Pharmacokinetics of a CCR5 inhibitor in rhesus macaques following vaginal, rectal and oral application

Abstract

This study measured and compared the pharmacokinetics of CMPD167, a small molecule antiretroviral CCR5 inhibitor with potential as an HIV microbicide, following vaginal, rectal and oral administration in rhesus macaques.
A vaginal hydroxyethylcellulose (HEC) gel, a rectal HEC gel, a silicone elastomer matrix-type vaginal ring and an oral solution, each containing CMPD167, were prepared and administered to rhesus macaques pretreated with Depo-Provera. CMPD167 concentrations in vaginal fluid, vaginal tissue (ring only), rectal fluid and blood plasma were quantified by HPLC-mass spectrometry.
CMPD167 concentrations measured in rectal fluid, vaginal fluid and blood plasma were highly dependent on both the route of administration and the formulation type. Although rectal and vaginal fluid concentrations were highest when CMPD167 was administered locally (via either gel or ring), lower concentrations of the drug were also measured in these compartments following administration at the remote mucosal site or orally. CMPD167 levels in the vaginal and rectal fluid following oral administration were relatively low compared with local administration.
The study provides clear evidence for vaginal-rectal and rectal-vaginal drug transfer pathways and suggests that oral pre-exposure prophylaxis with CMPD167 may be less efficacious at preventing sexual transmission of HIV-1 than topically applied products.

Full text

Pharmacokinetics of a CCR5 inhibitor in rhesus macaques following
vaginal, rectal and oral application
R. Karl Malcolm
1
*, Deborah Lowry
2
, Peter Boyd
1
, Leslie Geer
3
, Ronald S. Veazey
4
, Laurie Goldman
3
, P. J. Klasse
5
,
Robin J. Shattock
6
and John P. Moore
6
1
School of Pharmacy, Queen’s University Belfast, Belfast BT9 7BL, UK;
2
School of Life and Health Sciences, Aston University, Birmingham
B4 7ET, UK;
3
Particle Sciences, Bethlehem, PA, USA;
4
Tulane National Primate Research Center, Tulane University Health Sciences Center,
Covington, LA 70433, USA;
5
Department of Microbiology and Immunology, Weill Cornell Medical College, New York, NY 10021, USA;
6
Imperial College London, London, UK
*Corresponding author. Tel: +44-(0)28-9097-2319; E-mail: k.malcolm@qub.ac.uk
Received 9 August 2013; returned 2 October 2013; revised 12 November 2013; accepted 4 December 2013
Objectives: This study measured and compared the pharmacokinetics of CMPD167, a small molecule antiretro-
viral CCR5 inhibitor with potential as an HIV microbicide, following vaginal, rectal and oral administration in rhe-
sus macaques.
Methods: A vaginal hydroxyethylcellulose (HEC) gel, a rectal HEC gel, a silicone elastomer matrix-type vaginal ring
and an oral solution, each containing CMPD167, were prepared and administered to rhesus macaques pretreated
with Depo-Provera. CMPD167 concentrations in vaginal fluid, vaginal tissue (ring only), rectal fluid and blood
plasma were quantified by HPLC–mass spectrometry.
Results: CMPD167 concentrations measured in rectal fluid, vaginal fluid and blood plasma were highly depend-
ent on both the route of administration and the formulation type. Although rectal and vaginal fluid concentra-
tions were highest when CMPD167 was administered locally (via either gel or ring), lower concentrations of the
drug were also measured in these compartments following administration at the remote mucosal site or orally.
CMPD167 levels in the vaginal and rectal fluid following oral administration were relatively low compared with
local administration.
Conclusions: The study provides clear evidence for vaginal–rectal and rectal–vaginal drug transfer pathways and
suggests that oral pre-exposure prophylaxis with CMPD167 may be less efficacious at preventing sexual trans-
mission of HIV-1 than topically applied products.
Keywords: HIV microbicides, vaginal HEC gels, vaginal rings, rectal gels, pre-exposure prophylaxis, PrEP, CMPD167
Introduc tion
The development of topically applied microbicide formulations
able to reduce the incidence of sexually acquired HIV-1 infection
remains a priority within the prevention science field.
1,2
The signifi-
cant, albeit incomplete, protection provided by the vaginal admin-
istration of a water-based gel containing the reverse transcriptase
inhibitor tenofovir (CAPRISA 004 trial) illustrates the potential of
this app roach.
