Doxycycline is Neuroprotective Against Nigral DopaminergicDegeneration by a Dual Mechanism Involving MMP-3

Yuri Cho Æ Hyo Jin Son Æ Eun-Mee Kim Æ Ji Hyun Choi ÆSung Tae Kim Æ In Jung Ji Æ Dong Hee Choi ÆTong H. Joh Æ Yoon Seong Kim Æ Onyou Hwang

Received: 20 October 2008 / Revised: 15 May 2009 / Accepted: 18 June 2009 / Published online: 7 July 2009

� Springer Science+Business Media, LLC 2009

Abstract In Parkinson disease (PD), the dopaminergic

(DAergic) neurons in the substantia nigra undergo degen-

eration. While the exact mechanism for the degeneration is

still not completely understood, neuronal apoptosis and

inflammation are thought to play roles. We have recently

obtained evidence that matrix metalloproteinase (MMP)-3

plays a crucial role in the apoptotic signal in DAergic cells

as well as activation of microglia. The present study tested

whether doxycycline might modulate MMP-3 and provide

neuroprotection of DAergic neurons. Doxycycline effec-

tively suppressed the expression of MMP-3 induced in

response to cellular stress in the DAergic CATH.a cells.

This was accompanied by protection of CATH.a cells as

well as primary cultured mesencephalic DAergic neurons

via attenuation of apoptosis. The active form of MMP-3,

released under the cell stress condition, was also decreased

in the presence of doxycycline. In addition, doxycycline

led to downregulation of MMP-3 in microglial BV-2 cells

exposed to lipopolysaccharide (LPS). This was accompa-

nied by suppression of production of nitric oxide and TNF-

a, as well as gene expression of iNOS, TNF-a, IL-1b, and

COX-2. In vivo, doxycycline provided neuroprotection of

the nigral DAergic neurons following MPTP treatment, as

assessed by tyrosine hydroxylase immunocytochemistry

and silver staining, and suppressed microglial activation

and astrogliosis as assessed by Iba-1 and GFAP immuno-

chemistry, respectively. Taken together, doxycycline

showed neuroprotective effect on DAergic system both in

vitro and in vivo and this appeared to derive from anti-

apoptotic and anti-inflammatory mechanisms involving

downregulation of MMP-3.

Keywords Doxycycline � Matrix metalloproteinase-3 �Parkinson’s disease � Microglia � Dopaminergic neurons

Introduction

In Parkinson’s disease (PD), the dopaminergic (DAergic)

neurons in the substantia nigra (SN) undergo degeneration.

These DAergic neurons are particularly vulnerable due to

the presence of reactive oxygen species-generating mole-

cules, including dopamine (DA) itself and iron, as well as

low antioxidants. Various cellular stresses, therefore, can

cause activation of cascade of events leading to apoptotic

death of the DAergic neurons. In addition, SN contains a

relatively high number of microglia (Kim et al. 2000) and

neuroinflammation involving activated microglia plays an

active role in the pathogenesis of PD (Sanchez-Pernaute

et al. 2004; Walker and Lue 2005).

Matrix metalloproteinase (MMP)-3 is a member of the

MMP family, which has been widely recognized to be

involved in degradation and remodeling of the extracellular

matrix (Visse and Nagase 2003). More recently, evidence

that MMP-3 plays a critical role in DAergic neurodegen-

eration has been demonstrated. We have observed that

MMP-3 participates in the apoptotic signaling in DAergic

cells, and its pharmacological inhibition, suppression of

Yuri Cho and Hyo Jin Son contributed equally.

Y. Cho � H. J. Son � E.-M. Kim � J. H. Choi �S. T. Kim � I. J. Ji � D. H. Choi � O. Hwang (&)

Department of Biochemistry and Molecular Biology, University

of Ulsan College of Medicine, 388-1 Pungnap-dong, Songpa-ku,

Seoul 138-736, Korea

e-mail: oyhwang@amc.seoul.kr

T. H. Joh � Y. S. Kim

Department of Neurology and Neuroscience, Weill Medical

College of Cornell University, New York, NY 10021, USA

123

Neurotox Res (2009) 16:361–371

DOI 10.1007/s12640-009-9078-1

expression by siRNA, and gene deletion all lead to pro-

tection of the cells (Choi et al. 2008). In addition, MMP-3

is released from apoptotic DAergic cells (Kim et al. 2005b;

Choi et al. 2008), and this in turn can cause activation of

microglia and production of proinflammatory cytokines

and superoxide (Kim et al. 2005b; Kim et al. 2007).

Furthermore, MMP-3 is produced inside the activated

microglia as well, and mediates the generation of proin-

flammatory molecules (Mun-Bryce et al. 2002; Nut-

tall et al. 2007; Woo et al. 2008). Taken together,

modulation of MMP-3 may be used to protect against

neurodegeneration via direct neuroprotection, downregu-

lation of its release, and/or suppression of neuroinflam-

matory responses.

Currently no therapy has been definitively shown to

provide neuroprotection for PD and efforts are being made

to find ways to delay or halter the DAergic neurodegen-

eration. Doxycycline, a semi-synthetic tetracycline deriv-

ative, has been reported to downregulate MMPs (Boyle

et al. 1998; Brown et al. 2004; Burggraf et al. 2007) via a

mechanism that is distinct from its antibiotic activity

(Golub et al. 1994; Sapadin and Fleischmajer 2006; Soory

2008). Therefore, it was possible that doxycycline might

suppress MMP-3 induction and thus provide neuroprotec-

tion via both anti-apoptotic and anti-inflammatory mecha-

nisms. Doxycycline has a high penetration rate into the

brain and cerebrospinal fluid (Andersson and Alestig

1976), a prerequisite for a drug for the central nervous

system.

The present study shows the evidence that doxycycline

indeed downregulates the cellular stress-induced MMP-3

expression and suppresses apoptosis in DAergic cells. It

also downregulates MMP-3 release from these cells and

induction of MMP-3 in microglia in vitro. In addition,

doxycycline provides protection of the nigral DAergic

neurons in a 1-methyl-4-phenyl-1,2,3,6-tetrahydropyrimi-

dine (MPTP)-induced mouse PD model.

