ava i l ab l e a t www.sc i enced i r ec t . com

C l i n i ca l Immuno logy

www.e l sev i e r . com / l oca te / yc l im

Clinical Immunology (2014) 153, 40–48

Immunomodulation by the estrogenmetabolite 2-methoxyestradiol

Alexandra Stubelius⁎, Malin C. Erlandsson, Ulrika Islander, Hans Carlsten

Centre for Bone and Arthritis Research (CBAR), Department of Rheumatology and Inflammation Research,Institute of Medicine, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden

Received 23 December 2013; accepted with revision 19 March 2014Available online 29 March 2014

Abbreviations: E2, 17β-Estradiol;COMT, Catechol-O-methyltransferaseERE, Estrogen response elements;Orchidectomized; RA, Rheumatoid artgen receptor modulators.⁎ Corresponding author at: Centre fo

(CBAR), Department of RheumatologySahlgrenska Academy, University of GGöteborg, Sweden. Fax: +46 31 82 39

E-mail addresses: alexandra.stubel(A. Stubelius), malin.erlandsson@rheuulrika.islander@rheuma.gu.se (U. Islanhans.carlsten@rheuma.gu.se (H. Carls

http://dx.doi.org/10.1016/j.clim.2011521-6616/© 2014 The Authors. Pu(http://creativecommons.org/licenses

KEYWORDS2-Methoxyestradiol;Immune system;Immune regulation;17β-Estradiol;Uterus proliferation;Estrogen response elements

Abstract 2-methoxyestradiol (2me2), a metabolite of 17β-estradiol (E2), has been tested inphase II clinical cancer trials and models of inflammation. Its effects are only partly clarified. Weinvestigated the effects of 2me2 on the immune system, using ovariectomized or sham-operatedmice treated with a high and a low dose of 2me2 (2me2H and 2me2L), E2 or vehicle. Weinvestigated antagonism of tissue proliferation and estrogen response element (ERE) activation.Established immunomodulation by E2 was reproduced. 2me2L increased NK and T-cells frombone marrow, spleen and liver. Both 2me2H and E2 induced uterus proliferation inovariectomized mice, but no antagonistic effects on uteri growth were seen in intact animals.

Both E2 and 2me2H activated EREs.Immunomodulation by 2me2 is tissue-, and concentration dependent. E2 regulated the immunesystem more potently. The higher dose of 2me2 resulted in E2 like effects, important to considerwhen developing 2me2 as a drug.© 2014 The Authors. Published by Elsevier Inc. This is an open access article under the CCBY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

2me2, 2-Methoxyestradiol;; ERs, Estrogen receptors;ovx, Ovariectomized; orx,hritis; SERM, Selective estro-

r Bone and Arthritis Researchand Inflammation Research,othenburg, Box 480, 405 3025.ius@rheuma.gu.sema.gu.se (M.C. Erlandsson),der),ten).

4.03.011blished by Elsevier Inc. T/by-nc-nd/3.0/).

his

1. Introduction

2-methoxyestradiol (2me2) is an endogenous metaboliteof 17β-estradiol (E2). It exerts anti-angiogenic and anti-proliferative effects, and has been evaluated in phase II clinicaltrials for the treatment of cancers [1–5]. It has also been testedin animal models of rheumatoid arthritis (RA) [6,7] and,recently, multiple sclerosis [8]. In thesemodels, 2me2 inhibitedlocal inflammation, angiogenesis, and leukocyte infiltration, aswell as bone resorption. The mechanisms underlying thesedifferent effects are not clear. 2me2's known mode of actionsincludes disrupting microtubules leading to apoptosis, and

is an open access article under the CC BY-NC-ND license

41Immune regulation by 2-methoxyestradiol

regulation of hypoxia-inducible factor 1α (HIF-1α), active ininflammatory settings [5].

Estrogens regulate the immune system; both affecting T-,and B-cells, as well as down regulating NK cell cytotoxicity[9–14]. In some inflammatory models, many effects of 2me2resemble actions seen after exposure to E2 [7,15]. E2 isconverted to 2me2 through the P450 system and theenzyme catechol-O-methyltransferase (COMT). COMT defi-cient (COMT−/−) mice develop preeclampsia and showa dysregulation of their decidual NK cells [16]. Thepre-eclampsia was reversed when the COMT−/− mice weregiven 2me2. In addition, 2me2 levels are lower in patientswith preeclampsia [16–18]. In a recent study, we presenteddata showing that mice lacking COMT have an alteredimmune phenotype [19]. This suggests that not only E2, butalso 2me2, can regulate the immune system.

In vitro, 2me2 has a 500-fold lower affinity for theestrogen receptor α (ERα) than E2 [20]. During pregnancy,levels of 2me2 increase more than 1000-fold, and there isconflicting data regarding whether or not administration of2me2 can lead to ER activation in vivo [7,15,21,22]. ERαsignaling mediates the action of E2 on bone and uterus [23].In animal models, 2me2 treatment caused uterus prolifera-tion and bone remodeling, as well as T-cell regulation[7,15]. Since 2me2 is interesting not only from a physiolog-ical point of view, but also pharmacologically, there is needfor further studies on estrogen related effects by 2me2. Wetherefore studied whether 2me2 exerted immunomodulatingeffects in tissues and cells known to respond to E2. Inaddition, we tried to elucidate the mechanisms or pathwaysbehind the effects by giving 2me2 to estrogen responseelement-coupled luciferase-mice and treatingestrogen-intact animals with 2me2 in order to investigatecompetitive blocking of estrogen receptors.

