differential proteomic analysis reveals increased ... · differential proteomic analysis reveals...
TRANSCRIPT
75Vet. Res. 37 (2006) 75–87© INRA, EDP Sciences, 2005DOI: 10.1051/vetres:2005043
Original article
Differential proteomic analysis reveals increased cathelicidin expression in porcine
bronchoalveolar lavage fluid after an Actinobacillus pleuropneumoniae infection
Isabel HENNIG-PAUKAa,b, Ilse JACOBSENa, Frank BLECHAc, Karl-Heinz WALDMANNb, Gerald-Friedrich GERLACHa*
a Institute for Microbiology, Department of Infectious Diseases, University of Veterinary Medicine Hannover, Foundation, Hannover, Germany
b Clinic for Swine and Small Ruminants, University of Veterinary Medicine Hannover, Foundation, Hannover, Germany
c Department of Anatomy and Physiology, College of Veterinary Medicine, Kansas State University, Manhattan, Kansas, USA
(Received 23 December 2004; accepted 23 June 2005)
Abstract – Accurate definition of respiratory health in pigs is an important problem for swineproducers and veterinarians. In an approach to identify potential biomarkers, two-dimensional gelelectrophoresis and mass spectrometry on bronchoalveolar lavage fluid (BALF)-derived proteinsfrom pigs experimentally infected with Actinobacillus pleuropneumoniae were performed atdifferent time points post infection. Mock-infected pigs were used as a control. It was shown thatthe antimicrobial peptides, prophenin-2 and PR-39, and the calcium-binding protein calgranulin Cwere reproducibly upregulated in BALF of pigs chronically infected with A. pleuropneumoniae.Concentrations of PR-39 were significantly (p < 0.05) increased in BALF (median of 4.8 nM) butnot in serum (median of 2.5 nM) on day 21 after infection. A Receiver Operating Characteristics(ROC) plot showed that PR-39 in BALF is an accurate and easily accessible marker to detect clinicallyhealthy pigs convalescent from an experimental A. pleuropneumoniae infection. These results implythat PR-39 might have a potential as a general biomarker to determine porcine respiratory health.
antimicrobial peptide / PR-39 / BALF / porcine respiratory tract
1. INTRODUCTION
Detection of respiratory diseases in pigsis of growing importance. Due to changesin liability laws, swine producers may beheld responsible for disease-related lossesunless it can be determined that the animalswere healthy at the time they were sold.
Because several respiratory tract patho-gens, such as Mycoplasma hyopneumoniae,Haemophilus parasuis, Bordetella bronchi-septica, Pasteurella multocida, and Actino-bacillus (A.) pleuropneumoniae, can beresponsible for respiratory disease in pigsand can persist for extended periods of timein convalescent animals, tests facilitating
* Corresponding author: [email protected]
Article published by EDP Sciences and available at http://www.edpsciences.org/vetres or http://dx.doi.org/10.1051/vetres:2005043
76 I. Hennig-Pauka et al.
the detection of lung alterations independ-ent of the pathogen would be highly rele-vant.
Clinically healthy individuals convales-cent from respiratory disease are likely tostill exhibit alterations in the respiratorytract for a prolonged period of time. Thesealterations are reflected in changes withinthe epithelial lining fluid (ELF; [14]), a thinlayer covering the surface of the respiratorytract and presenting the first line of hostdefence against inhaled microbial invaders.Modifications in ELF have been exploitedin humans to diagnose lung diseases basedon measurements of phospholipid content,surface tension, cytological findings andprotein profiles [25]. However, approxi-mately 97% of ELF-proteins are serum-derived [19] and, therefore, may be of lim-ited use as a diagnostic marker for respira-tory infections.
The use of proteomic technologies hasintensified the search for lung disease mark-ers in human ELF [25], and currently a listof approximately 1000 proteins identifiedin human bronchoalveolar lavage fluid(BALF) is available (http://w3.umh.ac.be/~biochim/proteomic.htm; [26, 36]). Further-more, it was found that a broad spectrum ofantimicrobial agents such as defensins,inducible nitric oxide synthetase and LL37/CAP-18, a member of the cathelicidin fam-ily, are found in ELF [10], which, in part,are present in increased concentrations dur-ing infection [29]. Most of these agentsoriginate from emigrated phagocytes wheresome of them are stored as precursors incytoplasmic granules [38]. Although por-cine neutrophils lack defensins they arewell endowed with cathelicidins, includingprotegrins, prophenins, PR-39 [28] and por-cine myeloid antimicrobial peptides [31].
The definition of respiratory health inswine is limited by the difficulty of inter-preting data collected from the respiratorytract. In addition to genetically determinedresistance or susceptibility, many non-path-ogenic and facultative pathogenic microor-ganisms as well as abiotic factors influence
host immunity. Pathogenic microorgan-isms persisting on the respiratory epithe-lium of clinically healthy pigs have a higheconomic impact on swine production andcan be spread into previously uninfectedherds by carrier animals.
Our aim was to identify new diagnostictools for the definition of respiratory healthin swine, which are not restricted to bacte-riological examinations of BALF. To iden-tify suitable marker proteins in BALF weused an A. pleuropneumoniae infectionmodel. Using two-dimensional gel electro-phoresis and subsequent mass spectrome-try, we detected reproducible changes inprotein profiles between uninfected andconvalescent animals 3-weeks post infec-tion. Our findings identified three peptidesthat were up-regulated in convalescent ani-mals and showed that one of them, the anti-microbial peptide PR-39, has a potential asa diagnostic marker to detect lung diseasein convalescent pigs.