Materials and Methods

Chemicals

Culture media, fetal bovine serum (FBS), L-glutamine,

trypsin/EDTA, and penicillin–streptomycin were pur-

chased from GibcoBRL (Gaithersburg, MD). Lipopoly-

saccharide (LPS), tetrahydrobiopterin (BH4), doxycycline,

and MPTP were purchased from Sigma Chemical (St.

Louis, MO). First strand cDNA synthesis kit for RT-PCR

was purchased from MBI Fermentas (Ontario, Canada).

Protein concentration of samples was determined by

Bradford assay (BioRad, Hempstead, UK). Antibodies used

were as follows: polyclonal anti-MMP-3 (R&D System,

Minneapolis, MN), anti-tyrosine hydroxylase (TH) (Protos,

New York, NY), anti-Iba-1 (Wako Chemicals, Osaka,

Japan) and anti-activated caspase-3 (Cell Signaling Tech-

nology, Beverly, MA) and monoclonal anti-TH (R&D

system), and anti-glial fibrillary acidic protein (GFAP; Cell

Signaling Technology, Beverly, MA) and anti-b-actin

(Santa Cruz Biotechnology, Santa Cruz, CA). Alexa Fluor

488 goat anti-rabbit IgG and Alexa Fluor 546 donkey anti-

mouse IgG were from Molecular Probes (Eugene, OR). In

situ cell death detection kit for TUNEL staining was from

Roche Diagnostics GmbH (Penzberg, Germany). Vecta-

stain ABC kit and biotinylated secondary antibodies were

from Vector Laboratories (Burlingame, CA). Ready-Set-

Go mouse TNF-a ELISA kit was from eBioscience (San

Diego, CA). Chemiluminescence detection system was

obtained from Pierce Chemical (Rockford, IL). All other

chemicals were reagent grade and were purchased from

Sigma Chemical or Merck (Rahway, NJ). [3H]DA was

from Amersham Biosciences (Piscataway, NJ) and Taq

polymerase was from Roche Applied Science (Indianapo-

lis, IN). Laminin, Trizol reagent, and superscript II reverse

transcriptase were obtained from Invitrogen (Carlsbad,

CA).

Animals

All procedures were performed in compliance with the

Laboratory Animal Manual of the University of Ulsan

Asan Institute for Life Sciences, which follow the guide-

lines set forth by the Society for Neuroscience. Male

C57Bl/6 mice and pregnant Sprague-Dawley rats were

obtained from Orient Charles River Technology (Seoul,

Korea).

Cell Cultures

The BV-2 mouse microglial cells, originally produced by

Dr. V. Bocchini (Blasi et al. 1990) were grown and

maintained in Dulbecco’s modified Eagle’s medium with

10% FBS, 100 IU/l penicillin, and 10 lg/ml streptomycin

at 37�C in a humidified incubator under 5% CO2. For

experiments, BV-2 cells were plated at a density of

2.5 9 105 cells/well in 24-well polystyrene plates or

2.6 9 106 cells/60 mm plate. CATH.a cells (Suri et al.

1993) were grown in RPMI 1640 containing 8% horse

serum, 4% FBS, and penicillin–streptomycin. They were

plated at a density of 1.5 9 105 cells/well in 24-well

polystyrene plates or 1.5 9 106 cells/60 mm plate. After

24 h, the cells were fed with fresh medium and treated. For

primary culture of mesencephalic neurons, the ventral

mesencephalon was removed from 14-day gestation rat

embryo and incubated with 0.01% trypsin in HBSS for

15 min at 37�C. After trituration, 3 9 105 cells were plated

362 Neurotox Res (2009) 16:361–371

123

on each polystyrene cover slide that had been precoated

with 100 lg/ml poly-L-lysine and 4 lg/ml laminin and

placed in a 24-well culture plate. The cells were maintained

at 37�C in a humidified atmosphere with 5% CO2 in neu-

robasal medium supplemented with B-27, 2 mM gluta-

mine, 100 IU/l penicillin, and 10 lg/ml streptomycin. On

day 5 or 6 in vitro, the neurons were fed with fresh medium

and treated.

Cell Staining

For immunofluorescence staining, cells were fixed in cold

4% paraformaldehyde in 0.1 M phosphate buffered saline

(PBS), pH 7.4 for 30 min at room temperature. After

washing twice in PBS, the cells were incubated for 1 h in

blocking solution (5% FBS and 0.3% Triton X-100 in 0.1

M PBS), followed by overnight incubation with appropri-

ate primary antibody diluted in incubation solution (5%

FBS and 0.2% Triton X-100 in 0.1 M PBS) at 4�C. After

two washes in PBS, the samples were incubated at room

temperature for 1 h with appropriate fluorescence-labeled

secondary antibody in the incubation solution. For TH

immunostaining, mouse monoclonal anti-TH antibody

(1:5009) and Alexa Fluor 546 goat anti-mouse IgG

(1:2009) were used. For caspase-3, rabbit polyclonal anti-

caspase-3 antibody (1:1009) and Alexa Fluor 488 goat

anti-rabbit IgG (1:2009) were used. As a negative control,

the samples were incubated with the secondary antibody

only. The cells were washed in PBS, mounted on a glass

slide, and viewed by confocal microscopy (TCS-ST2;

Leica, Wetzlar, Germany). TUNEL staining was performed

according to the manufacturer’s protocol using fluorescein

and analyzed by confocal microscopy.

DA Uptake Assay

The primary cultured neurons were rinsed with Krebs–

Ringer solution and incubated in Krebs–Ringer solution

containing 1 mM ascorbate, 2 mM b-alanine, and 100 lM

pargyline at 37�C for 5 min in a humidified 5% CO2

atmosphere. The neurons were further incubated in the

presence of 25 nM [3H]DA (45 Ci/mmol) for 15 min.

Control neurons were treated in the same solution but

containing the DA uptake blocker bupropion (10 lM). The

cells were then washed with ice-cold Krebs–Ringer solu-

tion and lysed with 0.5 M NaOH. Radioactivity was

measured using a liquid scintillation counter (Beckman,

Fullerton, CA).

RT-PCR

Reverse transcription (RT) reaction was performed using

5 lg total RNA isolated from BV-2 or CATH.a cells.