2. Materials and methods

2.1. Animals and experimental procedures

Female C57/Bl6 mice were purchased from Charles River,Ry, Denmark. All mice were housed in a temperature-controlled room with a 06.00–18.00 h light cycle andconsumed a soy-free diet (R70, Lantmännen, Stockholm,Sweden) and tap water ad libitum. Luciferase reporter micecoupled to estrogen response elements (ERE-luciferasemice) enabling in vivo detection of activation of ERE andclassic estrogen signaling were a kind gift from Jan-ÅkeGustafsson and Paul T van der Saag [24]. The EthicsCommittee on Animal Cares and Use in Gothenburg approvedall procedures.

At nine weeks of age, female mice were ovariectomized(ovx) or sham operated, and male mice were orchidectomized(orx) under isoflurane (Baxter Medical Ab, Kista, Sweden)anesthetization by inhalation. Carprofen (Orion Pharma AB,Animal Health, Sollentuna, Sweden) was administered subcu-taneously, as postoperative analgesic.

2.2. Drug administration

The dose of 2me2 was chosen based on previous studieswhere 2me2 and E2 exhibited anti-inflammatory properties

in the collagen induced arthritis model, and reducedatherosclerotic lesions [7,25]. Slow release pellets with2-methoxyestradiol, 17β-estradiol or vehicle were im-planted subcutaneously (Innovative Research of America,Sarasota, FL, USA). Mice were given vehicle, 2me2 low dose6.66 μg/day (2me2L; 0.4 mg/pellet/60 days), 2me2 highdose 66.6 μg/day (2me2H; 4 mg/pellet/60 days) or E20.83 μg/day (0.05 mg/pellet/60 days). For the sham study,mice were divided into vehicle, 2me2 low dose 6.66 μg/day(0.4 mg/pellet/60 days) and E2 (0.05 mg/kg/day). Administra-tion coincided with the ovx/sham operation, continuingfor three weeks, until mice were terminated. For the ERE-luciferase study, subcutaneous injections of 17β-estradiol-3-benzoate (E2; 0.06 mg/kg; Sigma Aldrich, Stockholm, Swe-den) 2me2 low dose 6.66 μg (2me2L; 0.38 mg/kg), 2me2 highdose 2me2H; 66.6 μg (3.8 mg/kg) or Miglyoil 812 (OmyaPeralta, Hamburg Germany; 100 μl/mouse) were administered4 h prior to termination.

2.3. Tissue collection

Mice were terminated three weeks after pellet implanta-tion. At time of termination, mice were anesthetized withketamine (Pfizer AB, Täby, Sweden) and medetomidine(Orion Pharma AB), bled and killed by cervical dislocation.Thymus, liver, spleen and uterus were weighed, and thenthymus, spleen and whole bone were put in PBS. The liverwas put into RPMI 1640 (without phenol red, PAA Laborato-ries GmbH, Pashing, Austria).

2.4. Single cell preparation andimmune-phenotyping by flow cytometry

Single cell suspensions were prepared from the spleen,thymus, liver, and bone marrow as previously described[19]. Cells from spleen, marrow, and liver were subjected toerythrocyte lysis in Tris-buffered 0.83% NH4Cl. All cells werewashed with PBS and put on ice until the total number ofleukocytes was calculated using an automated cell counter(Sysmex, Kobe, Japan).

Expression of cell surface markers was detected usingfluorochrome conjugated antibodies following Fc-blocking(anti-mouse CD16/CD32; clone 2.4G2); anti-CD122 FITC(Clone TM-Beta 1), anti-CD3 PerCp (Clone I45-2C11), anti-NK1.1 APC (Clone PKI36), anti-CD27 PE (Clone LG.3A10),anti-Ly49C/I PE (Clone 5E6; BD Biosciences, San Diego, CA,USA) anti-CD19 PerCp (Clone 6D5; Biolegend, San Diego, CA,USA). Fluorochrome-minus-one (FMO) was used as control.Cells were acquired by a FACSCanto II (BD Biosciences)equipped with Diva Software (BD Biosciences), and data wasfurther processed using FlowJo software (Three Star, Inc.,Ashland, VA, USA). Data is presented as percentage positiveand viable leukocytes.

2.5. NK cell cytotoxicity assay

To test NK cells capacity to lyse cells, the NK cell cytotoxiccapability was assayed using the CytoTox 96 non-radioactivecytotoxicity assay (Promega, Madison, WI, USA) using YAC-1as target cells, as described previously [19]. Briefly, splenicNK cells were negatively enriched using the NK Cell Isolation

Ovx+v

ehicl

e

Ovx+2

me2

L

Ovx+2

me2

H

Ovx+E

20

10

20

30

40

50 ******

***

B c

ells

(%

CD

19+

)

Sham

+veh

icle

Sham

+veh

icle

Ovx+v

ehicl

e

Ovx+2

me2

L

Ovx+2

me2

H

Ovx+E

20

10

20

30

40B

one

mar

row

cel

ls (

#106 )

* ***

******

A B C

B cells(CD19)

Figure 1 Effects on B lymphopoesis: a classic estrogen regulated immune parameter. Nine week old female mice were either ovx orsham operated and exposed to vehicle, 2me2L (6.66 μg/day), 2me2H (66.6 μg/day), or E2 (0.83 μg/day) for three weeks. The effectson cell number in the bone marrow (A) were evaluated. B-cell development in the bone marrow was evaluated using FACS andantibodies against CD19 (B cells; B). C shows a representative FACS plot of CD19+ cells from the bone marrow. Dots representindividuals; bars represent the arithmetic mean per group. *p b 0.05, ***p b 0.001 versus ovx + vehicle treated mice.

