Group Title: BMC Infectious Diseases
Title: Different inflammatory responses are associated with Ureaplasma parvum-induced UTI and urolith formation
CITATION PDF VIEWER THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00099934/00001
 Material Information
Title: Different inflammatory responses are associated with Ureaplasma parvum-induced UTI and urolith formation
Physical Description: Book
Language: English
Creator: Reyes, Leticia
Reinhard, Mary
Brown, Mary
Publisher: BMC Infectious Diseases
Publication Date: 2009
 Notes
Abstract: BACKGROUND:Epidemiologic studies show a strong association between Ureaplasmas and urogenital tract disease in humans. Since healthy humans can be colonized with Ureaplasmas, its role as a pathogen remains controversial. In order to begin to define the role of the host in disease, we developed a rodent model of urinary tract infection (UTI) using Fischer 344 (F344) rats. Animals were inoculated with sterile broth, 101, 103, 105, 107, or 109 log CFU of a rat-adapted strain of Ureaplasma parvum.RESULTS:Infected animals exhibited two distinct profiles, asymptomatic UTI and UTI complicated with struvite urolithiasis. Inoculum dose of U. parvum affected the incidence of UTI, and 50% to 57% of animals inoculated with = 107 CFU of U. parvum remained infected (p < 0.04). However, inoculum dose did not influence immune response to U. parvum. Asymptomatic UTI was characterized by a minimal immune response that was predominantly monocytic and lymphocytic, with limited lesions, and elevated urinary levels of IFN-?, IL-18 and MCP-1 (P = 0.02). UTI complicated with struvite formation was characterized by an exaggerated immune response that was mostly neutrophilic (P = 0.0001), with lesions that showed extensive uroepithelial hyperplasia (P = 0.0001), and a predominance of IL-1a, IL-1ß, and GRO/KC in the urine (P = 0.02). Animals with asymptomatic UTI also had a significantly high rate of kidney infection (P = 0.0005).CONCLUSION:Complications associated with U. parvum infection are primarily dependent upon host-specific factors rather than Ureaplasma microbial load. The immune response in F344 rats is similar to that which occurs in humans with ureaplasmal associated disease. Therefore, this model of infection is a useful tool for elucidating U. parvum-host interactions that confer UTI and disease.
General Note: Start page 9
General Note: M3: 10.1186/1471-2334-9-9
 Record Information
Bibliographic ID: UF00099934
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: Open Access: http://www.biomedcentral.com/info/about/openaccess/
Resource Identifier: issn - 1471-2334
http://www.biomedcentral.com/1471-2334/9/9

Downloads

This item has the following downloads:

PDF ( 2 MBs ) ( PDF )


Full Text



BMC Infectious Diseases .ioll



Research article

Different inflammatory responses are associated with Ureaplasma
parvum-induced UTI and urolith formation
Leticia Reyes*, Mary Reinhard and Mary B Brown


Address: Department of Infectious Disease & Pathology, College of Veterinary Medicine, University of Florida, Gainesville, FL, USA
Email: Leticia Reyes* lreyes@ufl.edu; Mary Reinhard ReinhardM@vetmed.ufl.edu; Mary B Brown Mbbrown@ufl.edu
* Corresponding author


)
Central


Published: 26 January 2009
BMC Infectious Diseases 2009, 9:9 doi:10.1186/1471-2334-9-9


Received: 24 August 2008
Accepted: 26 January 2009


This article is available from: http://www.biomedcentral.com/1471-2334/9/9
2009 Reyes et al; licensee BioMed Central Ltd.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.ore/licenses/by/2.0),
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.



Abstract
Background: Epidemiologic studies show a strong association between Ureaplasmas and
urogenital tract disease in humans. Since healthy humans can be colonized with Ureaplasmas, its role
as a pathogen remains controversial. In order to begin to define the role of the host in disease, we
developed a rodent model of urinary tract infection (UTI) using Fischer 344 (F344) rats. Animals
were inoculated with sterile broth, 101, 103, 105, 107, or 109 log CFU of a rat-adapted strain of
Ureaplasma parvum.
Results: Infected animals exhibited two distinct profiles, asymptomatic UTI and UTI complicated
with struvite urolithiasis. Inoculum dose of U. parvum affected the incidence of UTI, and 50% to 57%
of animals inoculated with > 107 CFU of U. parvum remained infected (p < 0.04). However,
inoculum dose did not influence immune response to U. parvum. Asymptomatic UTI was
characterized by a minimal immune response that was predominantly monocytic and lymphocytic,
with limited lesions, and elevated urinary levels of IFN-y, IL-18 and MCP-I (P 0.02). UTI
complicated with struvite formation was characterized by an exaggerated immune response that
was mostly neutrophilic (P 0.0001), with lesions that showed extensive uroepithelial hyperplasia
(P < 0.0001), and a predominance of IL-lI~, IL-I3P, and GRO/KC in the urine (P < 0.02). Animals
with asymptomatic UTI also had a significantly high rate of kidney infection (P < 0.0005).
Conclusion: Complications associated with U. parvum infection are primarily dependent upon
host-specific factors rather than Ureaplasma microbial load. The immune response in F344 rats is
similar to that which occurs in humans with ureaplasmal associated disease. Therefore, this model
of infection is a useful tool for elucidating U. parvum-host interactions that confer UTI and disease.