Polymerase chain reaction (PCR) was performed, using the

cDNA product as a template, at 94�C for 1 min, 60�C for

1 min, and 72�C for 1 min for MMP-3 and at 94�C for

30 s, 60�C for 30 s, and 72�C for 30 s for the others. After

30 cycles of amplification, the PCR product was subjected

to agarose gel electrophoresis. RT-PCR against b2M was

performed as an internal control. Analysis of PCR products

on 1.5% agarose gels showed single-band amplification

products with respective expected sizes. The following

primers were used for PCR: MMP-3 (forward, CTGG

TACCAACCTATTCCTGGTTGC; reverse, TCTCTTTAT

TAGAAATGGCAGCATCG), iNOS (forward, ATGTCC

GAAGCAAACATCAC; reverse, TAATGTCCAGGAAG

TAGGTG), TNF-a (forward, CAGACCCTCACACTCA-

GATCATCTT; reverse, CAGAGCAATGACTCCAAAG-

TAGACCT), IL-1b (forward, ATGGCAACTGTTCCTG

AACTCAACT; reverse, CAGGACAGGTATAGATTCT

TTCCTTT), COX-2 (forward, CAGCAAATCCTTGCT

GTTCC; reverse, TGGGCAAAGAATGCAAACATC), and

b2M (forward, GCTATCCAGAAAACCCCTCAA; reverse,

CATGTCTCGATCCCAGTAGACGGT).

Western Blot

Equal amounts (20 lg) of protein were separated on 10%

SDS polyacrylamide gel and transferred onto polyvinyli-

dene difluoride-nitrocellulose filters membrane. Mem-

branes were blocked for 1 h in 20 mM Tris-HCl containing

137 mM NaCl and 0.05% Tween 20 (TBST) containing

5% skim milk at room temperature, incubated overnight

with primary antibody against MMP-3 (1:1000) or b-actin

(1:1000) at 4�C followed by horseradish peroxidase-con-

jugated secondary antibodies for 1 h at room temperature.

Protein bands were detected by chemiluminescence.

Lactate Dehydrogenase (LDH) Assay

Degrees of cell death were assessed by activity of LDH

released into the culture medium using the method previ-

ously described (Choi et al. 2004). Aliquots (50 ll) of cell

culture medium were incubated at room temperature in the

presence of 0.26 mM NADH, 2.87 mM sodium pyruvate,

and 100 mM potassium phosphate buffer (pH 7.4) in a total

volume of 200 ll. The rate of NAD? formation was

monitored for 5 min at 2-s intervals at 340 nm using a

microplate spectrophotometer (SPECTRA MAX 340 pc;

Molecular Devices, Menlo Park, CA).

Griess Reaction

Accumulated nitrite, an oxidative metabolite of NO, was

measured in the cell supernatant by the Griess reaction as

Neurotox Res (2009) 16:361–371 363

123

previously described (Seo et al. 2005). Briefly, 200 ll

aliquots of cell medium was mixed with 100 ll of Griess

reagent (1% sulfanilamide and 0.1% naphthylethylenedi-

amine dihydrochloride in 2.5% H3PO4) in a 96-well

microtiter plate, and the absorbance was read at 540 nm

using the plate reader. The nitrite concentration was

determined using a curve calibrated with sodium nitrite

standards.

Measurement of TNF-a Levels

TNF-a was measured by ELISA according to the protocol

provided by the manufacturer. The concentration of TNF-awas calculated according to a standard curve.

MPTP Mice

Male C57Bl/6 mice weighing 23–25 g were maintained in

a temperature- and humidity-controlled room with a 12-h

light–dark cycle and food and water available ad libitum.

The animals were divided into four groups: saline treated

control, MPTP treated, MPTP plus doxycycline treated,

and doxycycline treated groups (n = 10). Doxycycline was

administered intraperitoneally at 45 mg/kg twice at 12-h

interval. The first doxycycline administration was done

30 min before the first MPTP injection. MPTP was injected

intraperitoneally 4 times in a single day, 2 h apart at

20 mg/kg dissolved in saline. The animals were killed

7 days after the first injection; they were deeply anesthe-

tized (80 mg/kg ketamine and 20 mg/kg xylazine, injected

intraperitoneally) and transcardially fixed in 4% parafor-

maldehyde. Brains were promptly removed and postfixed

in 4% paraformaldehyde. After cryoprotection, the tissue

was cut into 20-lm sections on a vibrating-blade micro-

tome (VT 1000S; Leica, Nussloch, Germany), which were

stored in 0.08% sodium azide in PBS at 4�C until analysis.

Immunohistochemistry

Immunohistochemistry of the nigral and striatal brain tis-

sue sections was performed as described previously

(Hwang et al. 1998), using antisera against TH (1:1,0009),

Iba-1 (1:3009), or GFAP (1:1009), followed by Vecta-

stain ABC kit and biotinylated secondary antibody.

Quantitation of the TH-positive neurons was done by the

method described previously (Kim et al. 2006).

Amino-Cupric-Silver Staining

Amino-cupric-silver degeneration staining was carried out

by the method as described previously (de Olmos et al.

1994; Kim et al. 2005a). For counterstaining, the sections

previously stained and mounted on gelatin-coated slides

were immersed in 0.5% neutral red for 15 min, rinsed

twice in distilled H2O, dehydrated, cleared in pure xylene,

and coverslipped using DPX as a mounting medium. For

quantitation, we randomly selected a total of 2500 small

circles (10 lm in diameter) from microscopic fields of SN

and striatum (1009 magnification) sections. Their optical

density was measured using Science Lab 2001 Image

Gauge Version 4.0 software (Fuji Photo, Tokyo, Japan). To

account for differences in background staining intensity,

the background optical density was represented by the

average of the smallest 25 values in 2500 circles, the mean

of which was subtracted from the optical density of silver

stain-positive area to provide the final optical density

value.

Data Analyses

Data are expressed as mean ± SEM of independent

experiments. Comparisons of three of more groups were

analyzed by one-way ANOVA (analysis of variance) and

Student-Newman-Keuls comparison test. Statistical tests

were carried out using PRISM (GraphPad Software, San

Diego, CA). A value of P \ 0.05 was considered statisti-

cally significant.