2. MATERIALS AND METHODS
2.1. Bacterial strains and media
Actinobacillus pleuropneumoniae sero-type 7 clinical isolate AP76 [1] and A. pleu-ropneumoniae serotype 2 clinical isolateC5934 [34] were used for challenge exper-iments. Both strains were cultured and pre-pared for experimental aerosol infectionstudies as described previously [3].
2.2. Clinical study and lung lavage protocol
A total of 48 specific-pathogen-free pigsof both sexes, approximately 7- to 9-weeksold, were obtained from closed herds witha long record of respiratory health. All ani-mals were tested for the absence of A. pleu-ropneumoniae antibodies using an ApxIIA-ELISA [24]. The animals were experimen-tally infected with an A. pleuropneumoniaeserotype 7 (32 pigs) or a serotype 2 clinicalisolate (8 pigs) in an aerosol infection model
Antimicrobial peptides in porcine BALF 77
[3]; eight pigs were used for a mock-infec-tion. The challenge dose was titrated toinduce infection but not fatal disease (102
to 103 bacteria per liter aerosol in the chal-lenge chamber). BALF was collected asdescribed previously [3] from all pigs priorto infection, from 28 pigs in the acute stageof infection (6 pigs infected with A. pleu-ropneumoniae serotype 2 on day 4 and22 pigs infected with A. pleuropneumoniaeserotype 7 on day 7 post infection), andfrom all 37 surviving pigs before necropsyon day 21 post infection. BALF was takenfrom control animals before and on day 7and day 21 after mock-infection.
BALF samples were processed immedi-ately for bacterial culture and for cytologi-cal and electrophoretical examinations.BALF cells were counted and cytospotswere prepared and stained with 10% May-Grünwald/Giemsa solution for each differ-ential count; 8% polymorphonuclear neu-trophils (PMN) in the differential cell countwas considered the cut-off value for healthypigs [13]. Pigs with a higher cell count inthe initial BALF were not included in thestudy (4 animals infected and 2 animalsfrom the control group); two pigs not show-ing any pathomorphological lung altera-tions in combination with constant negativebacteriological results were excluded. Post-mortem analyses as well as bacteriologicaland serological examinations of all pigswere performed as previously described [3,17]. All animals were cared for in accord-ance with the principles outlined in theEuropean Convention for the Protection ofVertebrate Animals Used for Experimentaland Other Scientific Purposes (ETS123).
2.3. Selection of BALF-samples for two-dimensional gel electrophoresis
BALF samples from a subgroup of tenpigs (six pigs infected with A. pleuropneu-moniae serotype 7 and four infected withA. pleuropneumoniae serotype 2) wereselected for two-dimensional polyacryla-mide gel electrophoresis (2-D PAGE). Allpigs in the subgroup showed clinical signs
such as depression and increased body tem-peratures after infection. Upon necropsy onday 21, all of these pigs had lung lesions,which were evaluated as previously described[17]. Briefly, each lung lobe was assesseda total possible lesion score of 5, resultingin a maximum score of 35 for the entirelung. Individual lesions were mapped on asimplified lung chart in which each lobewas subdivided into triangles. The numberof affected triangles was counted and thescore was calculated as a fraction of five forthis lobe. All pigs in this study had lunglesion scores between 0.1 and 13.4. A. pleu-ropneumoniae was reisolated from pneu-monic lesions, tonsils or lung lymph nodesin 9 of the 10 pigs.
BALF samples (taken before infection,4 days (A. pleuropneumoniae serotype 2) or7 days (A. pleuropneumoniae serotype 7) and21 days after infection) were analysed in trip-licate by 2-D PAGE. Briefly, BALF cellswere removed by centrifugation (5 000 × g,10 min), and proteins in the supernatantwere precipitated overnight at 4 °C withtrichloric acetic acid (TCA; 10% w/v finalconcentration) and harvested by centrifuga-tion (5 000 × g, 10 min). The pellet was dis-persed by sonication in ice-cold acetone,frozen for 2 h at –20 °C, and centrifugedagain as described above. The pellets wereair-dried and resuspended in rehydrationsolution (8 M urea, 2% CHAPS). Proteinconcentrations were determined (MicroBCA®, Uptima Interchim, France) andadjusted to 275 µg/mL in rehydration solu-tion supplemented with dithiothreithol (DTT;18 mM final concentration), IPG-buffer(pH 3–10, 0.5%; Amersham Biosciences,Uppsala, Sweden) and bromphenol blue(0.001%). Each sample (360 µL) was loadedon an Immobiline Dry Strip® (pH 3–10,18 cm; Amersham Biosciences) and run ina Pharmacia Biotech IPGphor system (11 hrehydration followed by a multistep pro-gram of 2 h at 100 V, 1 h at 500 V, 1 h at1000 V, 2 h at 4000 V, a linear gradientspanning from 4000 V to 8000 V in 2 h, anda final step at 8000 V for 6 h). After isoe-lectric focussing, the strips were stored at
78 I. Hennig-Pauka et al.
–20 °C or processed directly for the seconddimension. Briefly, the strips were placedin an equilibration buffer (50 mM Tris-HCl,pH 8.8, 6 M urea, 30% glycerol, 2% SDS,65 mM DTT) for 15 min followed by 15 minin an equilibration buffer containing iodoa-cetamide (244 mM) instead of DTT. Sepa-ration in the second dimension was carriedout on sodium dodecyl sulphate (SDS)PAGE (12.7% acrylamide, 0.3% bisacryla-mide). Gels were stained with colloidalCoomassie G-250 and with silver nitrate(Silver Staining Kit, Sigma, St. Louis, MO,USA) according to the manufacturer’sinstructions. Spot detection was performedusing the 2-D PAGE software (Phoretix®,Nonlinear Dynamics, Newcastle, UK).