Background
Ureaplasma species are among the most common isolates
from the human urogenital tract [1,21. Although ureaplas-
mas can be isolated from healthy individuals, epidemio-
logic studies have shown a strong association between
Ureaplasmas and various diseases including non-gono-
coccal urethritis (NGU), bacterial vaginosis, infertility,
prostatitis, epididymitis, urinary tract infection (UTI),


nephrolithiasis, postpartum endometritis, chorioamnio-
nitis, spontaneous abortion, premature birth, stillbirth
and neonatal pneumonia [1-41. Animal studies have dem-
onstrated the ability of ureaplasmas to induce pneumo-
nia, pyelonephritis, and struvite uroliths (urinary tract
stones primarily composed of magnesium, ammonium,
and phosphate) [5-10]. The most common host immune
response to Ureaplasma during disease states involves ele-


Page 1 of 12
(page number not for citation purposes)







http://www.biomedcentral.com/1471-2334/9/9


vated pro-inflammatory cytokines, most notably IL-la,
IL- 13, IL-6, IL-8, MCP-1 andTNF-a, accompanied by infil-
tration of neutrophils and macrophages at sites of infec-
tion [1,10-121. However, little work has been done to
characterize the immune response during uncomplicated
infections. Therefore, the complex interactions between
Ureaplasma and the host that lead to simple colonization
versus inflammation and disease are largely unknown. In
a recent study, we showed that the inbred rat strain Fischer
344 (F344) is susceptible to UTI induced by a rat adapted
strain of Ureaplasma parvum isolated from the urine of a
patient with recurrent UTI [13]. As part of that study, we
found that 60% of infected F344 rats developed struvite
uroliths, which were associated with an exaggerated
inflammatory response that is similar to what has been
reported in other disease states caused by Ureaplasma
infection [1,5,6,11,12]. Interestingly, the other 40% of
F344 rats developed uncomplicated UTI that was charac-
terized by low concentrations of pro-inflammatory
cytokines in urine as well as mild to moderate lesions in
the lower urinary tract. Since F344 rats are an inbred
strain, this particular infection model would be useful for
identifying the host/Ureaplasma interactions that confer
disease or asymptomatic infection without confounding
variables that would be introduced by genetic variability.

In the study reported here, we examined the innate
immune response to UTI induced with varying microbial
concentrations of U. parvum in the F344 rat. By applying
an integrated approach that combines histopathology
with cytokine profiling, we were able to identify innate
immune response profiles that were significantly different
between an uncomplicated UTI and a UTI accompanied
by struvite formation. Our findings provide insights into
innate immune responses that are likely involved in the
development of complicated disease with Ureaplasma.

Methods
Ureaplasma preparation and culture
A host-adapted strain of U. parvum, designated strain 257-
48 was used for the entire study [13]. Fifty mls of U. par-
vum in logarithmic growth phase was aliquoted into 1 ml
volumes and stored at -80C. This stock was used for all
experiments.

For infection studies, one ml of the working stock was
grown in 45 ml of 10B broth for 12 to 16 hours at 37C.
The Ureaplasma culture was pelleted by centrifugation at
10,000 x g, at 4C, for 50 minutes. Due to the delicate
nature of Ureaplasma, the pellet was resuspended in 15 to
20 ml of fresh 10B broth instead of saline, to give a final
concentration of 109 CFU per ml then serially diluted to
produce various inocula that contained 107, 105, 103, and
101 CFU per ml. The CFU of all inocula (including all


serial dilutions) were confirmed by culture on A8 agar. For
each infection experiment, at least two animals were
included in each U. parvum dose group and experiments
were replicated a minimum of 5 times.

Inocula and animal tissues were serially diluted 10-fold in
10B broth to 10-10 and 10-5, respectively. For CFU determi-
nation, 20 gl from each sample and its corresponding
dilutions were plated on A8 agar. Agar plates were incu-
bated at 37 C in 5% CO2; broth cultures were incubated
at 37C in ambient air. Agar cultures were incubated for
at least 5 days before colonies were counted to determine
CFU.

Animals
Specific pathogen free F344 virgin female rats were pur-
chased from a commercial vendor (Charles River, Indian-
apolis, IN). All animals ranged in weight from 178-200
grams. Animal colonies were monitored and found free of
the following pathogens: Sendai virus, H-1 virus, rat
corona virus, sialodacroadenitis virus, reovirus type 3, Kil-
ham rat virus, Hantaan virus, M. pulmonis, respiratory
and enteric bacterial pathogens, endoparasites and
ectoparasites. All animals were handled in accordance
with procedures approved by the University of Florida
Institutional Animal Care and Use Committee.

All animals were handled within a biosafety laminar flow
hood. Rats were housed in Microisolator (Lab Products,
Inc., Maywood, NJ) cages in the same room under the
same temperature and light conditions. Control animals
were always handled before infected and housed in sepa-
rate microisolator cages in order to prevent contamina-
tion with Ureaplasma. All food, water, bedding, and caging
were autoclaved before use.

Rats were anesthetized and inoculated with sterile broth
or U. parvum inoculum into the bladder as previously
described [13]. For each infection experiment, a mini-
mum of two rats per inoculum dose were infected, so that
each dose was represented in each experiment.

Necropsy
Rats were necropsied at two weeks post-infection as previ-
ously described [13]. Prior to euthanasia, free catch urine
was collected for cytokine analysis. The bladder was proc-
essed for histopathologic evaluation. Each kidney was
transected sagittally so that a portion of the renal pelvis
was present in each section. One half of each kidney was
processed for histopathologic evaluation. The remaining
halves of the right and left kidneys for an individual ani-
mal were combined, minced in sterile 1 OB broth, and the
medium was aseptically removed and cultured for U. par-
vum.



Page 2 of 12
(page number not for citation purposes)


BMC Infectious Diseases 2009, 9:9







http://www.biomedcentral.com/1471-2334/9/9


Stone analysis
Bladder calculi were submitted to a commercial labora-
tory (Louis C. Herring and Co., Orlando, FL) and ana-
lyzed by integrated crystallography.

Histopathology
Bladder and kidney tissues were fixed in a paraformalde-
hyde-lysine-periodate [14] solution for 24 hours, then
washed 3 times in sterile saline and transferred to 70%
ethanol prior to processing. Tissues were processed rou-
tinely and stained with hematoxylin and eosin (H&E).

Bladder lesions were scored by a system developed in a
previous study [13]. Epithelial changes in bladder tissues
were scored as: 0 for none, 1 for minimal hyperplasia,
ulceration or effacement of epithelium by inflammation;
2 for mild hyperplasia and rare dense inflammatory infil-
trates, and 3 for the same changes noted in a score of 2 but
accompanied with marked erosion and/or ulceration of
the epithelial surface. Scoring for cell types that comprised
the inflammatory infiltrate was: 1 for primarily mononu-
clear cells lymphocytess, plasma cells and macrophages),
2 for mononuclear cells and neutrophils, and 3 for mono-
nuclear cells, neutrophils and fibrous infiltrates. Kidney
tissues were scored on the basis of total area affected,
which was: 1 for less than 10%, 2 for 10 to 50%, and 3 for
greater than 50%.