Results

Doxycycline Downregulates MMP-3 Gene Expression

in DAergic Cells

We have previously established that exposure to BH4, an

endogenous molecule present in DAergic cells serving as a

cofactor for DA synthesis, causes oxidative stress (Choi

et al. 2000; Choi et al. 2003; Lee et al. 2007) and upreg-

ulation of MMP-3 (Choi et al. 2008) in cultured DAergic

CATH.a cells as well as primary cultured mesencephalic

DAergic neurons. Using this system, we first tested whe-

ther doxycycline might suppress the MMP-3 induction. As

shown by the RT-PCR data in Fig. 1a, the dramatic

upregulation of MMP-3 mRNA by the exposure to BH4

was suppressed by the co-treatment with doxycycline. At

1 lM doxycycline, the MMP-3 induction was completely

blocked. Accurate quantitation could not be done because

the bands corresponding to MMP-3 for untreated and

doxycycline-treated cells were barely detectable.

The suppression of MMP-3 induction by doxycycline

was also apparent at the protein level, as determined by

western blot analysis (Fig. 1b). We have previously

observed that under this cellular stress condition, most

MMP-3 is cleaved to the catalytically active 48 kD protein

(actMMP-3) (Choi et al. 2008). Whereas BH4 caused a

dramatic increase in actMMP-3 (6.86 ± 0.72 fold), this

364 Neurotox Res (2009) 16:361–371

123

was inhibited by doxycycline in a dose-dependent manner.

One micromolar doxycycline attenuated the increase in

actMMP-3 by 67%, and complete inhibition was achieved

at 5, 10, and 20 lM.

Doxycycline Protects Cultured DAergic Cells

Since MMP-3 participates in cell death (Choi et al. 2008)

and doxycycline suppressed MMP-3, it was possible that

doxycycline might provide protection. CATH.a cells trea-

ted with BH4 showed an increase in cell death (184 ± 10%

of untreated control; Fig. 2a), determined by activity of

LDH released into the medium. Co-treatment with doxy-

cycline at concentrations as low as 1 lM was able to

attenuate the cell death (126 ± 6% of untreated cells), and

almost complete protection was achieved at 5, 10, and

20 lM.

The protective effect of doxycycline on DAergic neu-

rons was also evident in primary cultured mesencephalic

neurons. Upon exposure to BH4, the ability of the DAergic

neurons to take up DA, used as an index of plasma mem-

brane integrity, was decreased to 60 ± 3% of untreated

control (Fig. 2b). Co-treatment with doxycycline (1 lM)

was able to completely block this effect. Therefore,

Fig. 1 Doxycycline suppresses the cell stress-induced MMP-3 pro-

duction in DAergic cells. CATH.a cells were exposed to 100 lM BH4

in the presence of various concentrations of doxycycline. The cells

were harvested after 6 h (a) or 24 h (b) and the levels of intracellular

MMP-3 mRNA (a) and protein (b) were assessed by RT-PCR and

western blot analysis, respectively. b2M and b-actin were used as

respective internal controls. For b, quantitation was made by

densitometry and the data are expressed as fold induction of untreated

control ± SEM; *** P \ 0.001 versus BH4-treated; and ??? P \0.001 versus untreated control

Fig. 2 Doxycycline protects DAergic cells from cell damage. aCATH.a cells were exposed to 100 lM BH4 for 24 h in the presence

of various concentrations of doxycycline and the degrees of cell death

were assessed by LDH activity in the culture medium. The data are

expressed as % of untreated control ± SEM; *** P \ 0.001 versus

BH4-treated; ??? P \ 0.001 versus untreated control; b Primary

cultured rat mesencephalic neurons were treated with 10 lM BH4 in

the presence or absence of 1 lM doxycycline for 24 h. The degree of

cell damage was assessed by DA uptake assay, expressed as percent

of untreated control ± SEM; *** P \ 0.001 versus BH4-treated; and??? P \ 0.001 versus untreated control

Neurotox Res (2009) 16:361–371 365

123

doxycycline exerted a direct protective effect on both

CATH.a cells and primary cultured mesencephalic DAer-

gic neurons.

Doxycycline Blocks Apoptotic Death of DA Cells

We tested whether doxycycline’s protective effect might be

achieved by interfering with apoptotic signaling. We have

previously shown that upon BH4 exposure, the TH-positive

DAergic neurons, among primary cultured mesencephalic

neurons, preferentially undergo apoptosis (Lee et al. 2007).

As shown in Fig. 3a, an increase in TUNEL staining was

indeed observed after exposure to BH4 in TH-immuno-

positive DAergic neurons, and this was completely blocked

by the co-treatment with doxycycline (1 lM). The increase

in the immunoreactivity of cleaved caspase-3, another

marker of apoptosis, was also blocked by doxycycline

(Fig. 3b).

Doxycycline Blocks the Release of actMMP-3

from Apoptotic DAergic Cells

Since actMMP-3 is released from apoptotic DAergic cells

(Kim et al. 2005b; Choi et al. 2008) and can exert further

effect extracellularly, we tested whether doxycycline might

be able to lower the amount of released actMMP-3. The

actMMP-3 level was indeed dramatically increased in the

culture medium of cells exposed to BH4 (12.6 ± 0.5 fold

increase in 24 h; P \ 0.001 vs. untreated control) (Fig. 4),

and this was prevented by the co-treatment with doxycy-

cline. Complete blockade of actMMP-3 release was

observed at 1 lM doxycycline.

Doxycycline Suppresses the Production of MMP-3

and Proinflammatory Molecules in Activated Microglia

We tested whether doxycycline might downregulate MMP-

3 expression in activated microglia as well. As shown in

Fig. 5a, in the mouse microglial cell line BV-2, exposure to

LPS led to a dramatic increase in MMP-3 mRNA level, and

this was suppressed by co-treatment with doxycycline. At

10 lM doxycycline, the LPS-induced MMP-3 increase was

completely abolished.

Whether the suppression of MMP-3 is accompanied by

downregulation of production of the proinflammatory

molecules was then tested. As shown in Fig. 5b, generation

of NO, elevated after LPS exposure (8.7 ± 0.7 fold of

untreated cells), was inhibited by co-treatment with 5 and

10 lM doxycycline (58 and 35% of LPS alone, respec-

tively). The level of TNF-a was also dramatically increased

after exposure to LPS (11.0 ± 0.9 fold of untreated cells),

and this was attenuated by doxycycline in a dose-depen-

dent manner (Fig. 5c). Ten micromolar doxycycline caused

lowering of the induction (52% of LPS alone. As shown in

Fig. 5a, the levels of RT-PCR products of iNOS and TNF-

a, greatly induced by the LPS treatment, were attenuated

by the co-treatment with 10 lM doxycycline (43 and 35%

Fig. 3 Doxycycline blocks apoptotic death of DAergic neurons.

