H), 6.58 (s,
1H), 5.93 (s, 2H), 4.17 (s, 2H), 3.75 (t, J = 5.8 Hz, 2H), 3.63 (s, 6H), 2.73 (t, J = 5.8 Hz, 2H). 13C
NMR Combined rotamers (101 MHz, DMSO) δ 164.7, 156.7, 156.6, 146.3, 146.2, 146.1, 146.1,
130.8, 130.7, 128.1, 127.8, 126.9, 126.5, 115.0, 114.8, 108.9, 108.8, 107.0, 106.4, 104.8, 104.7,
101.1, 101.0, 56.2, 47.8, 43.9, 43.7, 29.3, 28.4. HRMS (ESI) m/z calcd for: C19H19NO5 [M+H]+ =
342.1341, found: 342.1335.

5.2 Pharmacology
5.2.1 Cell culture conditions
The BV2 microglial cell line a generous gift from Dr. Yen (Fort Wayne, IN). Cells were grown in
DMEM:F-12 media (1:1) supplemented with 10% heat-inactivated FBS and 100 U/mL penicillin
and 100 μg/mL streptomycin in 150 cm2 culture flasks in a humidified atmosphere of 5% CO2.
The media was replaced every 2 days, and cells were subcultured once a confluence of 70-80%
was reached. All test compounds were dissolved in DMSO and diluted in media (final DMSO
concentration of 0.1% v/v).
Primary MG were generated as previously described.25 Briefly, cerebral cortical cells from 1-2
days old neonatal mice were plated in 75-cm2 culture flasks DMEM/F12 (1:1) supplemented with
10% heat-inactivated FBS, containing 2mM glutamine and 1X antibiotic/antimycotic (complete
medium). After days 3 and 6 of plating, the medium was removed and replaced with fresh
complete medium containing 5 ng/ml GM-CSF. MG were then harvested at day 13 or 14 by
shaking the flasks at 350 rpm for 30 min at 37°C. Collected cells were subsequently plated for
use.
5.2.2 Cytokine quantification assay
BV2 cells were plated in 12 well plates at a density of 5 x 105 cells/mL and allowed to attach
overnight. Cells were then treated with either DMSO or test compounds for one hour, followed
by LPS (E. coli O55:B5, 100 ng/mL, Sigma) treatment for 24 hours. Media was then collected,
centrifuged to remove floating cells, and either assayed immediately or frozen at -80 oC for
future use. TNF or IL-6 levels were determined using Ready Set Go! ELISA kit from
Thermofisher.
5.2.3 Nitrate quantification assay

BV2 cells were plated in 12 well plates at a density of 5 x 105 cells/mL and allowed to attach
overnight. Cells were then treated with either DMSO or test compounds for one hour, followed
by LPS treatment for 24 hours. Media was then collected, centrifuged to remove floating cells,
and either assayed immediately or frozen at -80 oC for future use. Media samples were assayed
for nitrite levels using the Greis assay kit from Thermofisher.
5.2.4 QPCR analysis
After collection primary MG were plated (5 x 105 cells/mL) and allowed to attach overnight.
Cells were then treated with either DMSO or test compound for one hour, followed by LPS (100
ng/mL) for 2 and 4 hours. Following treatment, primary MG were subjected to RNA extraction,
purification, and quantification using the Trizol reagent (Thermofisher), followed by cDNA
synthesis using the High Capacity Reverse Transcription cDNA kit (Thermofisher). Expression of
Il-1β, Il-12p35, Il-12p40, Il-23p19, GM-CSF, TNF, and Il-6 were detected by quantitative PCR as
previous described.26 The primers used were Il-1β: sense 5’-CCCTGCAGCTGGAGAGTGTGGA-3’
and antisense 5’-TGTGCTCTGCTTGGAGGTGCTG-3’; Il-12p35: sense 5’-
CTGTGCCTTGGTAGCATCTATG -3’ and anti-sense 5’-GCAGAGTCTCGCCATTATGATTC-3’; Il-12p40:
sense 5’-TGGTTTGCCATCGTTTTGCTG-3’ and antisense 5’-ACAGGTGAGGTTCACTGTTTCT-3’; Il-
23p19: sense 5’-TGCTG GATTGCAGAGCAGTAA-3’ and anti-sense 5’-
GCATGCAGAGATTCCGAGAGA-3’; GM-CSF: sense 5’-ATGCCTGTCACGTTGAATGAAG-3’ and anti-
sense 5’-GCGGGTCT GCACACATGTTA-3’; TNF: sense 5’-ATGGCCTCCCT-CTCATCA-GT-3’ and
antisense 5’-CTTGGTGGTTTGCTACGACG-3’; IL-6: sense 5’-TCCTCTCTGCAAGAGACTTCCATCC-3’
and antisense 5’-GGGAAGG-CCGTGGTTGTCACC-3’.
6. Statistical analysis
Results are expressed as mean ± SEM. Comparisons among multiple groups were done by one-
way ANOVA, followed by Dunnett’s post hoc test. Statistical significance was determined as p
values ≤ 0.05. GraphPad Prism 5 was used to perform all analyses.
Funding