2.4. Identification of differentially expressed proteins by mass spectrometry
Prominent spots expressed in BALF fromday 21 post infection samples were excisedfrom gels, digested with trypsin and the result-ing peptides were analysed by mass spec-trometry (AgeLab Pharma GmbH, Hamburg,Germany). Peptide mass fingerprint analy-ses were performed using the MASCOTprogram (www.matrixscience.com), andsignificant matches were identified usingthe SwissProt database. Scores were calcu-lated by the molecular weight search scorealgorithm (MOWSE) and are significant(p < 0.05) if greater than 69. Thus, a searchresult would be expected to occur at randomwith a frequency of less than 5% [27].
2.5. Western Blot analysis
BALF proteins were separated by dis-continuous SDS-PAGE (10.8% acrylamideand 0.29% bisacrylamide) and blotted ontoa nitrocellulose membrane using a Protean IIMinigel system (Bio-Rad, Munich, Ger-many) following standard procedures [30].Mouse monoclonal antibodies (mouse mAb)against PR-39, which recognise both maturePR-39 and its precursor (cathelin-contain-ing PR-39; [32, 33, 39]), were used as pri-
mary antibodies. Blots were developed usinga peroxidase-conjugate, and a chemilumi-nescent substrate (Super Signal Ultra®, Pierce,Illinois, USA).
2.6. Influence of PR-39 on bacterial growth and antigen expression
In order to investigate whether PR-39was bactericidal against A. pleuropneumo-niae, a growth inhibition zone assay onsolid medium was performed essentially asdescribed by Boman et al. [5]. Briefly,100 µL of A. pleuropneumoniae AP76 sus-pension with an optical density at 660 nm(OD660) of 0.1 in NaCl (150 mM) wereplated onto supplemented PPLO-Agar, andfilter paper discs saturated with differentconcentrations of PR-39 (0.024 nM to8 µM) were placed onto the agar surface.Cultures were incubated overnight at 37 °Cin a CO2-incubator (5% CO2) and bacterialgrowth around the paper discs was evalu-ated. In addition, minimal inhibitory con-centrations (MICs) of PR-39 for A. pleurop-neumoniae and E. coli (as the control) weredetermined by the broth microdilutionmethod according to the NCCLS guidelinesin document M31-A2 (2002) [7, 33]. PR-39concentrations ranging from 1.0 to 40 µMwere tested for their effect on bacterialgrowth (2–8 × 105 CFU /mL) in cation-adjusted Mueller-Hinton-Boullion (MHBII),PPLO medium with Isovitalex (1%) andTween 80 (0.1%) and Vast FastidiuousMedium (VFM) in 96-well plates.
The influence of PR-39 on bacterial anti-gen expression was investigated by 2-DPAGE. Briefly, 100 mL of an A. pleurop-neumoniae AP76 culture was grown (shak-ing at 200 rpm) to an OD660 of 0.3. This cul-ture was split into two 50 mL aliquots, andPR-39 was added to one aliquot to a finalconcentration of 0.5 µM; for control cul-tures, an equal volume of NaCl (150 mM)was added. This PR-39 concentration waschosen based on the average PR-39 concen-tration in BALF from pigs 21 days postinfection and an estimated dilution factor ofBALF with respect to ELF of 6 to 40 [21].
Antimicrobial peptides in porcine BALF 79
Cultures were further incubated while shak-ing for 1 h, and grown to an OD660 ofapproximately 0.6. Harvesting and treat-ment of bacterial cells for 2-D PAGE wasdone as described previously [22].
2.7. PR-39 capture ELISA
A PR-39 capture ELISA was conductedessentially as previously described [39].Briefly, 96-well microtitre plates (Max-iSorbTM Surface, Nunc, Roskilde, Denmark)were coated overnight at 4 °C with mousemAb against PR-39 (1.4 µg/mL) in carbon-ate buffer (0.1 M; pH 9.8). After washingtwice with PBST (phosphate buffered salinesupplemented with 0.05% Tween 20; PBSis 10 mM Na2HPO4, 137 mM NaCl, 2.7 mMKCl, 1.76 mM KH2PO4) plates were blockedfor 2 h at 22 °C in PBST containing 1%bovine serum albumin (BSA). All BALFsamples, antibodies and conjugates wereused in a dilution buffer; the dilution bufferwas PBS containing 1% BSA and 0.01%cetrimonium bromide (CETAB, Sigma, St.Louis, MO, USA). This capture ELISA wasused to determine PR-39 concentrations inBALF using a reference-standard-method[6] based on serial two-fold dilutions ofBALF (1:2 to 1:128) and synthetic PR-39(12.5 ng/mL initial concentration). Briefly,PR-39 mouse mAb coated plates were incu-bated with BALF. After four washes, a ratpolyclonal antibody to PR-39 (30 µg/mL)was added [33, 39]. After three washes, theplates were incubated with the conjugate(peroxidase-labelled goat-anti-rat Ab (Dian-ova, Hamburg, Germany)). All incubationsteps were done for 2 h (shaking) at 22 °C.After three washes, the plates were incu-bated with substrate SeramunBlue® fast (3,3’, 5, 5’ tetramethylbenzidine, 1.2 mM, andhydrogen peroxide, 3 mM; Seramun Diag-nostika Gmbh, Wolzig, Germany) for 10 minat 22 °C followed by the addition of a stopsolution (H2SO4, 2 M) according to themanufacturer’s instructions. Colour devel-opment was determined at 450 nm using amicroplate reader (SLT-Spectra, SLT Lab-instruments Deutschland GmbH, Crailsheim,
Germany). Differences in PR-39 concentra-tions were evaluated using the Signed Ranktest for Paired Samples (SAS® statisticalsoftware). The potential use of the PR-39capture ELISA for diagnostics was investi-gated by constructing a Receiver OperatingCharacteristics (ROC) plot resulting from thecalculation of sensitivities and specificitiesfor different potential cut-off-values [40].