Representative double immunostainings for a TH (red) and TUNEL

(green) and b TH (red) and active caspase-3 (green) of primary

cultured rat mesencephalic neurons exposed to 10 lM BH4 in the

presence or absence of doxycycline for 24 h. Scale bars = 8 and

20 lm for a and b, respectively. (Color figure online)

Fig. 4 Doxycycline prevents cell stress-induced actMMP-3 release

from DAergic cells. CATH.a cells were exposed to 100 lM BH4 in

the presence of various concentrations of doxycycline. After 24 h, the

levels of released actMMP-3 (48 kD) were determined by western

blot analysis of the culture medium

366 Neurotox Res (2009) 16:361–371

123

of LPS alone, respectively). The gene expression of

another proinflammatory cytokine IL-1b was also dramat-

ically upregulated by LPS (12.3 ± 1.0 fold of untreated

cells), and this was also blocked in the presence of doxy-

cycline to 35% of LPS treated cells. In addition, gene

expression of cyclooxygenenase-2 (COX-2), another

proinflammatory marker, was also upregulated by LPS to

8.8 ± 1.0 fold of untreated cells, and this was blocked in

the presence of doxycycline to 12%. Interestingly, doxy-

cycline alone had some downregulating effect on the basal

mRNA levels of IL-1b and COX-2.

Doxycycline Protects Nigral DAergic Neurons

in MPTP Mouse Model

We asked whether doxycycline might provide neuropro-

tection on the nigral DAergic neurons in vivo using the

well-established MPTP-injected animal PD model. As

shown in the representative photographs (Fig. 6a), while a

large decrease in the number of TH-immunopositive nigral

DAergic neurons were noted in the mice treated with

MPTP, co-treatment with doxycycline attenuated this

loss. Quantitative analysis showed that the numbers of

TH-positive neurons were 51 ± 3 and 81 ± 3% of vehicle-

treated control in the MPTP-treated and MPTP-doxycy-

cline-treated animals, respectively.

We also assessed the degree of degeneration in both SN

and striatum by amino-cupric-silver stain (de Olmos et al.

1994; Kim et al. 2005a). As shown in Fig. 6b, MPTP

caused increases in silver staining in both brain regions,

and this was attenuated by the doxycycline treatment.

Quantitative analyses revealed that in the SN, the optical

densities of MPTP-treated and MPTP-doxycycline-treated

animals were 147 ± 7 and 86 ± 4% of vehicle-treated

control, respectively. In the striatum, MPTP caused an

increase in silver staining optical density to 133 ± 3% of

untreated control and this was again suppressed by doxy-

cycline to 79 ± 2%. Taken together, doxycycline co-

treatment significantly prevented the nigrostriatal neurons

and fibers from degeneration.

Doxycycline Downregulates Nigral Microglia

Activation and Astrogliosis in MPTP Mouse Model

We also tested whether the doxycycline treatment might

affect microglial activation in vivo by immunohistochem-

istry against the microglial marker Iba-1. As shown in

Fig. 7a, compared to the vehicle-treated control animals

showing resting state morphology, microglia in MPTP-

treated animals have become activated, evidenced by

Fig. 5 Doxycycline prevents the induction of various inflammatory

genes in activated microglia. a The cells were exposed to 0.2 lg/ml LPS

for 6 h in the presence of 10 lM doxycycline and the relative amounts

of mRNA for MMP-3, iNOS, TNF-a, IL-1b, and COX-2 were measured

by RT-PCR, using b2M as an internal control. BV-2 cells were also

exposed to 0.2 lg/ml LPS in the presence of various concentrations of

doxycycline for 24 h and the amounts of NO (b) and TNF-a (c) were

assessed by the Griese method and ELISA, respectively. The data are

expressed as % of untreated control ± SEM; ??? P \ 0.001 versus

untreated control, *** P \0.001 versus LPS-alone

b

Neurotox Res (2009) 16:361–371 367

123

increased size and number as well as Iba-1 immunoreac-

tivity. On the other hand, the co-treatment of doxycycline

with MPTP resulted in a decrease in the degree of Iba-1

immunoreactivity almost to the level of control, suggestive

of downregulation of microglial activation. The effect of

doxycycline on astrocytes was also assessed by immuno-

histochemistry against the astrocyte marker GFAP

(Fig. 7b). A marked increase in GFAP-immunoreactivity,

indicative of astrogliosis, was observed in the SN of

MPTP-treated animals. In comparison, in the animals

co-treated with doxycycline, the GFAP immunoreactivity

was significantly reduced to a level comparable to the

vehicle-treated control.

Discussion

In the present study we demonstrate both in vitro and

in vivo that doxycycline, a widely prescribed antibiotic,

provides neuroprotection for the DAergic system. This

protective action seems to be mediated by multiple

mechanisms: (1) direct protection of DAergic cells by

lowering of the availability of intracellular MMP-3, which

participates in cell death signaling; (2) lowering of the

amount of released MMP-3, thereby precluding MMP-

3-induced activation of microglia; and (3) direct inhibition

of the production of MMP-3 and proinflammatory mole-

cules in LPS-activated microglia.

Because apoptosis involves a cascade of events, inter-

fering with the apoptotic signaling has been an important

issue in the field of neurodegeneration. Identification of

specific cellular targets participating in the apoptotic path-

way would be essential. Although caspases would be a good

target, the currently available caspase inhibitors are not

suitable because they must be injected directly into the brain

in large quantities (Charriaut-Marlangue 2004). We have

previously identified MMP-3 as a participant of apoptotic

pathway and therefore a potential cellular target (Choi et al.