42 A. Stubelius et al.

kit (Miltenyi Biotec GmbH, Germany). An effector vs targetratio of 1:1, 2.5:1 and 10:1 was used. Each sample wasincubated at 37°C in a humidified atmosphere of 5% CO2 for4 h. The data is presented as “% specific lysis”.

2.6. Confocal microscopy

Cells from the cytotoxicity assay (target cells together witheffector cells) were fixed in 2% paraformaldehyde, blockedwith Fc-blocking (BD Biosciences), and stained with biotinanti-mouse NK1.1 (Biolegend) and streptavidin Alexa Fluor 488(Life Technologies Europe BV, Bleiswijk, The Netherlands). Ascontrols, cells were stained with secondary antibody only, ordirectly mounted with DAPI excluding both primary andsecondary antibodies. Cells were mounted on slides usingProlong Gold Antifade Reagent with DAPI (Life Technologies).Cells were visualized using an LSM 700 confocal microscope

Figure 2 Modulation of T-, and NK cells: proposed 2me2 targetesham operated and exposed to vehicle, 2me2L (6.66 μg/day), 2me2Hby E2 and 2me2 of T- (CD3+), and NK cells (CD122+CD3−NK1.1+), an(Ly49C/I) in the bone marrow (A, D, G), spleen (B, E, H), and liver (C,tested for their ability to lyse YAC-1 cells measuring the release ofrepresentative image of the NK cells together with YAC-1 cells from tand green cells are stained for NK1.1. In A-I, dots represent individscale, and geometric mean on a logged scale. In J, dots represent thrange. *p b 0.05, ** p b 0.01, ***p b 0.001 versus ovx + vehicle treat

through a 40× plan-NEOFLUAR objective with 2 timesaveraging.

2.7. Protein preparation and luciferase analysis

Mice were sacrificed 4 h after injections; organs wereharvested and individually frozen in liquid nitrogen. Proteinpreparation and luciferase activity were performed aspreviously described, relating activity to amount of protein[24]. Luciferase activation is proportional to the activationof gene transcription when ERs bind to EREs and is thereforeconsidered a measurement of signaling via the classic pathway.

2.8. Statistical analysis

Statistical evaluations were performed using IBM SPSSStatistics software 22.0.0.0 and GraphPad Prism version6.0c. One-way ANOVA with Tukey's post hoc test was used

d immune cells. Nine week old female mice were either ovx or(66.6 μg/day), or E2 (0.83 μg/day) for three weeks. ModulationNK cell activation receptor (CD27), and an inhibitory receptorF, I) was evaluated. NK cells from the spleen were enriched andlactate dehydrogenase as compared to controls (J). K shows ahe cytotoxicity assay. Blue represents nuclear staining with DAPIuals; bars represent the arithmetic mean per group on a lineare arithmetic mean of 5–12 mice per group and error bars depicted mice.

Sham

+veh

icle

Ovx+v

ehicl

e

Ovx+2

me2

L

Ovx+2

me2

H

Ovx+E

20

204050

60

70

80

90

100

% C

D27

+N

K c

ells

***

Sham

+veh

icle

Ovx+v

ehicl

e

Ovx+2

me2

L

Ovx+2

me2

H

Ovx+E

20

20

40

60

% L

y49c/

i+N

K c

ells

*****

Sham

+veh

icle

Ovx+v

ehicl

e

Ovx+2

me2

L

Ovx+2

me2

H

Ovx+E

21

10

100

% C

D3+

*

Sham

+veh

icle

Ovx+v

ehicl

e

Ovx+2

me2

L

Ovx+2

me2

H

Ovx+E

20.1

1

10

% C

D12

2+C

D3- N

.K1.

1+

******** ***

1:1

2,5:

110

:10

10

20

30

% S

peci

fic ly

sis

E2Ovx+vehicle 2me2L

*

Sham

+veh

icle

Ovx+v

ehicl

e

Ovx+2

me2

L

Ovx+2

me2

H

Ovx+E

20

20

40

60

80

% C

D27

+N

K c

ells

***

Sham

+veh

icle

Ovx+v

ehicl

e

Ovx+2

me2

L

Ovx+2

me2

H

Ovx+E

20

20

40

60

80

% C

D3+

******

***

Sham

+veh

icle

Ovx+v

ehicl

e

Ovx+2

me2

L

Ovx+2

me2

H

Ovx+E

20

2

4

6

8

% C

D3+

*****

***

Bone marrow Spleen LiverA

D

G

B

H

C

I

K

T c

ells

NK

cel

l rec

epto

rs

S ham

+veh

icle

Ovx+v

ehicl

e

Ovx+2

me2

L

Ovx+2

me2

H

Ovx+E

20

2

4

6

8

10%

CD

122+

CD

3- N.K

1.1+

**

E

Sham

+veh

icle

Ovx+v

ehicl

e

Ovx+2

me2

L

Ovx+2

me2

H

Ovx+E

20

5

10

15

20

% C

D12

2+C

D3- N

.K1.