Detection of urinary cytokines
Urine from control and infected rats was analyzed for the
presence of cytokines with a multiplex antibody-immobi-
lized bead immunoassay (Lincoplex KIT, Linco Research,
Inc., St. Charles, MO). The manufacturer's protocol was
followed for the simultaneous detection of the following
cytokines and chemokines: GM-CSF, IL-la, IL-1 3, IL-6, IL-
10, IL-12p70, IFN-y, IL-18, GRO/KC (growth related
oncogene/keratinocyte chemoattractant- the rat analog
for human IL-8), and TNF-a within the same aliquot of
urine. Briefly, a standard cocktail was serially diluted in
order to develop a standard curve for each analyte that
ranged from 3.2 to 2000 pg/ml. Urine samples were
diluted in assay buffer to obtain a total volume of 60 gl
per well, and run in duplicate as previously described
[13].


Table 1: Colonization of rat urinary tract by U. parvum.


Data analysis
Data from multiple experiments were grouped together in
order to make statistical analysis possible. Wherever pos-
sible, data were analyzed by one-way AN OVA when more
than two groups were included in the analysis. Fisher's
PLSD test was used when the ANOVA indicated a signifi-
cant difference among group means. An unpaired stu-
dent's T test was used for comparisons between two
groups. Contingency table analysis was used for compari-
sons of groups involving nominal data (positive vs. nega-
tive). Cytokine pattern recognition analysis was
performed using JMP Genomics 3.0 (SAS Institute, Cary,
NC). Datasets were initially evaluated by distribution
analysis and normalized prior to analysis by one-way
ANOVA using row by row modeling and Fischer C correc-
tion for multiple comparisons. For all analyses a probabil-
ity of P < 0.05 was considered significant.

Results
Impact of inoculum dose of U. parvum on colonization
rates in the urogenital tract
Ureaplasmas were not isolated from any site from any con-
trol rat (data not shown). Neither uroliths nor crystals
were detected at any site in animals inoculated with sterile
10B broth. There was no statistical difference in the log
CFU of U. parvum isolated from culture positive animals
among the various inoculum groups. The log CFU iso-
lated from the bladder tissue of culture positive animals
was 1.59 + 1.4 (mean SD), and 1.48 + 1.1 from kidney
tissue. However, there was a statistical difference in the
frequency of animals that remained infected 2 weeks post-
inoculation (Table 1). Animals inoculated with 107 or
higher CFU had the greatest frequency of bladder infec-
tions at time of necropsy (P < 0.04).

Clinical profiling of animals inoculated with U. parvum
Animals inoculated with U. parvum were divided into
three groups (Negative, UTI, or Struvite) stratified on the
basis of culture status and urolith status (Table 2). The
Negative group consisted of animals found to be culture
negative and urolith negative at time of necropsy. As
expected, the frequency of animals that were culture neg-
ative after inoculation with U. parvum decreased as the log
CFU of the inoculum increased. The UTI group represents


Number of rats positive (%) inoculated with log CFU U. parvuma


Bladder
Kidney
Both Sitesc


4/14 (29)
1/14 (7)
1/14 (7)


103
2/14 (14)
2/14 (14)
1/14 (7)


105
2/14 (14)
3/14 (21)
1/14 (7)


107
7/14 (50)
1/14 (7)
2/14 (14)


a) Data was collected at 2 weeks post-inoculation, and is a combination of 5 separate experiments.
b) Probability values were derived by G statistic, P values greater than 0.05 were considered not significant (NS).
c) Both sites refer to bladder and kidney.


Page 3 of 12
(page number not for citation purposes)


109
8/14 (57)
6/14 (43)
4/14 (29)


P value
0.04
NS
NS


BMC Infectious Diseases 2009, 9:9







http://www.biomedcentral.com/1471-2334/9/9


Table 2: Distribution of animals grouped within Negative, UTI, or Struvite profiles.

Inoculum dose of U. parvum (log CFU)


103
11/14 (79)
2/14 (14)
1/14 (7)


105
9/14 (64)
4/14 (29)
1/14 (7)


107
6/14 (43)
3/14 (21)
5/14 (36)


109
3/14 (21)
6/14 (43)
5/14 (36)


P valueb
0.012
NS
NS


a) Profiles are based on culture status and urolith status of animals at time of necropsy. Negative group consists of animals found to be culture
negative and urolith negative at time of necropsy. UTI group consists of animals that were culture positive in the bladder and/or kidney but urolith
negative. Struvite group consists of animals found to have bladder uroliths regardless of U. parvum culture status at time of necropsy. Data is a
combination of 5 separate experiments.
b) P values obtained by G statistic obtained by contingency analysis. NS = not significant.


animals that were culture positive in the bladder and/or
kidney at time of necropsy but were found to be negative
for uroliths. Eleven of 17 (64.7%) animals within the UTI
group had kidney infections at time of necropsy, which
was significantly greater (P < 0.0005) than the number of
renal infections within the urolith group [1 of 14 (7%)
animals]. All animals within the Struvite group had blad-
der stones, which were composed of 89 to 95% magne-
sium ammonium phosphate (struvite), 1 to 5% calcium
phosphate (carbonate apatite) and 3 to 10% protein and
blood. No animal had macroscopic evidence of kidney
stones at time of necropsy. Although not statistically sig-
nificant (P < 0.11), animals inoculated with 107 or greater
CFU tended to have the highest frequency of struvite uro-
liths.

The inoculum dose of U. parvum did not impact the distri-
bution of animals that developed either UTI or struvite
uroliths (Table 2). Further, there was no statistical differ-
ence between the log CFU cultured from bladder tissue
among these two clinical groups. The log CFU (mean
SD) of U. parvum isolated from the bladder of animals
within the UTI group was 1.59 + 2 and 1.53 + 1 from the
bladder of animals within the Struvite group.