3. RESULTS
3.1. Clinical study
Within 2 days post infection 30 pigsshowed an increase in body temperatureabove 40 °C and developed clinical signssuch as inappetence and coughing. Threepigs (one infected with A. pleuropneumoniaeserotype 2 and two infected with A. pleuro-pneumoniae serotype 7) died within thefirst 3 days post infection. BALF samplesfrom ten pigs surviving until the end of theexperiment were selected for further inves-tigation (Tab. I).
3.2. Protein patterns in BALF of healthy and A. pleuropneumoniae infected pigs
Approximately 80 spots were found tobe differentially expressed in BALF samplestaken from individual pigs before as well ason day 4 or 7 and day 21 after infection. Acomparison of individual BALF samplesfrom day 4 or 7 after infection with BALFsamples prior to infection did not show aconsistent protein pattern among differentpigs irrespective of the day or the serotype(data not shown). In contrast, BALF fromday 21 after infection consistently containedincreased amounts of 12 proteins in the trip-licate gels; in each pig at least 8 spots wereincreased with each individual spot beingpresent in at least 7 out of 10 pigs (Fig. 1).However, the severity of the lung lesionsand the number of additional spots in theindividual pigs were not correlated.
80 I. Hennig-Pauka et al.Ta
ble
I. C
linic
al, p
atho
logi
cal a
nd m
icro
biol
ogic
al p
aram
eter
s of
pig
s se
lect
ed f
or 2
-D P
AG
E o
f B
AL
F sa
mpl
es.
Pig
Infe
ctio
n w
ith
(A. p
p. s
erot
ype)
Ave
rage
bod
y te
mpe
ratu
re
days
1–3
p. i
nf.
Lun
g le
sion
sc
orea
PMN
% a
cute
(d
ay 4
/7 p
. inf
.)PM
N %
day
21
p. i
nf.
PR-3
9 on
day
21
p. i
nf. (
nM)
A. p
p. in
BA
LF
day
4/7
p. in
f. b
A. p
p. in
BA
LF
day
21 p
. inf
. bA
. pp.
in lu
ng ti
ssue
da
y 21
p. i
nf.b
12
40.0
0.3
7.5
16.5
18.6
00
02
239
.80.
511
8.7
0.4
00
33
239
.31.
17.
34.
50.
52
03
42
40.3
0.1
105.
51.
12
01
57
40.8
3.15
12.2
22.
41
03
67
40.8
5.21
7.5
22.8
30.3
10
37
740
.012
.99
1142
34.7
00
18
740
.213
.38
8.2
52.2
75.9
10
19
740
.47.
4915
.736
19.6
03
310
740
.712
.56.
315
66.8
03
3
a T
he s
chem
e of
lung
lesi
on m
appi
ng a
nd e
valu
atio
n de
scri
bed
by H
anna
n et
al.
[17]
and
stip
ulat
ed in
the
Eur
opea
n Ph
arm
acop
oeia
for
the
test
ing
of A
. ple
uro-
pneu
mon
iae
vacc
ines
was
use
d.b
0, n
o de
tect
able
col
onie
s; 1
, les
s th
an 1
00 c
olon
ies;
2, m
ore
than
100
col
onie
s bu
t no
conf
luen
t gro
wth
; 3, c
onfl
uent
gro
wth
.
Tab
le I
I. P
rote
in id
entif
icat
ion
by Q
-TO
F-m
ass
spec
trom
etry
.
Spo
t nu
mbe
rPr
otei
n na
me
Acc
essi
on
num
ber
The
oret
ical
M
r b/ I
Pc
Mat
chin
g pe
ptid
esC
over
age
(%)
Mat
ched
pep
tide
sa
6C
algr
anul
in C
P80
310
1060
8/5.
672
27L
GH
YD
TL
IK--
----
----
----
----
----
----
IFQ
NL
DA
NQ
DE
QV
SF K
9P
R-3
9P
8005
419
464/
9.96
547
ME
TQ
RA
SL
CL
GR
----
----
---
----
---
----
----
-- -
---L
NE
QSS
EA
NLY
RL
LE
L
DQ
PPK
AD
ED
P G
TPK
PVSF
TV
K--
----
---
----
----
-E N
GR
----
---
----
----
-- -
----
----
- -
RR
PR
PP
YL
P R
PR
PPP
FFP
P R
----
----
- --
-FP
PR
FP
G K
-10
Pro
phen
in-2
pr
ecur
sor
P51
525
2583
9/10
.56
221
LN
EQ
SS
EA
NLY
RL
LE
L D
QPP
KA
DE
DP
GT
PKPV
SFT
V K
----
----
- R
PP
EL
CD
FK
E N
GR
a M
asco
t se
arch
res
ults
(w
ww
.mat
rixs
cien
ce.c
om)
base
d on
de-
novo
-seq
uenc
ing
by Q
-TO
F m
ass
spec
trom
etry
ana
lysi
s of
the
exc
ised
pro
tein
spo
ts s
how
n in
Fig
ure
1. D
ashe
s in
dica
te m
issi
ng a
min
o ac
ids.
One
-let
ter
code
s ar
e us
ed f
or a
min
o ac
ids.
b R
elat
ive
mol
ecul
ar m
ass
in D
alto
n.c
Isoe
lect
ric
poin
t.