2008). MMP-3 is dramatically induced in DAergic neurons

in response to cellular stress, such as oxidative stress and ER

stress, and participates in the apoptotic signaling upstream

of caspase-3 and downstream of JNK; downregulation of

MMP-3 by various approaches can provide protection in an

in vitro model (Choi et al. 2008). Although compounds that

inhibit the catalytic activity of MMP-3 including N-isobu-

tyl-N-(4-methoxyphenylsulfonyl)-glycylhydroxamic acid

Fig. 6 Doxycycline protects

nigral DAergic neurons and

striatal fibers from MPTP. Mice

were treated with vehicle,

MPTP, or MPTP ? doxycycline

as described in ‘‘Materials and

Methods’’ and the nigral (SN)

and striatal (STR) sections were

subjected to TH

immunocytochemistry (a) and

silver staining (b). Scale

bars = 200 lm (a) and 40 lm

(b). Densitometric analyses

were performed and the data are

expressed as % of vehicle-

treated control ± SEM.?? P \ 0.01 vs untreated

control; ** P \ 0.01 vs MPTP-

treated

368 Neurotox Res (2009) 16:361–371

123

(NNGH) are available and would be effective in vitro, their

high cost and low penetration into the brain make them

unsuitable as a neuroprotective drug. In addition, adminis-

tration of MMP-3-specific siRNA would involve a rather

invasive method. We report in the present study that the drug

doxycycline, which is known to enter the brain efficiently

(Andersson and Alestig 1976), can downregulate MMP-3

and provide DAergic neuroprotection.

We show the evidence that doxycycline provides direct

protection on DAergic neurons and that this is via inhib-

iting apoptosis, as demonstrated by the decreases in

TUNEL staining and active caspase-3. Doxycycline’s

protective effect has been observed in stroke models pre-

viously, and this has been thought to occur mainly by

inhibiting microglial activation. Lai and Todd (2006)

reported, in hypoxia-activated microglia model, that

doxycycline confers neuroprotection by regulating mi-

croglial, not neuronal, factors. We show for the first time

using isolated and glia-free cell culture systems that

doxycycline has a direct protective effect on DAergic cells

as well.

That actMMP-3 is released from apoptotic DAergic

cells (Kim et al. 2005b; Choi et al. 2008) and this in turn

causes microglial activation (Kim et al. 2005b) has been

shown. We observe in the present study that the release of

actMMP-3 is decreased in the presence of doxycycline.

The decrease in the amount of released actMMP may be

due to the lowered intracellular availability. However, one

cannot completely eliminate the possibility that a mecha-

nism for actMMP-3 release that is separate from but

coordinated with the intracellular apoptotic signaling may

exist. In fact, actMMP-3 release was observed to be com-

pletely abolished at 1 lM doxycycline (Fig. 4), but this

attenuated the intracellular actMMP-3 accumulation by

only 50% (Fig. 1b).

The proinflammatory and cytotoxic substances released

from activated microglia is thought to exert secondary

damage to nearby neurons that are otherwise healthy.

Because MMP-3 seems to play an important role in neur-

oinflammatory responses by eliciting the production of

TNF-a, IL-1b, and NO (Mun-Bryce et al. 2002; Kim et al.

2005b; Cauwe et al. 2007; Woo et al. 2008), the lowering

of its availability may preclude the secondary and perpetual

damage caused by neuroinflammation. The source of

MMP-3 activity can be activated microglia themselves as

well as the apoptotic DAergic neurons. MMP-3 expression

is increased in activated microglia (Rosenberg 2002; Woo

et al. 2008; the present data), and both actMMP-3 protein

and MMP-3 catalytic activity have been observed in the

medium of LPS-activated microglia (Woo et al. 2008). We

show in the present study that doxycycline suppresses the

LPS-induced increase in MMP-3 gene expression and that

this is accompanied by the downregulation of production of

NO and TNF-a, and the expression of iNOS, TNF-a, and

IL-1b.

Neuroprotective effect of doxycycline on DAergic

neurons was confirmed in vivo, in the MPTP-treated ani-

mal model of PD. The MPTP-induced degeneration of the

nigral DAergic neurons was attenuated by the co-treatment

with doxycycline. We have previously shown that MPP?,

the metabolite of MPTP, induced expression of MMP-3 in

primary cultured mesencephalic DAergic neurons and that

the nigral DAergic neurons of MMP-3 knockout animals

were relatively resistant to MPTP (Kim et al. 2007). Thus,

the protective effect of doxycycline shown in the MPTP-

treated animals is likely due to combination of both sup-

pression of apoptotic signal in the neurons and suppression

of microglial production of proinflammatory molecules,

both involving MMP-3 downregulation. Astrogliosis in the

MPTP-treated animals was also found to be suppressed by

the doxycycline co-treatment. It is possible that the soluble

factors that are released from activated microglia and

stimulate astrogliosis (Mohri et al. 2006; Rohl et al. 2007;

Zhang et al. 2009) are reduced under the condition that

microglia activation is suppressed by the doxycycline

administration.

Because doxycycline also suppresses other MMPs, it is

possible that inhibition of their activities also contributes to

the neuroprotective effect of doxycycline. However, we

have observed that in our BH4-induced CATH.a cell death

Fig. 7 Doxycycline suppresses microglial activation and astrogliosis

in SN of MPTP treated mice. Mice were treated with vehicle, MPTP,

or MPTP ? doxycycline as described in ‘‘Materials and Methods’’

and the nigral sections were subjected to Iba-1 (a) and GFAP (b)

immunohistochemistry. Scale bars = 50 lm

Neurotox Res (2009) 16:361–371 369

123

model (Choi et al. 2008) as well as ER stress model (Kim

et al. unpublished results), no changes were observed in the

mRNA levels of MMP-1, MMP-7, and MMP-9, belonging

to the collagenase, matrilysin, and gelatinase subtypes,

respectively, as well as MMP-2, which is usually present

ubiquitously. Therefore, at least in our cell system, the

levels of other MMP subtypes do not seem to be different

between healthy and apoptotic conditions and therefore

may not play an important role in the cell death

mechanism.

It is well known that another tetracycline derivation,

minocycline, also possesses the neuroprotective effect as

doxycycline (Stirling et al. 2005; Domercq and Matute

2004; Thomas et al. 2003). However, the drug is also

known to cause side effects such as vestibular toxicity

(Cunha 2006) and effects on bone and teeth formation.

During the on-going clinical trial for PD, the issue of

decreased tolerability of minocycline has recently been

raised to be a concern (NINDS NET-PD Investigators

2008). That is, due to dizziness, nausea, headache, joint

pain, and tooth discoloration, 23% of subjects prematurely

discontinued the minocycline treatment during the study

period. On the other hand, doxycycline has the least toxic

side effects of the tetracyclines (Toth et al. 1988; Smith and

Leyden 2005). Although doxycycline and minocycline

have similar chemical structures, they have been reported

to exhibit differences in patterns of neuroprotection in

some models of cerebral ischemia (Yrjanheikki et al. 1998;

Fox et al. 2005; Jantzie et al. 2005) and in antioxidant

activity (Kraus et al. 2005). In addition, doxycycline is a

more potent inhibitor of MMPs (Gilbertson-Beadling et al.