3
A major challenge for microbicide development
is to increase the degree of protection seen in that trial. Possible
improvements could come from the use of a different antiretro-
viral (ARV)
4
or a combination of ARVs with different mechanisms
of action,
5
or by applying longer-lasting gels
6,7
and/or sustained
release devices such as vaginal rings.
8 – 10
Rectal del ivery of
microbicides may also help protect both women and men against
this route of sexual transmission. Recently, there has been consid-
erable interest in administering ARVs orally to achieve the same
goals [i.e. oral pre-exposure prophylaxis (PrEP)].
11,12
Correlating efficacy data (from animal or human studies) with
post-application ARV concentrations in relevant biological fluids
and tissues is critical to understanding the drug levels required
for protection and to guide improvements t o the formulation.
In pharmacokinetic (PK) studies of topically (vaginal or rectal )
applied microbicide formulations, it is particularly important to
measureARVconcentrationsinthetissueandfluidcom-
partments at the local site of product administration, since
vaginal microbicides are primarily intended to prevent vaginal
HIV-1 trans mission and rectal microbicides rectal tran smission.
# The Author 2013. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved.
For Permissions, please e-mail: journals.permissions@oup.com
J Antimicrob Chemother 2014; 69: 1325–1329
doi:10.1093/jac/dkt506 Advance Access publication 30 December 2013
1325
by guest on October 29, 2015http://jac.oxfordjournals.org/Downloaded from

However, oral PrEP aims to deliver ARVs to both the vaginal
and rectal compartments via the systemic circulation.
12
Moreover, it is also possible that administering an ARV formula-
tion to the vagin a might protect against rectal transmission (and
vice versa). Hence, we felt it would be useful to extend the meas-
urement of drug concentrations to include biological compart-
ments beyond t he site of product application. Therefore, we
carried out a series of PK experiments in rhesus macaques
using the CCR5-targeted entry inhibitor C MPD167. This com-
pound, from the same general class as the licensed drug mara-
viroc, provides substantial protection against vag inal challenge
of macaques when delivered as a water-based vaginal gel or
as oral PrEP.
13,14
Methods
The International Partnership for Microbicides supplied CMPD167.
Macaque studies were performed at the Tulane National Primate
Research Cente r in accordance with recommendations in the Guide for
the Care and Use of Laboratory Animals of the NIH and following
approval from the Tulane University Institutional Animal Care and
Use Committee. Each macaque (n ¼ 24, in four groups of six animals)
received a single 30 mg intramuscular injection of Depo-Provera
30 days prior to administrat ion of CMPD167, to synchronize their men-
strual cycles and thin the vaginal mucosa. Three different CMPD167
formulations were tested: a 3 mL volume of a 2.2% w/w hydroxyethylcel-
lulose (HEC; grade HHX) gel containing 5 mM (3mg/mL;9mgtotal
dose) CMPD167 was admini stered either vaginally or rectally; a mat rix-
type silicone elastomer ring (overall diameter 25.0 mm, cross-sectional
diameter 6.0 mm) containing 400 mg CMPD167 was inserted vaginally
and removed after 28 days continuous use; and an aqueous CMPD167
tartrate buffer solution was given orally by gavage (20 mg/kg dose;
mean macaque weight 7.0 kg; weight range 5.5–10.2 kg; CMPD167
oral dose range 110–204 mg). The ring and gels were prepared as
described previously.
8,15
Vagi nal fluid (Weck-Cel inserted fol lowing
wiping of the vaginal surface with moistened gauze) , vaginal tissue
(pinch biopsy) and blood were sampled at various times. Tissue biopsies
wereplacedin1mLofdistilledwaterandfrozenwithin30minof
sampling. CMPD167 concentrations were quantified using gradient
reverse-phase HPLC–mass spectrometry according to previously
described methods.