Antimicrobial peptides in porcine BALF 81
3.3. Identification of differentially expressed BALF-proteins
The eight most prominent spots wereanalysed by mass spectrometry, and pep-tide mass fingerprint analysis revealed sig-nificant matches for three protein spots inthe SwissProt database (Tab. II). The pro-teins were identified as prophenin-2 precur-sor and PR-39, both cathelicidin antimicro-bial peptides, and calgranulin C, a memberof the highly variable S100-protein familyof calcium-binding proteins.
Western Blot analysis confirmed the dif-ferential expression of PR-39 that wasobserved in 2-D PAGE. Precursor and proc-essed forms were detected reproducibly inBALF taken from pigs on day 21 afterinfection but not in BALF taken prior toinfection (Fig. 2).
3.4. Influence of PR-39 on growth and antigen expression of A. pleuropneumoniae
Because PR-39 is a bactericidal peptide,we investigated whether the in vitro growthof A. pleuropneumoniae could be inhibitedwith concentrations of PR-39 that would bepresent in the ELF of infected pigs. Despiteusing up to 8 µM concentrations (more than10 times the concentration calculated to bepresent in the ELF (0.5 µM)), no inhibitionof growth was observed in the inhibitionzone assay. Using the microdilution method,an MIC of 5 µM was determined for A. pleu-ropneumoniae in the VFM medium only; inE. coli the MIC was 1 µM with all media.
Because it was reported recently that vir-ulence gene expression in Salmonella Typh-imurium is regulated by cationic antimi-crobial peptides, the protein profiles of
Figure 1. Silver stained 2-D-PAGE of BALF from a representative pig prior to infection (A) andon day 21 post infection with A. pleuropneumoniae (B). Spots 3, 5 and 7 were present in seven pigs,spots 4, 6 and 8 in eight pigs, spots 1, 9, and 11 in nine pigs and spots 2, 10 and 12 consistently inall pigs. In addition to spots 6, 9 and 10, amino acid sequences were determined for spots 2, 5 and12 using Q-TOF mass spectrometry and data base analyses (www. matrixscience.com), but no sig-nificant matches were detected.
82 I. Hennig-Pauka et al.
A. pleuropneumoniae cultured in the pres-ence and absence of PR-39 were investi-gated. The PR-39 concentration used was0.5 µM, which is within the concentrationrange in BALF from infected animals (0.4to 75.9 nM) and concentrations in ELF areestimated to be approximately 6 to 40 foldhigher. A comparison of protein patternsof bacterial whole cell lysates did not showclear differences, implying that PR-39 inthe concentrations present in ELF ofA. pleuropneumoniae infected pigs doesnot inhibit growth or regulate gene expres-sion of the pathogen.
3.5. PR-39 is significantly increased in BALF of A. pleuropneumoniae convalescent pigs on day 21 post infection
Concentrations of PR-39 were increased(p < 0.05) in BALF on day 21 after infectionwhen compared to concentrations beforeinfection and in mock-infected control ani-mals (Fig. 3). The increase of PR-39 inBALF from day 4 and 7 after infection was
not significant and no difference in PR-39concentrations was found between BALFof day 4 after infection with A. pleuropneu-moniae serotype 2, BALF of day 7 after infec-tion with A. pleuropneumoniae serotype 7,
Figure 3. PR-39 concentrations in BALF (B) and serum/plasma (S) of infected pigs and in BALFof mock-infected pigs (BC). PR-39 concentrations were quantified using an ELISA. The box rep-resents the 50% between 25% and 75% quartiles. The line inside the box indicates the median. Thetop and bottom lines denote maximum and minimum values. The numbers on the top of the boxesindicate the number of animals examined. * p < 0.05, Signed rank test for Paired Samples.
Figure 2. Western blot analysis of syntheticPR-39 (10 ng; lane 1), BALF-samples from pigsprior to infection (lanes 2, 4, 6) and 21 days postinfection (lanes 3, 5, 7). The blot was developedusing mouse anti-PR-39 mAb and a chemilumi-nescent peroxidase substrate. The high molec-ular weight bands also visible on the blot arecaused by non-specific reactions of the peroxi-dase conjugate.
Antimicrobial peptides in porcine BALF 83
and BALF of day 7 after mock infection. Fur-thermore, the concentrations of PR-39 inserum/plasma remained unchanged (Fig. 3).
3.6. Potential use of PR-39 as a diagnostic marker for respiratory health in convalescent pigs
To investigate whether concentrationsof PR-39 in BALF could be used as a diag-nostic marker to detect convalescent pigsafter an experimental A. pleuropneumoniaeinfection, the severity of the lung lesions,the relative number of PMN in BALF andthe concentration of PR-39 in BALF werecompared. Concentrations of PR-39 inBALF on day 21 after infection were cor-related significantly to the lung lesion score(Spearman correlation coefficient 0.6, p <0.001). Furthermore, a comparison of PR-
39 concentrations and the relative numberof PMN in BALF, which is currentlyaccepted as the sole diagnostic marker forrespiratory health in pigs, revealed that thenumber of PMN was clearly increased onday 4 and 7 post infection but decreasedagain on day 21 post infection. To assess thepossible diagnostic value of both concen-trations of PR-39 and relative number ofPMNs on day 21 post infection, a ReceiverOperating Characteristic (ROC) plot wasconstructed (Fig. 4). Concentration of PR-39,using a cut-off value of 1 nM, identified con-valescent pigs with lung alterations causedby A. pleuropneumoniae with 90% sensitiv-ity and 70% specificity in the chronic stageof infection (21 days after infection). Thisis clearly superior to the relative number ofPMN, which, at the recommended cut-offof 8%, has a sensitivity of only 40%.