This work was funded by the Manchester University Pharmacy Programs Internal Research Grant
to DB and by the Anna Yoder Multiple Sclerosis Fund to JY.
Declarations of interest
The authors declare there are no conflicts of interest.
Abbreviations.
BB: berberine
CNS: central nervous system
EAE: experimental autoimmune encephalomyelitis
LPS: lipopolysaccharide
MS: multiple sclerosis
Keywords. Berberine, neuroinflammation, multiple sclerosis
References
1. Compston A, Coles A. The Lancet.372:1502-1517.
2. Brück W, Porada P, Poser S, et al. Ann. Neurol. 1995;38:788-796.
3. Symptoms and Diagnosis. http://www.nationalmssociety.org/Symptoms-Diagnosis. (2017).
4. Imenshahidi M, Hosseinzadeh H. Phytother. Res. 2016;30:1745-1764.
5. Luo J, Chen R, Zeng S, et al. Biochem. Biophys. Res. Commun. 2017;490:927-932.
6. Jiang Y, Wu A, Zhu C, et al. Neurol. Res. 2013;35:360-368.
7. Li Z, Geng Y-N, Jiang J-D, et al. J. Evid. Based. Complementary Altern. Med. 2014;2014:12.
8. Battu SK, Repka MA, Maddineni S, et al. AAPS PharmSciTech. 2010;11:1466-1475.
9. Godugu C, Patel AR, Doddapaneni R, et al. PLoS One. 2014;9:e89919.
10. Wang K, Feng X, Chai L, et al. Drug. Metab. Rev. 2017;49:139-157.
11. Li Y-H, Yang P, Kong W-J, et al. J. Med. Chem. 2009;52:492-501.
12. Jin X, Yan T-H, Yan L, et al. Drug Des. Devel. Ther. 2014;8:1047-1059.
13. Zhang S, Wang X, Yin W, et al. Bioorg. Med. Chem. Lett. 2016;26:4799-4803.
14. Nechepurenko I, Boyarskikh U, Khvostov M, et al. Chem. Nat. Compd. 2015;51:916-922.
15. Huang WJ, Singh OV, Chen CH, et al. Helv. Chim. Acta. 2004;87:167-174.
16. Stenlake John BStenlake JB. Eur. J. Med. Chem. 1981;16:515-524.
17. Kollias G, Douni E, Kassiotis G, et al. Immunol. Rev. 1999;169:175-194.
18. Selmaj K, Raine CS, Cannella B, et al. J. Clin. Invest. 1991;87:949-954.
19. Selmaj KWRaine CS. Ann. Neurol. 1988;23:339-346.
20. Selmaj K, Raine CS, Farooq M, et al. J. Immunol. 1991;147:1522-1529.
21. Krei K, Fredrikson S, Fontana A, et al. J. Neuroimmunol. 1991;31:147-153.
22. Maimone D, Guazzi GC, Annunziata P. J. Neurol. Sci. 1997;146:59-65.
23. Smith KJ, Lassmann H. Lancet Neurol. 2002;1:232-241.
24. Fletcher JM, Lalor SJ, Sweeney CM, et al. Clin. Exp. Immunol. 2010;162:1-11.
25. Kuo P-C, Yu IC, Scofield BA, et al. Brain, Behav., Immun. 2017;62:180-192.
26. Kuo PC, Scofield BA, Yu IC, et al. J. Am. Heart Assoc. 2016;5:e002610.