8
Rectal fluid was similarly sampled by Weck-
Cel and quantified using the following HPLC– mass spectrometry
method. Internal standard (D5-CMPD167) was added to 150 mL of rectal
fluid and proteins were precipitated by the addition of 450 mLofaceto-
nitrile. After vortex mixing and centrifugation (3000 rpm, 10 min, 208C), a
400 mL aliquot of the supernatant was evaporated to dryness under
nitrogen at 408C. Residues were reconstituted in 20:80 (v:v)
methanol:10 mM ammonium formate ( pH 3.5) for analysis. Analysis
was performed using a Shimadzu Prominence
w
HPLC system ( Kyot o,
Japan),aBDSHypersilC8column(50×2.1 mm, 5 mm; Thermo
Scientific, Walt ham, MA, USA) and an API3200
w
(AB Sciex,
Framingham, MA, USA) triple quadrupole mass spectrometer. The
API3200
w
mass spectrometer was used in positive TurboIonSpray
w
mode with a source temperature of 6508C. The mobile phase consisted
of 5 0 :50 10 mM ammonium format e (pH 3.5):methanol with 0.1% for-
mic acid, a 50%–95% organic gradient and a 0.5 mL/min flow rate.
CMPD167 and the internal standard were detected using mul tiple reac-
tion monitoring and the precursor product ion transitions were m/z
575.5 444.3 for CMPD167 and 580.6 449. 4 for D5-CMPD167. The lin-
ear range was 0.5–1000 ng/mL.
Where appropriate, data were statistically analysed using either a
Mann–Whitney U-test or one-way ANOVA followed by post hoc analysis
using the Tukey – Kramer multiple comparison s test. In all cases, a
P value of ,0.05 was considered significant. Analysis was conducted
using GraphPad Prism.
Results and discussion
Vaginal fluid concentrations
CMPD167 concentrations in vaginal fluid were highest in the vagi-
nal gel group; they peaked at 10
6
ng/mL 15 min after gel appli-
cation and steadily decreased to 2.8×10
4
ng/mL by 24 h
(Figure 1a). The matrix-type vaginal ring device provided a similar
mean CMPD167 vaginal fluid concentration (2.8×10
4
ng/mL) at
the 1 h timepoint (the earliest sampled in this group) as the vagi-
nal gel at 24 h and comparable concentrations (range: 1.6×10
4
–
2.3×10
5
ng/mL) were then sustained out to 672 h (28 days; final
sampling timepoint). Rectal application of the same HEC gel also
resulted in relatively high vaginal fluid CMPD167 concentrations
(ranging from 2.2×10
5
at 1 h to 3. 6 ×10
4
ng/mL at 24 h;
Figure 1a), with values consistently, but not always significantly,
lower than those obtained with the vaginal gel at each sampling
timepoint. The vaginal gel AUC was 2.7-fold higher than that for
the rectal gel (Table 1). Hence, there must be a transfer or diffusion
of the ARV from the rectal to the vaginal compartment. Although
this drug transfer mechanism is expected to be time dependent,
the high variability in vaginal fluid concentrations following rectal
gel application resulted in a T
max
value that was not statistically
different (P¼ 0.18) from that for the vaginal gel (Table 1). Oral
administration of a CMPD167 solution yielded only relatively low
vaginal fluid concentrations that ranged from 1.3×10
3
to
4.6×10
3
ng/mL over the 24 h period. The time taken for vaginal
fluid concentr ations to reach peak levels (T
max
) was dependent on
the route of administr a tion and the type of formulation: vaginal gel
(0.4 h), rectal gel (4.1 h), oral solution (9.4 h) ,vaginal ring (24 h)
(Table 1). However, the rank order of decreasing C
max
values was
vaginal gel. vaginal ring. rectal gel. oral solution (Table 1).
Rectal fluid concentrations
As expected, the highest concentrations of CMPD167 in rectal
fluid were measured in animals receiving the rectal gel, with
mean CMPD167 levels ranging from 2.9×10
4
ng/mL (8 h) to
2.0×10
5
ng/mL (15 min) (Figure 1c). In the other groups, rectal
CMPD167 concentrations were lower, in the following rank
order: vaginal gel (range: 1.4–6.7×10
3
ng/mL). oral solution
(6.1×10
1
–3.1×10
3
ng/mL). vaginal ring (1.4×10
1
–1.9×10
3
ng/
mL). Following rec tal gel administration, the rectal and vaginal
fluid concentrations were similar (compare Figure 1cand
Figure 1a), prod ucing rectal and vaginal AUC values of 1 .4×10
6
and 9.9×10
5
ng
.
h/mL, respectively (Table 1). These data suggest
there is an efficient drug transport pathway from the rectal to the
vaginal compartment, more so than applies in the converse direc-
tion (see Figure 1c). With the vaginal ring device, rectal CMPD167
levels were maintained over the 14 day study period (Figure 1c),
bu t were two t o t hree orders of magnitude lower than vaginal
fluid concen trations (Figure 1a). The relatively long T
max
valu e
for CMPD167 concentrations in rectal fluids following oral admin-
istration (24 h; Table 1) suggests that C
max
is largely determined
by gastrointestinal transit rather than diffusion from the systemic
compartment, with implications for oral dosing studies measuring
rectal tissue concentrations.