1+

*F

J

NK

cel

ls

43Immune regulation by 2-methoxyestradiol

Sham

+veh

icle

Ovx+v

ehicl

e

Ovx+2

me2

L

Ovx+2

me2

H

Ovx+E

20.1

1

10

Ute

rus

wei

ght o

f tot

al b

ody

wei

ght (

%)

***p=0.07

******

Sham

+veh

icle

Ovx+v

ehicl

e

Ovx+2

me2

L

Ovx+2

me2

H

Ovx+E

20

100

200

300

Thy

mus

cel

ls (

#106 )

*** ***

******

A

Sham

+veh

icle

Ovx+v

ehicl

e

Ovx+2

me2

L

Ovx+2

me2

H

Ovx+E

20

2

4

6

Thy

mus

wei

ght o

f tot

al b

ody

wei

ght (

%)

*** ***

******

B C

Figure 3 Effects of E2 and 2me2 on classic estrogen responsive tissues. Nine week old female mice were either ovx or shamoperated and exposed to vehicle, 2me2L (6.66 μg/day), 2me2H (66.6 μg/day), or E2 (0.83 μg/day) for three weeks. Uterus (A) andthymus (B) were weighed at termination. Data is presented as % tissue weight of total body weight. Thymus cells were purified andleukocytes were counted (C). Dots represent individuals; bars represent the arithmetic mean per group on a linear scale, andgeometric mean on a logged scale. ***p b 0.001 versus ovx + vehicle treated mice.

44 A. Stubelius et al.

for comparison of three or more independent groups.Whenever Levene's test revealed unequal variances betweenthe groups, Dunnett's T3 post hoc test was used instead.Logarithmic transformations were used when appropriate toensure normal distribution of data. ANCOVA was used whenadjustments for covariates were needed, i.e. when twoexperiments were pooled. All analyses were two-tailed, andp values b 0.05 were considered statistically significant.

B

Sham

+veh

icle

Sham

+2m

e2L

Sham

+E2

16

18

20

22

24

26

Bod

y W

eigh

t (g)

**

Sham

+veh

icle

Sham

+2m

e2L

Sham

+E2

0

5

10

15

Ute

rus

wei

ght o

f tot

al b

ody

wei

ght (

%)

A

Figure 4 Effects of E2 and 2me2 on classic estrogen responsive teither ovx or sham operated and exposed to vehicle, 2me2L (6.66 μuterus (B), thymus (C), and liver (D) were weighed at termination. TDots represent individuals; bars represent the arithmetic mean per

3. Results

3.1. Immunomodulating effects by 2me2 and E2

3.1.1. B-cell development is affected by E2 and 2me2Loss of estrogens during menopause, or ovariectomy, leadsto bone loss and affects lymphopoesis [13]. In turn, exposureto estrogens affects B-cell maturation in the bone marrow,

Sham

+veh

icle

Sham

+2m

e2L

Sham

+E2

0

1

2

3

4

5

Thy

mus

wei

ght o

f tot

al b

ody

wei

ght (

%) **

DC

Sham

+veh

icle

Sham

+2m

e2L

Sham

+E2

2

4

6

8

Live

r w

eigh

t of t

otal

bod

y w

eigh

t (%

)

issues in ovary intact animals. Nine week old female mice wereg/day), or E2 (0.83 μg/day) for three weeks. Body weight (A),issue data is presented as % tissue weight of total body weight.group. **p b 0.01 versus vehicle treated mice.

Vehicl

e

2me2

L

2me2

H E21

10

Luci

fera

se a

ctiv

ity

****

Vehicl

e

2me2

L

2me2

H E21

10

Luci

fera

se a

ctiv

ity

******

Vehicl

e

2me2

L

2me2

H E2 1

10

Luci

fera

se a

ctiv

ity

*******

BASpleen

CLiverBone Marrow

Figure 5 Estrogen response elements-coupled luciferase activation by E2 and 2me2. Luciferase activity in tissues was evaluated inERE-coupled luciferase orchidectomized (orx) male mice 4 h after exposure to vehicle, 2me2L (6.66 μg), 2me2H (66.6 μg), or E2(1 μg). Data is presented as luciferase activity after relating to protein content, in (A) bone marrow, (B) spleen, and (C) liver. Dotsrepresent individuals; bars represent the geometric mean per group. *p b 0.05, ***p b 0.001 versus vehicle treated mice.

45Immune regulation by 2-methoxyestradiol

resulting in a diminished CD19+ population [14]. In concor-dance with previous studies, we could show that ovx led toan increased amount of cells in the bone marrow (Fig. 1A),and increased the CD19+-cell population (Fig. 1B and C).Treatment with 2me2L, 2me2H, and E2 for three weeksdiminished the number of cells in the bone marrow (Fig. 1A)compared to vehicle. Treatment with both E2 and 2me2H,but not 2me2L restored the CD19+ cell population equivalentto sham operated animals (Fig. 1B). In the spleen, B-cellsdiminished after treatment with E2, but not 2me2L (vehicle:56 ± 10%, 2me2L: 52 ± 5%, E2: 48 ± 3% (Mean ±SD). Vehiclevs. E2: p b 0.05, *; vehicle vs. 2me2L: n.s.).