Histologic characterization of the inflammatory response
to U. parvum
The extent and severity of bladder lesions as well as the
types of inflammatory cells present differed among
groups. There were no detectable lesions in bladder tissue
from control rats (see Figure 1, panel A). In animals inoc-
ulated with U. parvum, bladder lesions associated with
inflammation were highly variable. When present, lesions
consisted of infiltrates of lympho-plasma cells, macro-
phages, and neutrophils that were primarily located
within the epithelial layer (Figure 1, panel B and D). Mast
cells could be found within the submucosa regardless of
clinical profile (data not shown). Uroepithelial changes
ranged from spongiosis of epithelial cells with some
necrosis, exfoliation ofuroepithelium, or hyperplasia (see
Figure 1, panels B, C, and D). Animals with struvite uro-
liths had the most extensive lesions and the highest lesion


scores (Figure 2). In these animals, inflammation
extended into the submucosa and muscularis layers, and
was occasionally accompanied by venous congestion and
edema with a fibrinous reaction (Figure 1, panel B). Neu-
trophilic infiltrates were present in all of the animals
within the Struvite group. In contrast, infiltrates in ani-
mals within the Negative and UTI groups were predomi-
nantly lymphocytes, plasma cells and macrophages
(Figure 2, panel B). Animals with uroliths also had the
most extensive uroepithelial hyperplasia as shown in Fig-
ure 1, panel C. Animals within the Negative and UTI
groups had the mildest inflammatory changes (Figure 2).
However, somewhat surprisingly, more animals within
the Negative group had a higher degree of both inflamma-
tion and epithelial change than did animals within the
UTI group. Most of the animals in the UTI group exhibited
exfoliation of uroepithelium with some spongiosis.

Both the inflammatory cell type score and inoculating
dose of U. parvum influenced the lesion scores within the
Negative group. There was no difference in the lesions
scores pertaining to degree of inflammation and degree of
epithelial change among the inoculating dose groups
(data not shown). However, there was a significant differ-
ence (P < 0.006) in the inflammatory cell type score and
inoculating dose of U. parvum (Figure 2, panel D). A sig-
nificant number of animals that received log 5 CFU had a
mixed inflammatory cell infiltrate that included neu-
trophils as well as mononuclear cells lymphocytess,
plasma cells, and macrophages).

Kidney tissue from F344 rats were also evaluated for the
presence of inflammation. There were no significant
lesions present in the collecting ducts and renal pelvis of
control rats (Fig 3, panel A). Histopathologic findings in
kidney tissue from rats inoculated with U. parvum ranged
from minimal changes to varying degrees of inflammatory
infiltration that consisted of lymphocytes, plasma cells,
macrophages and neutrophils. In some animals, lesions
were characterized by scant multifocal areas of predomi-
nantly mononuclear cells that were present in the subepi-
thelial region of the renal pelvis (Figure 3, panel B). In


Page 4 of 12
(page number not for citation purposes)


Profile
Negative
UTI
Struvite


101
10/14 (71)
2/14 (14)
2/14 (14)


BMC Infectious Diseases 2009, 9:9







http://www.biomedcentral.com/1471-2334/9/9


Figure I
Summary of the bladder lesions found in F344 rats experimentally infected with U. parvum. Panel A is a 200x
magnification of bladder tissue from a Control rat. This sample represents a lesion score of 0 for degree of inflammation,
degree of epithelial change and inflammatory cell type. Panels B, C, and D are tissue sections from animals inoculated with U.
parvum that had a lesion score of 3 for degree of inflammation, degree of epithelial change and inflammatory cell type. Panel B
is a 200x magnification of bladder tissue from an animal within the struvite group. The asterisk demarcates the extensive
edema and fibrinous exudate infiltrating the submucosa. The black arrow points to uroepithelial effacement. Panel C is a 400x
magnification of extensive uroepithelial hyperplasia. Panel D is a magnified inset of Panel B that highlights the array of white
blood cells that comprised in the inflammatory cellular infiltrate in the tissues of infected animals. The blue arrow is pointing to
a neutrophil. The green arrow is pointing to a tissue macrophage. The yellow arrow is pointing to a plasma cell. The red arrow
is pointing to a lymphocyte.


these animals, the uroepithelial lining the pelvic space
was hyperplastic. Animals with the most severe kidney
lesions had extensive erosion of the uroepithelium, as
shown in Figure 3, panel D. In these animals, erosion of
the uroepithelial barrier was accompanied by hemorrhage
and infiltration with inflammatory cells (predominantly
neutrophilic) that spanned the pelvic luminal space,
through the epithelial layer and into the sub-epithelial
region of the pelvis (Figure 3, panel D). Other animals,


had inflammatory infiltration of the renal interstitium as
shown in Figure 3, panel C

Scoring of kidney tissues on the basis of total area affected
did not reveal any patterns that could be correlated to
local/active infection or clinical profile. Specifically, there
was no correlation between the total area affected score
and U. parvum culture status (data not shown). There also
was no correlation between the total area affected score


Page 5 of 12
(page number not for citation purposes)


BMC Infectious Diseases 2009, 9:9








http://www.biomedcentral.com/1471-2334/9/9


A P<0 0001
e3 a U *M-

2 P 0 0

1- 40000b -in-

Negative UTI Struvite




82 *Ow
E




C P
-


m -

Negative UTI Struvite

3- D
P<0.006
H-
0 2


E

Log 1 Log 3 Log 5 Log 7 Log 9
Inoculum Dose

Figure 2
Lesion scores of bladder tissue from F344 rats inocu-
lated with U. parvum. Lesion score analysis was done prior
to grouping each sample into a clinical profile (Panels A, B,
and C) or inoculum dose groups (Panel D). Data is a summa-
tion of 5 separate experiments. In panels A, B, and C, non-
parametric lesion scores were grouped according to the
clinical profile and analyzed by Kruskall- Wallis test (Control,
n = 6; Negative, n = 36; UTI, n = I I; Struvite, n = 10). Panel
D are the results of an inflammatory cell type lesion score
analysis performed on bladder tissues from animals in the
Negative group. Scoring for cell types that comprised the
inflammatory infiltrate was: I for primarily mononuclear cells
lymphocytess, plasma cells and macrophages), 2 for mononu-
clear cells and neutrophils, and 3 for mononuclear cells, neu-
trophils and fibrous infiltrates. Data is a summation of 5
separate experiments. Nonparametric lesion scores were
grouped according to the inoculum dose of U. parvum. Raw
lesion scores were analyzed by Kruskall- Wallis test (n = 10
for log I CFU, n = I I for log 3 CFU, n = 8 for log 5 CFU, n =
5 for log 7 CFU, and n = 3 for log 9 CFU). Values in all graphs
represent raw lesion scores for each biological replicate.
Horizontal bars demarcate the median value for each clinical
profile group.


and clinical profile (Negative, UTI, or Struvite), additional
file 1.