1995; Yao et al. 2007).

The effective concentration of doxycycline at which the

protective effect is observed in our hands is in the range of

concentration of plasma doxycycline in human [(3.75–

19.62 lM (Prall et al. 2002); 1.83–13.18 lM (Smith et al.

2008)]. Therefore, doxycycline, which is already available

in clinical situations and has protective action with minor

side effects, good tolerance, and excellent blood brain

barrier penetration, may be considered as a candidate

toward the therapeutic strategy for PD.

Acknowledgment This work was supported by Brain Research

Center of the 21st Century Frontier Research Program of the Ministry

of Science & Technology (2009K0012510) to O. Hwang.

References

Andersson H, Alestig K (1976) The penetration of doxycycline into

CSF. Scand J Infect Dis Suppl 9:17–19

Blasi E, Barluzzi R, Bocchini V, Mazzolla R, Bistoni F (1990)

Immortalization of murine microglial cells by a v-raf/v-myc

carrying retrovirus. J Neuroimmunol 27:229–237

Boyle JR, McDermott E, Crowther M, Willis AD, Bell PR, Thompson

MM (1998) Doxycycline inhibits elastin degradation and reduces

metalloproteinase activity in a model of aneurysmal disease.

J Vasc Surg 27:354–361

Brown DL, Desai KK, Vakili BA, Nouneh C, Lee HM, Golub LM

(2004) Clinical and biochemical results of the metalloproteinase

inhibition with subantimicrobial doses of doxycycline to prevent

acute coronary syndromes (MIDAS) pilot trial. Arterioscler

Thromb Vasc Biol 24:733–738

Burggraf D, Trinkl A, Dichgans M, Hamann GF (2007) Doxycycline

inhibits MMPs via modulation of plasminogen activators in focal

cerebral ischemia. Neurobiol Dis 25:506–513

Cauwe B, Van den Steen PE, Opdenakker G (2007) The biochemical,

biological, and pathological kaleidoscope of cell surface sub-

strates processed by matrix metalloproteinases. Crit Rev Bio-

chem Mol Biol 42:113–185

Charriaut-Marlangue C (2004) Apoptosis: a target for neuroprotec-

tion. Therapie 59:185–190

Choi HJ, Jang YJ, Kim HJ, Hwang O (2000) Tetrahydrobiopterin is

released from and causes preferential death of catecholaminergic

cells by oxidative stress. Mol Pharmacol 58:633–640

Choi HJ, Kim SW, Lee SY, Hwang O (2003) Dopamine-dependent

cytotoxicity of tetrahydrobiopterin: a possible mechanism for

selective neurodegeneration in Parkinson’s disease. J Neurochem

86:143–152

Choi HJ, Lee SY, Cho Y, Hwang O (2004) JNK activation by

tetrahydrobiopterin: implication for Parkinson’s disease. J Neu-

rosci Res 75:715–721

Choi DH, Kim EM, Son HJ, Joh TH, Kim YS, Kim D, Flint Beal M,

Hwang O (2008) A novel intracellular role of matrix metallo-

proteinase-3 during apoptosis of dopaminergic cells. J Neuro-

chem 106:405–415

Cunha BA (2006) New uses for older antibiotics: nitrofurantoin,

amikacin, colistin, polymyxin B, doxycycline, and minocycline

revisited. Med Clin North Am 90:1089–1107

de Olmos JS, Beltramino CA, de Lorenzo S (1994) Use of an amino-

cupric-silver technique for the detection of early and semiacute

neuronal degeneration caused by neurotoxicants, hypoxia, and

physical trauma. Neurotoxicol Teratol 16:545–561

Domercq M, Matute C (2004) Neuroprotection by tetracyclines.

Trends Pharmacol Sci 25:609–612

Fox C, Dingman A, Derugin N, Wendland MF, Manabat C, Ji S,

Ferriero DM, Vexler ZS (2005) Minocycline confers early but

transient protection in the immature brain following focal

cerebral ischemia-reperfusion. J Cereb Blood Flow Metab 25:

1138–1149

Gilbertson-Beadling S, Powers EA, Stamp-Cole M, Scott PS, Wallace

TL, Copeland J, Petzold G, Mitchell M, Ledbetter S, Poorman R

(1995) The tetracycline analogs minocycline and doxycycline

inhibit angiogenesis in vitro by a non-metalloproteinase-

dependent mechanism. Cancer Chemother Pharmacol 36:

418–424

Golub LM, Evans RT, McNamara TF, Lee HM, Ramamurthy NS

(1994) Non-antimicrobial tetracycline inhibits gingival matrix

metalloproteinases in Porphyromonas gingivalis-induced peri-

odontitis in rats. Ann NY Acad Sci 732:96–111

Hwang O, Baker H, Gross S, Joh TH (1998) Localization of GTP

cyclohydrolase in monoaminergic but not nitric oxide-producing

cells. Synapse 28:140–153

Jantzie LL, Cheung PY, Todd KG (2005) Doxycycline reduces

cleaved caspase-3 and microglial activation in an animal model

of neonatal hypoxia-ischemia. J Cereb Blood Flow Metab 25:

314–324

Kim WG, Mohney RP, Wilson B, Jeohn GH, Liu B, Hong JS (2000)

Regional difference in susceptibility to lipopolysaccharide-

370 Neurotox Res (2009) 16:361–371

123

induced neurotoxicity in the rat brain: role of microglia.

J Neurosci 20:6309–6316

Kim ST, Choi JH, Chang JW, Kim SW, Hwang O (2005a)

Immobilization stress causes increases in tetrahydrobiopterin,

dopamine, and neuromelanin and oxidative damage in the

nigrostriatal system. J Neurochem 95:89–98

Kim YS, Kim SS, Cho JJ, Choi DH, Hwang O, Shin DH, Chun HS,

Beal MF, Joh TH (2005b) Matrix metalloproteinase-3: a novel

signaling proteinase from apoptotic neuronal cells that activates

microglia. J Neurosci 25:3701–3711

Kim ST, Choi JH, Kim D, Hwang O (2006) Increased tyrosine

hydroxylase and neuromelanin in the SN of middle aged mice.