3.1.2. Peripheral T-, and NK-cells are affected by E2and 2me2Previous studies have indicated that both peripheral T-, andNK cells are targets for 2me2 activity [8,16]. Exposure to E2for three weeks minimum has also been shown to affect NKcells and inhibit their cytotoxicity [10,11]. We found thatthree weeks of treatment with E2 and both doses of 2me2produced an expansion of T-cells in the bone marrow and theliver (Fig. 2A,C). In the spleen however, only the low dose of2me2 expanded the T-cell population (Fig. 2B). We could alsoshow that the NK cell population in the bonemarrowdiminishedafter ovx and expanded after all treatments (Fig. 2D). Thedifference compared with vehicle was statistically significant,and the magnitude of the response comparable for all threetreatments. In the spleen and liver, only 2me2L expanded theNK cell population (Fig. 2E,F). To investigate these effectsfurther, we examined subpopulations of NK cells. We studiedNK cell expression of the activation receptor CD27 in the bonemarrow and spleen, and the inhibitory receptor Ly49C/I in theliver. CD27 is a tumor necrosis factor (TNF) receptor, andCD27hi NK cells have increased cytotoxic capacity, produce

more cytokines, andmigrate to a higher extent [26]. Since CD27also has been shown to be part of the developmental stepsfor the NK cell, we analyzed this receptor in the bone marrowas well as in the spleen. Treatment with E2 increased theexpression of CD27 in the bone marrow (Fig. 2G), while theexpression in the spleen decreased (Fig. 2H). The inhibitoryreceptor Ly49C/I is implied to convey memory functions ofNK cells specifically in the liver [27]. We found that ovx downregulated the expression of Ly49C/I, whereas treatment with E2and 2me2H up regulated the receptor (Fig. 2I).

Following the findings of an expanded NK cell populationby 2me2L in the spleen and a down regulation of CD27 by E2,we tested the cytotoxic capacity of NK cells from the spleen.We could repeat the classic E2 down regulation of NK cellcytotoxicity as previously reported, but we failed to showany effect by treatment with 2me2 at a low dose (Fig. 2J).

3.2. Estrogen signaling by 2me2

3.2.1. Tissue sensitivity for E2 and 2me2 after long-termadministration in ovx animals and intact animalsMost previous studies have focused on mechanisms relatedto 2me2's effects on microtubule disruption and have nottaken into account possible activation of ERs. Both theuterus and the thymus are tissues highly responsive toestrogen stimulation. Estrogens affect T-cell maturation inthe thymus, where thymic involution is dramatic, andmassive loss of cells ensues. Three weeks after ovx micedisplayed uterus involution (Fig. 3A) and thymic growth(Fig. 3B). Both treatment with 2me2H and E2 increasedproliferation of the uterus to the same extent (Fig. 3A), andat the same time induced atrophy of the thymus (Fig. 3B).Treatment with 2me2L caused uterus proliferation, but less

46 A. Stubelius et al.

than in sham-operated animals (Fig. 3A, p = 0.07), andinduced thymic involution (Fig. 3B–C).

3.2.2. 2me2 does not act as an antagonist on ERsERα mediates proliferation of the uterus [23]. The effect of aligand after binding to ERs, and its following transcriptionalregulation, is a complex systemdepending on interactions withdifferent co-activators and co-repressors [28]. In fact, selec-tive estrogen receptor modulators (SERM) can act as antago-nists on ERs when physiological E2 is present, depending onrecruitment of different co-molecules [29]. We investigatedpossible antagonistic effects of 2me2 on estrogen-sensitivetissues by giving 2me2L, vehicle or E2 to ovary intact shamoperated animals for three weeks. We used only 2me2L as2me2H resembled E2 effects. Treatment with E2 caused anincrease in body weight (Fig. 4A). However, no antagonisticeffectswere seen after treatmentwith 2me2L on uteri growth,thymic involution, or liver weight (Fig. 4B–D).

3.2.3. High dose of 2me2 activates estrogen responseelements (ERE) after short-term treatmentThe uteri weight after long-term treatment of 2me2 demon-strates a dose-dependent uterotrophic effect (Fig. 3A). Indeed,2me2H induced a uterus weight similar to that of mice exposedto E2 (Fig. 3A). We have previously published data showing thatthe high dose of 2me2 can activate ERE to the same extent asE2 in the uterus [7]. In order to elucidate if 2me2 could activateEREs also in other organs, in a system devoid of naturalestrogens, orchidectomized male luciferase-ERE reporter micewere given 2me2L (6.66 μg), or 2me2H (66.6 μg), vehicle, orE2 (1 μg) 4 h prior to termination. Data show activation of EREin bone marrow, spleen and liver after exposure to both E2 and2me2H (Fig. 5A,B,C). 2me2L only activated EREs in the bonemarrow (Fig. 5A). The same results were seenwhen repeated inovx females after administration with E2 and 2me2L (data notshown).

4. Discussion

The purposes of these studies were to investigate theimmunomodulating effects of 2me2, in comparison to thoseof E2. Similar to E2, 2me2 is an endogenous steroid, actingin several tissues on multiple targets [30,31]. 2me2 hasbeen evaluated in clinical trials of cancers, and models ofdifferent inflammatory diseases. We have identified threekey findings in the presented studies. First, 2me2 is a lesspotent immune-modulator than E2. Second, the higher theconcentration of 2me2, the more E2-like effects on immunecells and estrogen responsive tissues are detected. Finally,2me2 seems to regulate cells in a tissue-, and dose-dependent manner.