Urine cytokine analysis of clinical profiles
The relationships between specific cytokines and bladder
lesion scores among animals inoculated with sterile 10B
broth or U. parvum were examined by Spearman Correla-
tion analysis. There were no correlations between urine
cytokine levels and degree of inflammation score among
animals inoculated with sterile 10B broth (data not
shown). However, there was a significant correlation
(summarized in Table 3) between the degree of inflam-
mation score and urine concentrations of GRO/KC, IL-la,
IL- 1 P, IL- 10 and TNF-a. There was also a significant corre-
lation between degree of epithelial change and urine con-
centrations of GRO/KC and IL-1 l (Table 4). MCP-1 levels
negatively correlated with degree of epithelial change
(Table 4).

The cytokine profile in urine differed among groups.
Absolute concentrations of each individual cytokine in
urine were compared among clinical profiles (Control,
Negative, UTI and Struvite). Control animals had signifi-
cantly higher levels of MCP-1 than animals in either the
Negative or UTI groups (Figure 4). Animals in the Struvite
group had the highest levels of GRO/KC (Figure 4), IL-1 Ia,
IL-13P, IL-6, IL-10 and TNF-a (Figure 5) than did animals
in all other groups. There was no statistical difference in
the absolute amounts of urine GM-CSF, IFN-y, IL-2, IL-4,
IL- 12, and IL- 18 among groups (data not shown).

In order to identify distinctive chemokine/cytokine pat-
terns between clinical profiles associated with active infec-
tion, samples from animals within the Negative group
were excluded from this analysis. Each urine cytokine
multiplex from each animal was normalized prior to anal-
ysis by one-way ANOVA using row by row modeling and
Fischer C correction for multiple comparisons. Figure 6 is
a clustered heat map illustrating two cytokine profile clus-
ters that significantly differed between Control, UTI and
Struvite groups (P < 0.02). Both Control and UTI groups
showed a significant emphasis in IL-18 and MCP-1,
whereas the Struvite group showed a significant emphasis
in IL-la, IL-13P, and GRO/KC. Only the UTI group showed
a significant emphasis in IFN-y.

The impact of inoculum dose on the urine cytokine profiles
of animals in the Negative group
Urine cytokine data from control and culture negative ani-
mals was normalized prior to statistical analysis and ana-
lyzed as described above. There was a significant
difference in the overall pattern of IL-2, IL-4, IL-10, TNF-
a, IFN-y, and GRO/KC in the urine of culture negative ani-
mals that received different inoculating doses of U. par-
vum (P < 0.05). Figure 7 is a clustered heat map illustrating



Page 6 of 12
(page number not for citation purposes)


BMC Infectious Diseases 2009, 9:9







http://www.biomedcentral.com/1471-2334/9/9


Figure 3
Summary of the kidney lesions found in F344 rats experimentally infected with U. parvum. Panel A is a 200x mag-
nification of kidney tissue from a Control rat and demonstrates the lack of inflammatory lesions that are characteristic in ani-
mals inoculated with U. parvum. Panels B, C, and D are tissue sections from animals inoculated with U. parvum that had a lesion
score of 4 for total area affected. Panel B is a 400x magnification demonstrating the inflammatory infiltrate extending from the
renal pelvic space into the interstitium with uroepithelium largely intact. The black arrow is pointing to uroepithelial hyperpla-
sia. Panel C is a 400x magnification of renal tubules. The black arrow is pointing to the extensive inflammatory infiltrate
throughout the renal tubular interstitium. The yellow arrow is pointing to a glomerulus. Panel D is a 600x magnification of
renal uroepithelium at the edge of the pelvic space. The black arrow is pointing to extensive hemorrhage and disruption of the
uroepithelial barrier by a fibrinous inflammatory infiltrate.


two distinct cytokine profile clusters (green and red cluster
trees) among these animals. With the exception of log 5
and log 7 inoculating groups, there was an inverse rela-
tionship in the pattern of expression between the two
cytokine clusters.

Discussion
Ureaplasmas are an underappreciated pathogen of the
urogenital tract. Despite strong epidemiological evidence
and even experimental infections in humans that fulfilled
Koch's postulates [15], the etiologic role of Ureaplasmas is
confounded by the isolation of the microbe from the


lower urogenital tract of normal, asymptomatic individu-
als. In addition, the severity of disease for most mycoplas-
mal infections depends on the host immune response.
Therefore, experimental infections in genetically defined
animal models will be critical to unraveling the key inter-
actions in the host/parasite relationship that contribute to
disease severity. By using a combined approach involving
histopathology and cytokine profiling, we were able to
further characterize the immune response associated with
asymptomatic UTI and UTI complicated with struvite for-
mation.


Page 7 of 12
(page number not for citation purposes)


BMC Infectious Diseases 2009, 9:9







http://www.biomedcentral.com/1471-2334/9/9


Table 3: Urine chemokines/cytokines found to have a significant
correlation to Degree of Inflammationa