Figure 4. Receiver Operating Characteristic (ROC) plots for PR-39 and PMN on day 21 post infec-tion. Numbered symbols in the curves are potential cut-off values for PR-39 (squares) and polymor-phonuclear neutrophils (PMN, triangles) in BALF, which were positioned at their respective calcu-lated sensitivities and 100-specificities. The more distant the potential cut-off value on the curvefrom a diagonal between the X- and Y- axes, the lower the combined number of false-positives andfalse-negatives. The dotted lines indicate the sensitivity and 100-specificity-values for the suggestedcut-off value of PR-39 at 1 nM.
84 I. Hennig-Pauka et al.
4. DISCUSSION
An initial concern when proposing theuse of BALF for diagnostic purposes in pigsis whether it is practical. However, methodssuch as the sampling of BALF with dispos-able materials have been established in recentyears [8] and are now practical and cost-efficient, particularly when considering thehigh financial risks resulting from the intro-duction of respiratory tract pathogens intoclosed herds via convalescent animals. Asecond major concern for using this proce-dure is the high variability of BALF proteincontent [21]. However, the majority of pro-teins are serum-derived [19], and a BALF-based diagnostic scheme is considered prac-tical if it is based on ELF-specific compo-nents [25]. Therefore, the 2-D PAGE-basedapproach was considered to be a reasonablescreening method to identify possible ELF-specific marker proteins in the complexBALF samples.
In order to investigate changes in BALFprotein samples from healthy and convales-cent animals, we chose an establishedA. pleuropneumoniae infection model [3,4]. This model closely mimics the naturalinfection and commonly results in a lowmortality but high morbidity of infectedpigs. Furthermore, numerous experimentshave shown that lavaging pigs at weeklyintervals does not influence respiratoryhealth. The clinical data showed a typicalcourse of A. pleuropneumoniae disease.
Because 2-D PAGE has been reported tobe prone to methodological variation [15],we compared only protein profiles thatwere evident on triplicate gels. A compari-son of protein patterns of BALF-samplesfrom days 4 or 7 after infection to controlBALF taken prior to infection showed noconsistent differences. This finding likelyreflects individual differences in the courseof disease and in the acute local and sys-temic immune response commonly seen ingroups of outbred animals. The differencesseen in BALF taken from convalescent pigssuggests a consistent pattern in the late reac-
tion to respiratory disease, which likelyreflects the occurrence of consolidation andtissue repair processes. The finding that notall proteins observed to be upregulatedwere detected in the BALF of all animals,is likely due to individual differences and tothe generally low concentrations of the pro-teins in question. The identification of onlythree of the ten proteins analysed by massspectrometry is likely due to the lack ofavailability of the entire porcine genomesequence. All three proteins identified(PR-39, prophenin-2 and calgranulin C) arereleased by activated PMN. Both PR-39and prophenin-2 belong to the cathelicidinfamily of antimicrobial peptides (reviewedby Zanetti [37]). The antibacterial proper-ties of the amino terminal 60 residues ofprophenin-1, primarily directed against gramnegative organisms, have long been known[18, 23], and recently it has been shownthat, even in the presence of a lung sur-factant preparation, an 18-residue carboxyterminal fragment also possesses antibacte-rial activity [9, 35]. Calgranulin C belongsto a family of calcium- and zinc-bindingproteins and also possesses antimicrobial aswell as filariacidal activity [16]. BecausePR-39 has been reported to be increased inserum after infection [39] and, because thisantimicrobial peptide is involved in woundrepair processes [12], we hypothesised thatPR-39 might be an appropriate diseasemarker in BALF facilitating the detectionof convalescent animals, which still mightbe carriers of the pathogen. Therefore, amongthe three proteins identified, we regardedPR-39 to be the most promising candidatefor further investigation and performed alter-native methods to confirm our 2-D PAGEfindings. Western blot analysis (Fig. 2) anddata from a PR-39 capture ELISA (Fig. 3)confirmed our proteomic data showingincreased PR-39 in the BALF of pigs 21 dayspost infection. The finding that concentra-tions of PR-39 are not increased in serum,on the contrary to what had been reportedpreviously for gastrointestinal infections inpigs with Salmonella Typhimurium [39], islikely due to the stage of infection. Thus, in
Antimicrobial peptides in porcine BALF 85
acute systemic A. pleuropneumoniae infec-tion, serum concentrations of PR-39 alsoshow a slight increase (Fig. 3). In chronicinfection, however, the infection is local-ised in the lung and the involvement of PR-39 in wound repair processes might be itsprimary role [12] at this stage of disease.
The finding that PR-39 did not have bac-tericidal activity on A. pleuropneumoniaewas unexpected since the bactericidal activ-ity of PR-39 is well documented againstother gram-negative bacteria [5, 23]. Thislack of bactericidal activity was confirmedby the determination of the MIC-value whichwas found to be 5 times higher than theMIC-value of 1 µM determined for E. coliin accordance with previous reports [7, 33].The apparent resistance of A. pleuropneu-moniae against PR-39 was further con-firmed by the results of the 2-D PAGEwhere no impact on gene expression couldbe found as shown recently in SalmonellaTyphimurium using the synthetic cationicantimicrobial peptide polymyxin [2]. Thisresistance of A. pleuropneumoniae to PR-39 observed in vitro is in accordance withits ability to persist for extended periods oftime on the respiratory epithelium of pigs[11, 20]. The cause for this resistance remainsto be determined.