In our study we have used two doses of 2me2; 2me2L and2me2H, to investigate its efficacy, and compared theseeffects with an established immune-modulatory dose of E2.These doses of 2me2 were chosen based on our previousstudies showing dose-dependent amelioration of collageninduced arthritis and underlying inflammation, by both 2me2doses [7]. The 2me2L dose used in this study is equimolar of6 μg/day E2, which allowed us not only to study efficacy of2me2, but also its potency compared to E2.

Estrogens affect B lymphopoesis in the bone marrow [14].We therefore used this model to compare the immunemodulation by E2 and 2me2. Both E2 and 2me2H affected Blymphopoesis but not 2me2L in the bone marrow (Fig. 1B). Inthe spleen, only E2 diminished the B-cell population. Weconclude that E2 affects B lymphopoesis more potently than2me2.

According to recent publications, 2me2 targets T lym-phocytes and NK cells [8,16], and long-term treatment withestrogens can down regulate NK cell cytotoxicity [10,11]. Inour study, we found that both NK-, and T-cells wereregulated equally by both E2 and 2me2 in the bone marrow(Fig. 2A and D). In the peripheral organs however, adifferent pattern could be identified. NK cells in the liver,and both NK cells and T-cells in the spleen, expanded afterexposure to 2me2L, but not after E2 and 2me2H. However,only E2 downregulated NK cell cytotoxicity. As suggested byHayakawa et al. [26], this reduced killing capacity correlat-ed to a reduced CD27 expression, thus affecting TNFmediated functions. Reducing the activation by TNF couldbe a mechanism E2 employs to decrease NK cell cytotoxicity.The inhibitory Ly49C/I receptor on NK cells in the liver wasalso downregulated by E2. So even though we see a shift inthe NK cell population by 2me2L, only E2 regulates NK cellfunctionality. A recent study found that lower plasma levelsof E2 coincided with an increased tissue concentration of2me2 [32], suggesting a physiological interplay between thehormones. This steroid relationship regulated angiogenesis,and could be an explanation for the found regulation of theimmune response. As higher doses of 2me2 exert similaractions as E2, it is possible that administration of 2me2H isconverted into E2, thus regulating the immune system as themore potent form of E2. Then, a lower concentration mightbe needed to manifest 2me2's own regulatory effects. In anE2-free environment, we show that a lower 2me2 concen-tration did not activate EREs in the spleen and liver, (Fig.5B–C), and this coincided with a regulation of NK cellpopulations, corroborating a regulatory role for lower con-centrations of 2me2. Based on the consistency of ourfindings, we conclude that E2 is a more potent immunomod-ulator than 2me2. However, 2me2 seems to regulate T-, andNK cell populations, in a dose and tissue dependent manner.

Mechanistically, ERα-activation leads to some of the sideeffects seen in association with hormone replacementtherapy [33]. 2me2 has shown low toxicity when tested inclinical trials for cancers [2,4], and low affinity for the ERs invitro [20]. ERs bind to its ligand in the classic transcriptionpathway, form a receptor dimer, and translocate to thecell nucleus. There they form complexes together withco-regulatory proteins, and initiate transcription afterbinding to estrogen response elements (ERE). This results inagonism or antagonism, and is a complex system dependingon conformational changes of the receptors and subsequentrecruitment of different co-activators or co-repressors. Thiscomplex system is exploited by the selective estrogenreceptor modulators (SERM) [34]. SERMs distinguish them-selves from pure agonists as their effects differ in varioustissues, and thereby selectively stimulate or inhibit estrogenlike actions, often in the presence of physiological E2 [35].Indeed, the SERM Raloxifene is approved as a prophylactic forinvasive breast cancer in the US. We wanted to investigatepossible antagonistic effects after administration of 2me2 in

47Immune regulation by 2-methoxyestradiol

the presence of physiological E2. Based on the findings thatERα mediates the utero-proliferative effects of E2 [23], andthe thymus and liver are known to respond to E2, we usedthese as target organs. We treated ovary-intact mice withvehicle, E2 and 2me2 for three weeks. Neither the uteri,thymus, nor liver weights were affected after administrationof 2me2 in intact mice, indicating that 2me2 does notantagonize ERs in these tissues.

Further, we tested estrogen signaling by 2me2 in micewith the luciferase reporter gene coupled to EREs. Signalingthrough ERE was investigated 4 h after administration toprevent complete metabolism of 2me2 [36]. We coulddetect ERE activation by both E2 and 2me2H in the spleenand liver. In these organs, the low dose of 2me2 did notactivate ERE. As suggested previously, one possible mecha-nism for the ER activation seen after 2me2 administrationcould be the demethylation of 2me2 into 2-hydroxyestradiol,followed by conversion back into 17β-estradiol [37–39].Corroborating the demethylation theory, we found a dosedependent regulation of 2me2 resembling E2, i.e. the higherthe 2me2 dose, the more E2-like effects. If equilibriumexists between 2me2 and E2, allowing both hormones to bepresent at a specific location, the E2 effects are morepotent. The difference in hierarchy between the hormonesmight be a question of affinity to the ERs, or that thetranscriptional activity after ER activation is more potentthan for example HIF-1α signaling, which 2me2 regulates. Tofurther investigate the demethylation theory, we adminis-tered an even higher dose of 2me2 to ovx mice for threeweeks, 333 μg/day. At this dose however, the agent was toxic.