Cytokine Rhob Tied Z value Tied P value


GRO/KC
IL-laI
IL-1p3
IL-10
TNF-a


0.462
0.382
0.652
0.351
0.359


0.0004
0.003
0.000 I
0.007
0.006


a) Data was a combination of 5 separate experiments (n = 58).
b) Spearmann correlation analysis was performed on cytokine data
from animals inoculated with U. parvum regardless of their clinical
profile grouping. Listed Rho values have been corrected for tied
lesion scores.
The F344 rat strain is highly susceptible to development of
complicated UTI following experimental inoculation with
U. parvum [13]. In this study, we showed that varying the
inoculum dose of U. parvum significantly affected the fre-
quency of animals that remained colonized two weeks
after inoculation. Therefore, the initial microbial load is
important in establishing infection. However, once
infected, the proportion of animals that developed com-
plicated UTI in response to varying inocula of U. parvum
did not show a definitive dose response effect. More
importantly, the immune response to infection in this
group of animals with complicated UTI was consistent,
regardless of initial inoculating dose. For example, the
cytokine profile and urinary tract pathology of a struvite
positive animal that was inoculated with 101 CFU was


such as IFN-y, IL-18, and MCP-1 in the UTI group that
work synergistically to regulate monocyte/macrophage
activation [16-181. An intriguing finding was the signifi-
cant emphasis of IFN-y in the urine of animals with
asymptomatic UTI, since this cytokine is a potent priming
agent for macrophages [19]. This cytokine profile also
coincides with the cellular immune response in these ani-
mals that consisted of macrophages, lymphocytes, and
plasma cells, which resembles a profile that may be seen
during the healing or resolution phase of infection. We
cannot rule out that these animals could be displaying a
pre-resolution phase to infection, but there are indicators
suggesting that these animals have compromised immune
defense. For example, the immune profile of these ani-
mals was obtained while they were actively colonized
with U. parvum, and 65% exhibited an ascending infection
into the kidneys. Further, the microbial load of U. parvum
in animals with asymptomatic UTI was equivalent to ani-
mals in the Struvite group. Another intriguing feature in
animals with asymptomatic UTI was the overall lack of
uroepithelial proliferation that was present in varying
degrees in the Negative group as well as the Struvite group.
A primary defense mechanism of uroepithelium exposed
to bacteria involves desquamation, necrosis or apoptosis
followed by proliferation [20]. Therefore, the overall lack
of this response in animals with asymptomatic UTI also
implies that uroepithelial defense mechanisms may be
perturbed by U. parvum.


indistinguishable from a struvite positive animal that was F344 rats with struvite uroliths had a similar clinical pro-
inoculated with 109 CFU. This suggests that the initial file to what we have previously described [13]. Specifi-
microbial load of U. parvum is not a critical factor in the cally, these animals had the greatest concentration of pro-
development of complicated UTI in this rat strain. Fur- inflammatory cytokines in their urine (GRO/KC the rat
their, it supports the concept that, once infection is estab- analog of human IL-8, IL-la, IL-1, IL-6 and TNF-a) and
lished, the host inflammatory response is a key the most extensive inflammatory changes in bladder tis-
determinant of lesion severity in the urinary tract, sue. Since IL-13 is a known inducer of IL-8 and GRO
chemokines in human and murine epithelial cells [21-
As previously reported [13], animals with asymptomatic 23], it is not an unexpected finding that these cytokines
UTI had significantly less pro-inflammatory urine are closely linked in their expression. Cytokine pattern
cytokines and tissue damage when compared to rats with analysis showed this cytokine cluster is unique to animals
struvites. By profiling the entire cytokine milieu, we were with struvites. Moreover, there is a significant positive cor-
able to identify a significant predominance of cytokines relation between IL-1p, GRO/KC and the degree of his-
topathologic change, which suggests that IL-1 and GRO/
Table 4: Urine chemokines/cytokines found to have a significant KC are critical elements in a pro-inflammatory loop that
correlation to Degree of Epithelial Changea leads to chronic active inflammation, epithelial hyperpla-
sia and struvite formation as seen in struvite positive F344
Cytokine Rhob Tied Z value Tied P value rats. Most of the animals within the struvite group had

GRO/KC 0.3 70 2.79 0.0053 uroepithelial hyperplasia or erosion with hemorrhage and
IL-lp 0.407 3.08 0.0021 inflammation within the kidneys, yet none of these ani-
MCP-I -0.270 -2.04 0.042 mals had uroliths in the renal pelvis that could account
for these lesions. Therefore, although mechanical irrita-
a) Data was a combination of 5 separate experiments (n = 58). tion by the urolith itself may partially contribute to epi-
b) Spearmann correlation analysis was performed on cytokine data thelial erosion or hyperplasia in the bladder, it cannot
from animals inoculated with U. parvum regardless of their clinical entirely account for the lesions that were present in the
profile grouping. Listed Rho values have been corrected for tied urinary tract of these animals.
lesion scores. urnary tract of these animals.


Page 8 of 12
(page number not for citation purposes)


BMC Infectious Diseases 2009, 9:9







http://www.biomedcentral.com/1471-2334/9/9


p<0.0001


p<0.04


T


Control Neqative


-r


UTI Urolith


Figure 4
Urine chemokines detected in F344 rats inoculated
with sterile I OB or U. parvum. Data represent the mean
SD of a combination of 5 separate experiments. Urine
chemokine concentrations were grouped according to clini-
cal profile, control (n = 6), negative (n = 36), UTI (n = 16),
and Struvite (n = 13). P values within each graph were
obtained by one-way ANOVA. Fisher's PLSD test revealed
that GRO/KC concentrations in the struvite group were sig-
nificantly greater than the control, negative and UTI groups.
Fisher's PLSD test revealed that MCP- I concentrations in the
control group was significantly greater than the negative and
control groups.


The immune response of animals within the negative
group was highly variable and most likely comprises a
mixed population of rats, including animals that never
became colonized as well as animals that cleared the
infection at various time points post-inoculation. There-
fore, interpretation of data from this group of animals is
difficult and is done with caution. In spite of this limita-
tion, profiling urine cytokine data and bladder lesion
scores by inoculum dose was informative. The threshold
dose for successful colonization appears to be between log
5 and log 7 CFU, since 64% and 43% of rats respectively
were culture negative 2 weeks post-inoculation. The ani-
mals within the log 5 and log 7 CFU inoculation groups
also had the greatest flux in both pro-inflammatory (TNF-
a, IFN-y and GRO/KC) and anti-inflammatory (IL-4, and


8000-


Page 9 of 12
(page number not for citation purposes)


4000-


IL-10) cytokines. Interestingly, pattern analysis of urine
cytokine data showed two distinct clusters. The first clus-
ter identified in the negative group included IL-2, IL-4, IL-
10, and TNF-a; these cytokines were notable as they were
not part of the cytokine cluster groupings of U. parvum
infected animals. Therefore, these cytokines may be criti-
cal in directing a more efficient immune response that
leads to bacterial clearance with minimal tissue damage.
The second cytokine cluster in the negative group
included IFN-y and GRO/KC, which are significant
cytokines in the UTI and Struvite groups, respectively.
Except for animals that were inoculated with log 9 CFU of
U. parvum, the expression pattern of IFN-y and GRO/KC
was not inversely related as they were in culture positive
animals. This may be reflecting a more balanced immune
response than what is seen in U. parvum positive animals,
and we suggest that this balanced response may be critical
to resolution of infection and prevention of severe dis-
ease.