Malcolm et al.
1326
by guest on October 29, 2015http://jac.oxfordjournals.org/Downloaded from

Plasma concentrations
Plasma levels of CMPD167 were, as expected, very much lower in
the animals given the vaginal gel, rectal gel and vaginal ring, com-
pared with oral administration (Figure 1dandTable1). Plasma
concentrations following oral and gel administration showed typ-
ical first-order elimination kinetics, while those following vaginal
ring placement decreased relatively slowly (7.8 ng/mL at 4 h to
3.8 ng/mL at 24 h and 0.7 ng/mL at 672 h).
Vaginal tissue concentrations
CMPD167 tissue levels, only measured in vaginal ring recipients,
were maintained in the range 2.29×10
3
–1.72×10
4
ng/g
(Figu re 1b; levels were onl y measured out to 14 days). We have
previously reported similar tissue levels in macaques with a
CMPD167 ring.
8
In that study, the PK parameters were highly
dependent on Depo-Provera administration to the animals,
which modestly reduced vaginal concentrations of CMPD167
while increasing transfer of the compound into the plasma (rectal
concentrations were not measured).
Conclusions
Several points arise from this study. First, CMPD167 is effectively
transferred from the rectum to the vagina (rectal gel) and from
the vagina to the rectum (vaginal gel and ring), although the for-
mer route appears to be the more efficient. A simple mechanism
accounting for the differential rate of drug transfer between the
compartments is not immediately obvious. Although the thinner
simple columnar epithelial tissue of the rectum should offer
increased drug absorption from that compartment, thus facilitat-
ing the rectal to vaginal (and rectal to blood) drug transfer path-
way, it is also likely to similarly effect drug diffusing in the other
direction (vaginal to rectal). It is also possible that the concentra-
tion of drug at the mucosal tissue (and, in turn, the concentration
gradient established across the tissue) is higher following rec tal
gel administration compared with vaginal gel administration,
due to differences in how the gel spreads to cover the tissue
and the different fluid volumes within each compartment.
Finally, it is also possible that the in tercompartmental transfer
does not take place directly via the connecting tissue, but instead
follows a mechanism similar to the first uterine pass effect
10
6
(a)
(b)
(c)
(d)
CMPD167 in vaginal fluid (ng/mL)
CMPD167 in rectal fluid (ng/mL)
CMPD167 in vaginal tissue (ng/mL)
CMPD167 in plasma (ng/mL)
10
5
10
4
Vaginal gel
Vaginal ring
Rectal gel
Vaginal ring
Oral solution
Vaginal gel
Rectal gel
Vaginal ring
Oral solution
Vaginal gel
Rectal gel
Vaginal ring
Oral solution
0
0
8
16
24
48
Time (h)
Time (h)
Time (h)
Time (h)
72
96
168
336
504
672
8
16
24
48
72
96
168
336
504
672
0
8
16
24
48
72
96
168
336
504
672
0
8
16
24
48
72
96
168
336
504
672
10
3
10
6
10
3
10
2
10
1
10
0
10
–1
10
5
10
4
10
3
10
2
10
1
10
6
10
5
10
4
10
3
10
2
Figure 1. Vaginal fluid (a), vaginal tissue (b), rectal fluid (c) and plasma (d) concentrations (ng/mL) of CMPD167 following administration of a vaginal HEC
gel, a rectal HEC gel, an oral solution and a silicone elastomer vaginal ring. Values are means+SD (n¼ 6).
CMPD167 pharmacokinetics
1327
JA
C
by guest on October 29, 2015http://jac.oxfordjournals.org/Downloaded from

observed for direct vaginal –uterine drug transfer and attributed
to the overlapping network of blood vessels associated with the
arteriovenous plexus.