To summarize, there are a number of different E2metabolites whose physiological functions are yet to beunderstood. 2me2 functions include anti-angiogenesis,anti-apoptosis and anti-proliferation [5]. 2me2's role ininflammation has been investigated either in in vitrosystems, or in in vivo models of inflammation. Here, westudied the modulation of the immune system after threeweeks with 2me2 at 6.66 μg/day and 66.6 μg/day comparedto E2 0.83 μg/day. We found that 2me2 affected severalimmune-competent cells and organs in a dose dependentway. These effects are consistent with those of E2 in someorgans and target cells, but not uniformly so.

5. Conclusions

We conclude that 2me2 has immunomodulatory effects in adose dependent way, and the effects are partly similar tothat of E2modulation. As 2me2 has been investigated in clinicaltrials of cancer, and is under consideration for clinical trials ofother diseases, it is important to clarify all its potential immunemodulatory properties. Local tissue concentrations of 2me2could have a physiological immune regulatory role. Based onour data however, we cannot exclude that pharmacologicaldoses of 2me2, and exceeding the window of efficacy, couldlead to unwanted estrogen-related effects, which furtherstudies need to investigate.

Declaration of interest

The authors declare no competing interests.

Funding

This study was supported by grants from the Medical Facultyof Göteborg University (ALF), the Göteborg Medical Socie-ty, the Börje Dahlin Foundation, the Association againstRheumatism, Reumaforskningsfond Margareta, the COM-BINE Network, Ingabritt and Arne Lundbergs Foundation,and the Swedish Research Council (2011-2521).

Acknowledgments

We thank Paul T van der Saag and Jan-Åke Gustafsson forkindly providing us with the luciferase reporter mice. Wethank Margareta Rosenqvist for her excellent technicalassistance.

References

[1] J.Y. Bruce, et al., A phase II study of 2-methoxyestradiolnanocrystal colloidal dispersion alone and in combination withsunitinib malate in patients with metastatic renal cellcarcinoma progressing on sunitinib malate, Investig. NewDrugs 30 (2) (2012) 794–802.

[2] M.R. Harrison, et al., A phase II study of 2-methoxyestradiol(2ME2) nanocrystal(R) dispersion (NCD) in patients withtaxane-refractory, metastatic castrate-resistant prostate can-cer (CRPC), Investig. New Drugs 29 (6) (2011) 1465–1474.

[3] M.H. Kulke, et al., A prospective phase II study of 2-methoxyestradiol administered in combinationwith bevacizumabin patients with metastatic carcinoid tumors, Cancer Chemother.Pharmacol. 68 (2) (2011) 293–300.

[4] D. Matei, et al., Activity of 2 methoxyestradiol (Panzem NCD)in advanced, platinum-resistant ovarian cancer and primaryperitoneal carcinomatosis: a Hoosier Oncology Group trial,Gynecol. Oncol. 115 (1) (2009) 90–96.

[5] N.J. Mabjeesh, et al., 2ME2 inhibits tumor growth andangiogenesis by disrupting microtubules and dysregulatingHIF, Cancer Cell 3 (4) (2003) 363–375.

[6] E. Brahn, et al., An angiogenesis inhibitor, 2-methoxyestradiol,involutes rat collagen-induced arthritis and suppresses geneexpression of synovial vascular endothelial growth factor andbasic fibroblast growth factor, J. Rheumatol. 35 (11) (2008)2119–2128.

[7] A. Stubelius, et al., Role of 2-methoxyestradiol as inhibitorof arthritis and osteoporosis in a model of postmenopausalrheumatoid arthritis, Clin. Immunol. 140 (1) (2011) 37–46.

[8] G.S. Duncan, et al., 2-Methoxyestradiol inhibits experimentalautoimmune encephalomyelitis through suppression of immunecell activation, Proc. Natl. Acad. Sci. U. S. A. 109 (51) (2012)21034–21039.

[9] M.C. Erlandsson, et al., Oestrogen receptor specificity inoestradiol-mediated effects on B lymphopoiesis and immuno-globulin production in male mice, Immunology 108 (3) (2003)346–351.

[10] N. Hanna, M. Schneider, Enhancement of tumor metastasis andsuppression of natural killer cell activity by beta-estradioltreatment, J. Immunol. 130 (2) (1983) 974–980.

[11] N. Nilsson, H. Carlsten, Estrogen induces suppression of naturalkiller cell cytotoxicity and augmentation of polyclonal B-cellactivation, Cell. Immunol. 158 (1) (1994) 131–139.

[12] T. Marotti, et al., In vivo effect of progesteron and estrogen onthymus mass and T-cell functions in female mice, Horm.Metab. Res. 16 (4) (1984) 201–203.

[13] U. Islander, et al., Estrogens in rheumatoid arthritis; the immunesystem and bone, Mol. Cell. Endocrinol. 335 (1) (2011) 14–29.

48 A. Stubelius et al.

[14] R.H. Straub, The complex role of estrogens in inflammation,Endocr. Rev. 28 (5) (2007) 521–574.

[15] J.D. Sibonga, et al., Dose–response effects of 2-methoxyestradiolon estrogen target tissues in the ovariectomized rat, Endocrinol-ogy 144 (3) (2003) 785–792.