After confirming that PR-39 is increasedin BALF taken from pigs 21 days afterinfection, we investigated whether it couldbe used as a disease marker by constructinga Receiver Operating Characteristic (ROC)plot. This would establish a quantitativediagnostic test to differentiate, at differentcut-off values, between healthy and dis-eased individuals (Fig. 4). The optimal cut-off value determined for PR-39 (1 nM)would correspond to a concentration ofapproximately 40 nM in ELF; a value thatis well below bactericidal concentrations [5].
In summary, our findings show for thefirst time that concentrations of PR-39 inBALF is a potential diagnostic markerallowing the detection of clinically healthypigs convalescent from an experimentalA. pleuropneumoniae infection. Because the
introduction of respiratory tract pathogensinto closed herds via convalescent pigs is amajor risk to the swine industry, thismethod could become an important screen-ing aid for swine producers. Further studiesshould be conducted to investigate theapplicability of this approach to other res-piratory pathogens, e.g. Haemophilus para-suis, Mycoplasma hyopneumoniae, andviruses. Moreover the influence of abioticfactors (dust, NH3), which also impact res-piratory health in pigs, should be evaluated.
ACKNOWLEDGMENTS
This work was funded by the Sonderforsc-hungsbereich 587 (project A4) from the Deut-sche Forschungsgemeinschaft (DFG, Bonn,Germany). I. Jacobsen is a fellow of the Gradu-ate College 745 (project A1) of the DFG.
REFERENCES
[1] Anderson C., Potter A.A., Gerlach G.-F., Iso-lation and molecular characterization of spon-taneously occurring cytolysin-negative mutantsof Actinobacillus pleuropneumoniae serotype7, Infect. Immun. 59 (1991) 4110–4116.
[2] Bader M.W., Navarre W.W., Shiau W.,Nikaido H., Frye J.G., McClelland M., FangF.C., Miller S.I., Regulation of Salmonellatyphimurium virulence gene expression bycationic antimicrobial peptides, Mol. Micro-biol. 50 (2003) 219–230.
[3] Baltes N., Tonpitak W., Gerlach G.-F., Hennig-Pauka I., Hoffmann-Moujahid A., Ganter M.,Rothkoetter H.J., Actinobacillus pleuropneu-moniae iron transport and urease activity:effects on bacterial virulence and hostimmune response, Infect. Immun. 69 (2001)472–478.
[4] Baltes N., Hennig-Pauka I., Gerlach G.-F.,Both transferrin binding proteins are viru-lence factors in Actinobacillus pleuropneu-moniae serotype 7 infection, FEMS Micro-biol. Lett. 209 (2002) 283–287.
[5] Boman H.G., Agerberth B., Boman A., Mech-anisms of action on Escherichia coli of cecro-pin P1 and PR-39, two antibacterial peptidesfrom pig intestine, Infect. Immun. 61 (1993)2978–2984.
86 I. Hennig-Pauka et al.
[6] Butler J.E., Peterman J.H., Suter M., DierksS.E., The immunochemistry of solid-phasesandwich enzyme-linked immunosorbentassays, Fed. Proc. 46 (1987) 2548–2556.
[7] Chan Y.R., Zanetti M., Gennaro R., GalloR.L., Anti-microbial activity and cell bindingare controlled by sequence determinants in theanti-microbial peptide PR-39, J. Invest. Der-matol. 116 (2001) 230–235.
[8] Delbeck F., Tegeler R., Ganter M., Broncho-alveolar lavage on pig breeding farms, Dtsch.Tieraerztl. Wochenschr. 104 (1997) 374–378(in German).
[9] Dell'Angelica E.C., Schleicher C.H., SantomeJ.A., Primary structure and binding propertiesof calgranulin C, a novel S100-like calcium-binding protein from pig granulocytes, J. Biol.Chem. 269 (1994) 28929–28936.
[10] Diamond G., Legarda D., Ryan L.K., Theinnate immune response of the respiratory epi-thelium, Immunol. Rev. 173 (2000) 27–38.
[11] Dom P., Haesebrouck F., Ducatelle R.,Charlier G., In vivo association of Actinoba-cillus pleuropneumoniae serotype 2 with therespiratory epithelium of pigs, Infect. Immun.62 (1994) 1262–1267.
[12] Gallo R.L., Ono M., Povsic T., Page C., ErikssonE., Klagsbrun M., Bernfield M., Syndecans,cell surface heparan sulfate proteoglycans, areinduced by a proline-rich antimicrobial peptidefrom wounds, Proc. Natl. Acad. Sci. USA 91(1994) 11035–11039.
[13] Ganter M., Hensel A., Cellular variables inbronchoalveolar lavage fluids (BALF) inselected healthy pigs, Res. Vet. Sci. 63 (1997)215–217.
[14] Ganz T., Antimicrobial polypeptides in hostdefense of the respiratory tract, J. Clin. Invest.109 (2002) 693–697.
[15] Goerg A., Obermaier C., Boguth G., HarderA., Scheibe B., Wildgruber R., Weiss W., Thecurrent state of two-dimensional electro-phoresis with immobilized pH gradients,Electrophoresis 21 (2000) 1037–1053.
[16] Gottsch J.D., Eisinger S.W., Liu S.H., ScottA.L., Calgranulin C has filariacidal and filari-astatic activity, Infect. Immun. 67 (1999)6631–6636.
[17] Hannan P.C., Bhogal B.S., Fish J.P., Tylosintartrate and tiamutilin effects on experimentalpiglet pneumonia induced with pneumonicpig lung homogenate containing mycoplas-mas, bacteria and viruses, Res. Vet. Sci. 33(1982) 76–88.
[18] Harwig S.S., Kokryakov V.N., SwiderekK.M., Aleshina G.M., Zhao C., Lehrer R.I.,
Prophenin-1, an exceptionally proline-richantimicrobial peptide from porcine leuko-cytes, FEBS Lett. 362 (1995) 65–69.