The variable clinical outcome to experimental inoculation
with U. parvum in the F344 rat is an interesting phenome-
non since this is an inbred strain. In this study, both
genetic and environmental influences on disease were
minimized to the extent possible. All of the animals in
this study originated from the same colony. Further, rats
were housed under the same barrier maintained condi-
tions in order to minimize environmental variability.
Despite our efforts, it was common to find that a rat that
developed asymptomatic UTI had co-habited the same
cage with a rat that developed struvites or was culture neg-
ative at time of necropsy. Therefore, external environment
could not account for the varying clinical outcome in our
study. However, our experimental inoculation procedure
may be a critical source of variability. Although attempts
were made to reduce mechanical trauma caused by cathe-
rization, it is possible that the trauma may have been suf-
ficient to shift the immune response towards a pro-
inflammatory profile in a subset of animals. Once this
occurred, the pro-inflammatory cycle progressed until
infection was resolved (Negative group) or the study was
terminated (Struvite group). Another possible explana-
tion for our findings may involve the actual placement of
ureaplasmas within the urinary tract at time of inocula-
tion. For example, if the catheter disrupted the uroepithe-
lial barrier so that a sufficient number of ureaplasmas
were deposited into the submucosa instead of the
mucosal surface, this could elicit a different inflammatory
response cascade than what would normally occur if the
microorganisms were only present on the mucosal surface
of the bladder. The results of this study were similar to
what we have previously reported, thus showing the con-
sistency and reproducibility of this model of infection.
Moreover, the clinical outcome to ureaplasmal infection
in the F344 rat is similar to what occurs in humans. The


0.'-~-- -


1500-

E
0) 1000-


-3 500-


T


BMC Infectious Diseases 2009, 9:9







http://www.biomedcentral.com/1471-2334/9/9


5000-



S2500-

0
0-
3000-


-- 2000-
0.
Ca
S1000-
.-J

0-
3000-



0.

-J





1W

I-
0)

00
-j


400-

E 300-
0)
200-

Z 100-

0-


p<0.0001


p<0.0001


p<0.0012


p<0.012









p<0.001


Control Negative UTI Urolith


Figure 5
Urine cytokines detected in F344 rats inoculated
with sterile I OB or U. parvum. Data represent the mean
SD of a combination of 5 separate experiments. Urine
cytokine concentrations were grouped according to clinical
profile, control (n = 6), negative (n = 36), UTI (n = 16), and
Struvite (n = 13). P values within each graph were obtained
by one-way ANOVA. Fisher's PLSD test revealed that IL-Ia,
IL-1 P, IL-6, and TNF-a concentrations in the struvite group
were significantly greater than control, negative, and UTI
groups. Fisher's PLSD test revealed IL-10 concentrations in
the struvite group were significantly greater than negative
and UTI groups.


complex interactions between most mucosal pathogens
and the host that lead to uncomplicated colonization ver-
sus inflammation and disease are largely unknown. There-
fore, this model may be particularly useful for identifying
the molecular events that confer asymptomatic infection,
complicated infection as well as resolution of infection
with an opportunistic pathogen of the urogenital tract.

Conclusion
The complex interactions between Ureaplasma and the
host that lead to uncomplicated colonization versus
inflammation and disease are largely unknown. We char-
acterized the F344 rat immune response in the urinary
tract to varying inoculum concentrations of U. parvum.
Establishment of UTI was influenced by microbial load,
but the host immune response was independent of micro-
bial load. Two distinct innate immune profiles were iden-
tified with two different clinical outcomes: asymptomatic
UTI and complicated UTI with struvite formation. Asymp-
tomatic UTI was characterized by a minimal immune
response that was predominantly monocytic and lym-
phocytic and was accompanied by a significantly high rate
of kidney infection. UTI complicated with struvite forma-
tion was characterized by an exaggerated immune
response that was predominantly neutrophilic and was
accompanied by uroepithelial hyperplasia and extensive
tissue damage.

Competing interests
The authors declare that they have no competing interests.


Control

UTI
Struvite
KiIA^*LJ<


-2.521
-2016
-1.512
.1 008
-0.504
0

1.0082
1.5123
2 0164
2 5205


Figure 6
Global profiling of urine cytokines detected in con-
trol and U. parvum infected F344 rats. The clustered
heat map represents the standardized LS means for each
cytokine that had a significantly different pattern of expres-
sion among clinical groups (P _< 0.02). Values were obtained
by one-way ANOVA using a row by row modeling with
Fischer C correction for multiple comparisons. Two main
cytokine cluster patterns were identified in the analysis and
are demarcated by the green and red cluster tree. The
number of biological replicates were n = 6 for control, n = 16
for UTI group, n = 13 for the Struvite group.


Page 10 of 12
(page number not for citation purposes)


BMC Infectious Diseases 2009, 9:9








http://www.biomedcentral.com/1471-2334/9/9


Log 1
Control
Log 5
Log 7
Log 3
Log 9 ....
*v h.<- A


-1.777

_1.6w6
-1 .111
-0.555
4e- 16
0.5555
1.111
1.6685
12222
2.7775


Figure 7
Profiling the inflammatory response to different
doses of U. parvum in culture negative F344 rats. Panel
A is a clustered heat map representing the standardized LS
means for each cytokine with a significantly different pattern
of expression among infection dose groups (P 0.05). Values
were obtained by one-way ANOVA using a row by row
modeling with Fischer C correction for multiple compari-
sons. Two main cytokine cluster patterns were identified in
the analysis and are demarcated by the green and red cluster
tree. The number of biological replicates were n = 6 for con-
trol, n = 10 for log I CFU, n = I I for log 3 CFU, n = 8 for log
5 CFU, n = 5 for log 7 CFU, and n = 3 for log 9 CFU. The red
arrow is highlighting the pattern of cytokines present in the
urine of culture negative rats that were inoculated with log 5
CFU. Animals within this dose group were the only animals
to exhibit an obvious inflammatory cell infiltrate comprising a
mixture of mononuclear cells with neutrophils (P 0.006, Fig-
ure 2, Panel D).