16,17
Thinning of the epithelial tissue has previously been postulated
to increase transfer of CMPD167 from vag inal to systemi c com-
partments when macaques were pretreated with Depo-
Provera prior to ring administratio n.
8
Hence, it is possible that
rectal administration of an ARV microbicide formulation might
provide protection against both rectal and vaginal transmission,
a s upposition supported by a recent macaque study of rectally
and vaginally administered tenofovir gel.
18
The influence of
Depo-Provera on rectal absorption has not been reported.
Our second conclusion is that loc al concentrations of
CMPD167 following vaginal and rectal application are typically
one to t wo orders of magnitude greater than achieved by oral
dosing. Similar differences have also been observed in women
given vaginal gel or oral tablet formulations of tenofovir diso-
proxil fu marate.
19
Assuming that protection is only mediated
locally, these data suggest that it may be very diffi cult for orally
dosed ARVs to achieve the relatively high vaginal/rectal co ncen-
trations obtained with mucosal dosing. The implication is that
the efficacy of PrEP could be greater for vaginal/rectal dosing
than for oral delivery, assuming similar adherence rates.
Recently, weekly oral administration of two doses of maraviroc
to m acaques 24 h before and 2 h after rectal challenge provided
negligible protection, despite concentrations in rectal fluid reach-
ing as high as 10
5
ng/mL.
12
A more sustai ned, multiday regimen
of orally delivered CMPD167 was, however, significantly protec-
tive against vaginal challenge, suggesting that oral PrEP may
not be limited to local activity.
14
However, v aginal concentra-
tions of CMPD167 were not m easured in that study and the con-
centrations required for protection are unknown.
We also conclude that concentrations of CMPD167 in vagi-
nal fluid during the first few hours after administration were
greater for the vaginal gel compared with the ring, but they
then decreased to below the more sustai ned concentrations pro-
vided by the ring. The critical unknown is whether the vaginal
concentrations obtained with the ring are sufficient for protec-
tion. An answer would resolve whether the short-term h igh
bolus effec t of ge ls is m ore im portant than the longer PK tail
seen with rings.
Additional PK and challenge studies in the macaque that com-
pared various modes of ARV delivery, including, but not limited to,
CCR5 inhibitors, would provide useful informatio n on the short-
and longer-term concentration requirements for protection from
rectal and/or vaginal challenge. At present, mostly only inferences
can be made by extrapolating across studies that yield partial
datasets.
Funding
This work was supported by the National Institutes of Health (grant num-
ber U19 AI076982).
Transparency declarations
None to declare.
Table 1. Pharmacokinetic parameters for CMPD167 concentrations measured in macaque rectal fluid, vaginal fluid and blood plasma following administration of vaginal gel, rectal gel, vaginal ring
and oral solution
Formulation
type
CMPD167
dose/ring
loading
Rectal fluid Vaginal fluid Blood plasma
C
max
(ng/mL) T
max
(h) AUC
0–24
(ng
.
h/mL) C
max
(ng/mL) T
max
(h) AUC
0–24
(ng
.
h/mL) C
max
(ng/mL) T
max
(h)
AUC
0–24
(ng
.
h/mL)
Vaginal gel 9 mg 6995+9652 4.2+2.2 58420+68700 915680+267 810 0.4+0.3 2705 000+617500 55.98+30.00 2.5+1.6 447.4+495.8
Rectal gel 9 mg 207700+107900 0.25+0.0 1434000+1 732000 220200+347200 4.1+4.2 994600+2016 000 144.0+247.5 0.9+0.3 370.0+489.2
Vaginal ring 400 mg 453+456 18.0+9.4 5084+4410 225200+285900 24.0+0.0 2282000+ 2 410000 7.762+2.500 4.0+0.0 140.0+37.6
Oral solution 110–204 mg
(20 mg/kg)
a
3087+2865 24.0+0.0 33330+33100 7920+8745 9.4+11.4 64420+37970 841.3+577.5 1.8+1.8 4341+4042
AUC
0–24
, area under the plasma concentration–time curve from 0 to 24 h; C
max
, maximum concentration; T
max
, time to reach the C
max.
Values are means+SD (n¼ 6).
a
Total administered oral dose was dependent on macaque weight (see the Methods section).
Malcolm et al.