[16] K. Kanasaki, et al., Deficiency in catechol-O-methyltransferaseand 2-methoxyoestradiol is associated with preeclampsia,Nature 453 (7198) (2008) 1117–1121.

[17] A. Perez-Sepulveda, et al., Low 2-methoxyestradiol levels at thefirst trimester of pregnancy are associated with the developmentof preeclampsia, Prenat. Diagn. 32 (11) (2012) 1053–1058.

[18] S.O. Jobe, C.T. Tyler, R.R. Magness, Aberrant synthesis,metabolism, and plasma accumulation of circulating estrogensand estrogen metabolites in preeclampsia implications forvascular dysfunction, Hypertension 61 (2) (2013) 480–487.

[19] A. Stubelius, et al., Sexual dimorphisms in the immune system ofcatechol-O-methyltransferase knockout mice, Immunobiology217 (8) (2012) 751–760.

[20] T.M. LaVallee, et al., 2-Methoxyestradiol inhibits proliferationand induces apoptosis independently of estrogen receptorsalpha and beta, Cancer Res. 62 (13) (2002) 3691–3697.

[21] T.E. Sutherland, et al., 2-Methoxyestradiol is an estrogen receptoragonist that supports tumor growth in murine xenograft models ofbreast cancer, Clin. Cancer Res. 11 (5) (2005) 1722–1732.

[22] E. Josefsson, A. Tarkowski, Suppression of type II collagen-induced arthritis by the endogenous estrogen metabolite 2-methoxyestradiol, Arthritis Rheum. 40 (1) (1997) 154–163.

[23] M.K. Lindberg, et al., Estrogen receptor specificity for theeffects of estrogen in ovariectomized mice, J. Endocrinol. 174(2) (2002) 167–178.

[24] J.G. Lemmen, et al., Tissue- and time-dependent estrogenreceptor activation in estrogen reporter mice, J. Mol. Endocrinol.32 (3) (2004) 689–701.

[25] J. Bourghardt, et al., The endogenous estradiol metabolite 2-methoxyestradiol reduces atherosclerotic lesion formation infemale apolipoprotein E-deficient mice, Endocrinology 148 (9)(2007) 4128–4132.

[26] Y. Hayakawa, M.J. Smyth, CD27 dissects mature NK cells intotwo subsets with distinct responsiveness and migratorycapacity, J. Immunol. 176 (3) (2006) 1517–1524.

[27] J.C. Sun, J.N. Beilke, L.L. Lanier, Adaptive immune features ofnatural killer cells, Nature 457 (7229) (2009) 557–561.

[28] A.M. Brzozowski, et al., Molecular basis of agonism andantagonism in the oestrogen receptor, Nature 389 (6652)(1997) 753–758.

[29] Y.L. Wu, et al., Structural basis for an unexpected modeof SERM-mediated ER antagonism, Mol. Cell 18 (4) (2005)413–424.

[30] M. Schmidt, et al., Estrone/17beta-estradiol conversion to,and tumor necrosis factor inhibition by, estrogen metabolitesin synovial cells of patients with rheumatoid arthritis andpatients with osteoarthritis, Arthritis Rheum. 60 (10) (2009)2913–2922.

[31] S.P. Tofovic, et al., Estradiol metabolites attenuate renal andcardiovascular injury induced by chronic nitric oxide synthaseinhibition, J. Cardiovasc. Pharmacol. 46 (1) (2005) 25–35.

[32] P. Kohen, et al., 2-Methoxyestradiol in the human corpusluteum throughout the luteal phase and its influence on luteincell steroidogenesis and angiogenic activity, Fertil. Steril. 100(5) (2013) 1397–1404 (e1).

[33] J.E. Rossouw, et al., Risks and benefits of estrogen plusprogestin in healthy postmenopausal women: principal resultsfrom the Women's Health Initiative randomized controlledtrial, JAMA 288 (3) (2002) 321–333.

[34] B.L. Riggs, L.C. Hartmann, Selective estrogen-receptor modu-lators — mechanisms of action and application to clinicalpractice, N. Engl. J. Med. 348 (7) (2003) 618–629.

[35] M.C. Erlandsson, et al., Raloxifene- and estradiol-mediatedeffects on uterus, bone and B lymphocytes in mice, J.Endocrinol. 175 (2) (2002) 319–327.

[36] C.R. Ireson, et al., Pharmacokinetics and efficacy of 2-methoxyoestradiol and 2-methoxyoestradiol-bis-sulphamatein vivo in rodents, Br. J. Cancer 90 (4) (2004) 932–937.

[37] P.S. Crooke, et al., Estrogens, enzyme variants, and breastcancer: a risk model, Cancer Epidemiol. Biomarkers Prev. 15(9) (2006) 1620–1629.

[38] S. Dawling, N. Roodi, F.F. Parl, Methoxyestrogens exertfeedback inhibition on cytochrome P450 1A1 and 1B1, CancerRes. 63 (12) (2003) 3127–3132.

[39] B.T. Zhu, et al., Metabolic deglucuronidation and demethyla-tion of estrogen conjugates as a source of parent estrogens andcatecholestrogen metabolites in Syrian hamster kidney, atarget organ of estrogen-induced tumorigenesis, Toxicol.Appl. Pharmacol. 136 (1) (1996) 186–193.