[19] Hayem A., Proteins in bronchoalveolar lavagefluid, Ann. Biol. Clin. (Paris) 40 (1982) 231–233 (in French).
[20] Hennig I., Teutenberg-Riedel B., Gerlach G.-F., Downregulation of a protective Actinoba-cillus pleuropneumoniae antigen during thecourse of infection, Microb. Pathog. 26 (1999)53–63.
[21] Hennig-Pauka I., Ganter M., Gerlach G.-F.,Rothkoetter H.J., Enzyme activities, proteincontent and cellular variables in the pulmo-nary epithelial lining fluid in selected healthypigs, J. Vet. Med. A Physiol. Pathol. Clin.Med. 48 (2001) 631–639.
[22] Jacobsen I., Hennig-Pauka I., Baltes N., TrostM., Gerlach G.-F., Enzymes involved inanaerobic respiration appear to play a role inActinobacillus pleuropneumoniae virulence,Infect. Immun. 73 (2005) 226–234.
[23] Lee J.Y., Boman A., Sun C.X., Andersson M.,Jornvall H., Mutt V., Boman H.G., Antibac-terial peptides from pig intestine: isolation ofa mammalian cecropin, Proc. Natl. Acad. Sci.USA 86 (1989) 9159–9162.
[24] Leiner G., Franz B., Strutzberg K., GerlachG.-F., A novel enzyme-linked immunosorb-ent assay using the recombinant Actinobacil-lus pleuropneumoniae ApxII antigen for diag-nosis of pleuropneumonia in pig herds, Clin.Diagn. Lab. Immunol. 6 (1999) 630–632.
[25] Noel-Georis I., Bernard A., Falmagne P.,Wattiez R., Proteomics as the tool to search forlung disease markers in bronchoalveolar lav-age, Dis. Markers 17 (2001) 271–284.
[26] Noel-Georis I., Bernard A., Falmagne P.,Wattiez R., Database of bronchoalveolar lav-age fluid proteins, J. Chromatogr. B Analyt.Technol. Biomed. Life Sci. 771 (2002) 221–236.
[27] Pappin D.J., Hojrup P., Bleasby A.J., Rapididentification of proteins by peptide-mass fin-gerprinting, Curr. Biol. 3 (1993) 327–332.
[28] Ramanathan B., Davis E.G., Ross C.R.,Blecha F., Cathelicidins: microbicidal activ-ity, mechanisms of action, and roles in innateimmunity, Microbes Infect. 4 (2002) 361–372.
[29] Russell J.P., Diamond G., Tarver A.P.,Scanlin T.F., Bevins C.L., Coordinate induc-tion of two antibiotic genes in tracheal epithe-lial cells exposed to the inflammatory media-tors lipopolysaccharide and tumor necrosis
Antimicrobial peptides in porcine BALF 87
To access this journal online: www.edpsciences.org
factor alpha, Infect. Immun. 64 (1996) 1565–1568.
[30] Sambrook J., Fritsch E.F., Maniatis T.,Molecular Cloning: a Laboratory Manual,second edition, Cold Spring Harbor Labora-tory, Cold Spring Harbor, New York, 1989.
[31] Shi J., Ganz T., The role of protegrins andother elastase-activated polypeptides in thebactericidal properties of porcine inflamma-tory fluids, Infect. Immun. 66 (1998) 3611–3617.
[32] Shi J., Ross C.R., Chengappa M.M., BlechaF., Identification of a proline-arginine-richantibacterial peptide from neutrophils that isanalogous to PR-39, an antibacterial peptidefrom the small intestine, J. Leukoc. Biol. 56(1994) 807–811.
[33] Shi J., Ross C.R., Chengappa M.M., SylteM.J., McVey D.S., Blecha F., Antibacterialactivity of a synthetic peptide (PR-26) derivedfrom PR-39, a proline-arginine-rich neu-trophil antimicrobial peptide, Antimicrob.Agents Chemother. 40 (1996) 115–121.
[34] Tonpitak W., Baltes N., Hennig-Pauka I.,Gerlach G.-F., Construction of an Actinoba-cillus pleuropneumoniae serotype 2 prototypelive negative-marker vaccine, Infect. Immun.70 (2002) 7120–7125.
[35] Wang Y., Walter G., Herting E., Agerberth B.,Johansson J., Antibacterial activities of thecathelicidins prophenin (residues 62 to 79)and LL-37 in the presence of a lung surfactantpreparation, Antimicrob. Agents Chemother.48 (2004) 2097–2100.
[36] Wattiez R., Hermans C., Cruyt C., Bernard A.,Falmagne P., Human bronchoalveolar lavagefluid protein two-dimensional database: studyof interstitial lung diseases, Electrophoresis21 (2000) 2703–2712.
[37] Zanetti M., Cathelicidins, multifunctionalpeptides of the innate immunity, J. Leukoc.Biol. 75 (2004) 39–48.
[38] Zanetti M., Gennaro R., Romeo D., Catheli-cidins: a novel protein family with a commonproregion and a variable C-terminal antimi-crobial domain, FEBS Lett. 374 (1995) 1–5.
[39] Zhang G., Ross C.R., Dritz S.S., Nietfeld J.C.,Blecha F., Salmonella infection increases por-cine antibacterial peptide concentrations inserum, Clin. Diagn. Lab. Immunol. 4 (1997)774–777.
[40] Zweig M.H., Campbell G., Receiver Operat-ing Characteristic (ROC) plots – A fundamen-tal evaluation tool in clinical medicine, Clin.Chem. 39 (1993) 561–577.