1328
by guest on October 29, 2015http://jac.oxfordjournals.org/Downloaded from

References
1 Friend DR, Kiser PF. Assessment of topical microbicides to prevent HIV-1
transmission: concepts, testing, lessons learned. Antiviral Res 2013; 99:
391–400.
2 Shattock RJ, Rosenberg Z. Microbicides: topical prevention against HIV.
Cold Spring Harb Perspect Med 2012; 2: a007385.
3 Karim Q A, Karim SSA, Frohlich JA et al. Effectiveness and safety of
tenof ovir gel, an antiretroviral microbicide, for the prevention of HIV
infection in women. Science 2010; 329: 1168–74.
4 Klasse PJ, Shattock R, Moore KP. Antiretroviral drug-based microbicides
to prevent HIV-1 sexual transmission. Ann Rev Med 2008; 59: 455–71.
5 Balzarini J, Schols D. Combination of antiretroviral drugs as microbicides.
Curr HIV Res 2012; 10: 53–60.
6 Forbes CJ, Lowry D, Geer L et al. Non-aqueous silicone elastomer gels as
a vaginal microbicide delivery system for the HIV-1 entry inhibitor
maraviroc. J Control Release 2011; 156: 161–9.
7 CurranRM,DonnellyL,MorrowRJet al. Vaginal de livery of the
recombinant HIV-1 clade-C trimeric gp140 envelope protein CN54gp140
within novel rheolog ically structured vehicles el icits specific immune
responses. Vaccine 2009; 27: 6791–8.
8 Malcolm RK, Veazey RS, G eer L et al.SustainedreleaseoftheCCR5
inhibitors CMPD167 and maraviroc from vaginal rings in rhesus
macaques. Antimicrob Agents Chemother 2012; 56: 2251–8.
9 Nel A, Smythe S, Young K et al. Safety and pharmacokinetics of dapivirine
delivery from matrix and reservoir intravaginal rings to HIV-negative
women. J Acquir Immune Defic Syndr 2009; 51: 416–23.
10 Malcolm RK, Edwards K-L, Kiser P et al. Advances in microbicide vaginal
rings. Antiviral Res 2010; 88 Suppl 1: S30– 9.
11 Grant RM, Lama JR, Anderson PL et al. Preexposure chemoprophylaxis
for HIV prevention in men who have sex with men. N Engl J Med 2010; 363:
2587–99.
12 Massud I, Aung W, Martin A et al. Lack of prophylactic efficacy of oral
maraviroc in macaques despite high drug concentrations in rectal tissues.
J
Virol 2013; 87: 8952–61.
13 Veazey RS, Klasse PJ, Schader SM et al. Protection of macaques from
vaginal SHIV chal lenge by vaginally delivered inhibitors of virus–cell
fusion. Nature 2005; 438; 99–102.
14 Veazey RS, Springer MS, Marx PA et al. Protection of macaqu es from
vaginal SHIV challen ge by an orally delivered CCR5 inhibitor. Nat Med
2005; 11: 1293– 4.
15 Malcolm RK, Forbes CJ, Geer L et al. Pharmacokinetics and efficacy of a
vaginally ad ministered maraviroc gel in rhesus macaques. J Antimicrob
Chemother 2013; 68: 678–83.
16 Bulletti C, de Ziegler D, Flamigni C et al.Targeteddrugdeliveryin
gynaecology: the first uterine pass effect. Hum Reprod 1 997; 12:
1073–9.
17 Cicinelli E, de Ziegler D, Morgese S et al. ‘First uterine pass effec t’ is
observed when estradiol is placed in the upper but not lower third of the
vagina. Fertil Steril 2004; 81: 1414–6.
18 NuttallJ,KashubaA,WangRet al. Pharmacokinetics of
tenofovir following intravaginal and intrarectal administration of
tenofovir gel to rhesus macaques. Antimicrob Agents Chemother
2012; 56:103–9.
19 Hendrix CW, Chen BA, Guddera V et al. MTN-001: randomized
pharmacokineti c cross-over study comparing tenofovir vaginal gel and
oral table ts in va ginal tissue and other compartments. PLoS One 2013;
8: e55013.
CMPD167 pharmacokinetics
1329
JA
C
by guest on October 29, 2015http://jac.oxfordjournals.org/Downloaded from