Neurosci Lett 396:263–268

Kim YS, Choi DH, Block ML, Lorenzl S, Yang L, Kim YJ, Sugama

S, Cho BP, Hwang O, Browne SE, Kim SY, Hong JS, Beal MF,

Joh TH (2007) A pivotal role of matrix metalloproteinase-3

activity in dopaminergic neuronal degeneration via microglial

activation. FASEB J 21:179–187

Kraus RL, Pasieczny R, Lariosa-Willingham K, Turner MS, Jiang A,

Trauger JW (2005) Antioxidant properties of minocycline:

neuroprotection in an oxidative stress assay and direct radical-

scavenging activity. J Neurochem 94:819–827

Lai AY, Todd KG (2006) Hypoxia-activated microglial mediators of

neuronal survival are differentially regulated by tetracyclines.

Glia 53:809–816

Lee SY, Moon Y, Hee Choi D, Jin Choi H, Hwang O (2007)

Particular vulnerability of rat mesencephalic dopaminergic

neurons to tetrahydrobiopterin: Relevance to Parkinson’s dis-

ease. Neurobiol Dis 25:112–120

Mohri I, Taniike M, Taniguchi H, Kanekiyo T, Aritake K, Inui T,

Fukumoto N, Eguchi N, Kushi A, Sasai H, Kanaoka Y, Ozono K,

Narumiya S, Suzuki K, Urade Y (2006) Prostaglandin D2-

mediated microglia/astrocyte interaction enhances astrogliosis

and demyelination in twitcher. J Neurosci 9:4383–4393

Mun-Bryce S, Lukes A, Wallace J, Lukes-Marx M, Rosenberg GA

(2002) Stromelysin-1 and gelatinase A are upregulated before

TNF-alpha in LPS-stimulated neuroinflammation. Brain Res

933:42–49

NINDS NET-PD Investigators (2008) A pilot clinical trial of creatine

and minocycline in early Parkinson disease: 18-month results.

Clin Neuropharmacol 31:141–150

Nuttall RK, Silva C, Hader W, Bar-Or A, Patel KD, Edwards DR,

Yong VW (2007) Metalloproteinases are enriched in microglia

compared with leukocytes and they regulate cytokine levels in

activated microglia. Glia 55:516–526

Prall AK, Longo GM, Mayhan WG, Waltke EA, Fleckten B, Thompson

RW, Baxter BT (2002) Doxycycline in patients with abdominal

aortic aneurysms and in mice: Comparison of serum levels and

effect on aneurysm growth in mice. J Vasc Surgery 35:923–929

Rohl C, Lucius R, Sievers J (2007) The effect of activated microglia

on astrogliosis parameters in astrocyte cultures. Brain Res

1129:43–52

Rosenberg GA (2002) Matrix metalloproteinases in neuroinflamma-

tion. Glia 39:279–291

Sanchez-Pernaute R, Ferree A, Cooper O, Yu M, Brownell AL,

Isacson O (2004) Selective COX-2 inhibition prevents progres-

sive dopamine neuron degeneration in a rat model of Parkinson’s

disease. J Neuroinflammation 1:6

Sapadin AN, Fleischmajer R (2006) Tetracyclines: nonantibiotic

properties and their clinical implications. Am Acad Dermatol

54:258–265

Seo JW, Srisook E, Son HJ, Hwang O, Cha YN, Chi DY (2005)

Syntheses of NAMDA derivatives inhibiting NO production in

BV-2 cells stimulated with lipopolysaccharide. Bioorg Med

Chem Lett 15:3369–3373

Smith K, Leyden JJ (2005) Safety of doxycycline and minocycline: a

systematic review. Clin Ther 27:1329–1342

Smith VA, Khan-Lim D, Anderson L, Cook SD, Dick AD (2008)

Does orally administered doxycycline reach the tear film? Br J

Ophthalmol 92:856–859

Soory M (2008) A role for non-antimicrobial actions of tetracyclines

in combating oxidative stress in periodontal and metabolic

diseases: a literature review. Open Dent J 2:5–12

Stirling DP, Koochesfahani KM, Steeves JD, Tetzlaff W (2005)

Minocycline as a neuroprotective agent. Neuroscientist 11:308–

322

Suri C, Fung BP, Tischler AS, Chikaraishi DM (1993) Catecholam-

inergic cell lines from the brain and adrenal glands of tyrosine

hydroxylase-SV40 T antigen transgenic mice. J Neurosci 13:

1280–1291

Thomas M, Le WD, Jankovic J (2003) Minocycline and other

tetracycline derivatives: a neuroprotective strategy in Parkin-

son’s disease and Huntington’s disease. Clin Neuropharmacol

26:18–23

Toth A, Lesser ML, Naus G, Brooks C, Adams D (1988) Effect of

doxycycline on pre-menstrual syndrome: a double-blind random-

ized clinical trial. J Int Med Res 16:270–279

Visse R, Nagase H (2003) Matrix metalloproteinases and tissue

inhibitors of metalloproteinases: structure, function, and bio-

chemistry. Circ Res 92:827–839

Walker DG, Lue LF (2005) Investigations with cultured human

microglia on pathogenic mechanisms of Alzheimer’s disease and

other neurodegenerative diseases. J Neurosci Res 81:412–425

Woo MS, Park JS, Choi IY, Kim WK, Kim HS (2008) Inhibition of

MMP-3 or -9 suppresses lipopolysaccharide-induced expression

of proinflammatory cytokines and iNOS in microglia. J Neuro-

chem 106:770–780

Yao JS, Shen F, Young WL, Yang GY (2007) Comparison of

doxycycline and minocycline in the inhibition of VEGF-induced

smooth muscle cell migration. Neurochem Int 50:524–530

Yrjanheikki J, Keinanen R, Pellikka M, Hokfelt T, Koistinaho J (1998)

Tetracyclines inhibit microglial activation and are neuroprotective

in global brain ischemia. Proc Natl Acad Sci USA 95:15769–

15774

Zhang D, Hu X, Qian L, Wilson B, Lee C, Flood P, Langenbach R,

Hong JS (2009) Prostaglandin E2 released from activated

microglia enhances astrocyte proliferation in vitro. Toxicol Appl

Pharmacol 238(1):64–70

Neurotox Res (2009) 16:361–371 371

123