Authors' contributions
LR designed and executed animal infection studies, data
analysis and manuscript preparation. MR performed his-
topathologic evaluation of tissues and developed the
lesion scoring system implemented in this study. MBB
participated in the design and coordination of the study,
and helped draft the manuscript. All authors read and
approved the final manuscript.

Additional material


Additional file 1
Distribution of kidney lesion scores in F344 rats inoculated with var-
ying doses of U. parvum. Lesion score analysis of kidney tissue was based
on total area ii,.. ... The data provided represents the raw lesion scores
for each biological replicate. Horizontal bars demarcate the median value
for each clinical profile group. Lesion scoring of each sample was per-
formed without prior knowledge of the clinical profile. Data is a summa-
tion of5 separate experiments. Nonparametric lesion scores were grouped
according to the clinical profile and analyzed by Kruskall-Wallis test
(Control, n 6; Negative, n 36; UTI, n 11; Struvite, n 10).
Click here for file
[http://www.biomedcentral.com/content/supplementary/1471-
2334-9-9-S1.rtf]


Acknowledgements
This work was supported by the National Institutes of Health grants RO I
A145875 and K08 DK07565 I1. We also wish to acknowledge Ms. Janet Ste-
vens and LJ McDonnell for their assistance with microbial cultures and
cytokine immunoassays.

References
1. Schelonka RL, Waites KB: Ureaplasma infection and neonatal
lung disease. Semin Perinatol 2007, 3 1:2-9.
2. Volgmann T, Ohlinger R, Panzig B: Ureaplasma urealyticum-
harmless commensal or underestimated enemy of human
reproduction? A review. Archives of gynecology and obstetrics 2005,
273:133-139.
3. Skerk V, Marekovic I, Markovinovic L, Begovac J, Skerk V, Barsic N,
Majdak-Gluhinic V: Comparative randomized pilot study of azi-
thromycin and doxycycline efficacy and tolerability in the
treatment of prostate infection caused by Ureaplasma urea-
lyticum. Chemotherapy 2006, 52:9-1 I.
4. Hedelin H: Uropathogens and urinary tract concretion forma-
tion and catheter encrustations. International journal of antimicro-
bial agents 2002, 19:484-487.
5. Viscardi RM, Atamas SP, Luzina IG, Hasday JD, He JR, Sime PJ, Coal-
son JJ, Yoder BA: Antenatal Ureaplasma urealyticum respira-
tory tract infection stimulates proinflammatory, profibrotic
responses in the preterm baboon lung. Pediatric research 2006,
60:141-146.
6. Viscardi RM, Kaplan J, LovchikJC, He JR, Hester L, Rao S, Hasday JD:
Characterization of a marine model of Ureaplasma urealyti-
cum pneumonia. Infect Immun 2002, 70:5721-5729.
7. Pickering WJ, Birch DF, Kincaid-Smith P: Biochemical and histo-
logic findings in experimental pyelonephritis due to Urea-
plasma urealyticum. Infect Immun 1990, 58:3401-3406.
8. Grenabo L, Brorson JE, Hedelin H, Pettersson S: Concrement for-
mation in the urinary bladder in rats inoculated with Urea-
plasma urealyticum. Urological research 1985, 13:195-198.
9. Larsson PA, Cano M, Grenabo L, Brorson JE, Hedelin H, Pettersson
S, Johansson SL: Morphological lesions of the rat urinary tract
induced by inoculation of mycoplasmas and other urinary
tract pathogens. Urologia internationalis 1989, 44:210-217.
10. Goldenberg RL, CulhaneJF, lamsJD, Romero R: Epidemiology and
causes of preterm birth. Lancet 2008, 371:75-84.
I I. Goncalves LF, Chaiworapongsa T, Romero R: Intrauterine infec-
tion and prematurity. Mental retardation and developmental disabil-
ities research reviews 2002, 8:3-13.
12. Viscardi R, Manimtim W, He JR, Hasday JD, Sun CC, Joyce B, Pierce
RA: Disordered pulmonary myofibroblast distribution and
elastin expression in preterm infants with Ureaplasma urea-
lyticum pneumonitis. Pediatr Dev Pathol 2006, 2:143-15 1.
13. Reyes L, Reinhard M, O'Donell LJ, Stevens J, Brown MB: Rat strains
differ in susceptibility to Ureaplasma parvum-induced uri-
nary tract infection and struvite stone formation. Infect Immun
2006, 74:6656-666.
14. Whiteland JL, Shimeld C, Nicholls SM, Easty DL, Williams NA, Hill TJ:
Immunohistochemical detection of cytokines in paraffin-
embedded mouse tissues. j Immunol Methods 1997, 210:103-108.
15. Taylor-Robinson D, Csonka GW, Prentice MJ: Human intra-ure-
thral inoculation of Ureaplasmas. QJ Med 1977, 46:309-26.
16. Dinarello CA, Fantuzzi G: Interleukin-18 and host defense
against infection. The journal of infectious diseases 2003, 187(Suppl
2):S370-384.
17. Shtrichman R, Samuel CE: The role of gamma interferon in anti-
microbial immunity. Current opinion in microbiology 2001,
4(3):25 1-259.
18. Yoo JK, Kwon H, Khil LY, Zhang L, Jun HS, Yoon JW: IL-I 18 induces
monocyte chemotactic protein-I production in macro-
phages through the phosphatidylinositol 3-kinase/Akt and
MEKIERKI/2 pathways. j Immunol 2005, 175(12):8280-8286.
19. MaJ, Chen T, Mandelin J, Ceponis A, Miller NE, Hukkanen M, Ma GF,
Konttinen YT: Regulation of macrophage activation. Cell Mol
Life Sci 2003, 60(1 1):2334-2346.
20. Mulvey MA, Schilling JD, Martinez JJ, Hultgren SJ: Bad bugs and
beleaguered bladders: interplay between uropathogenic
Escherichia coli 2000. and innate host defenses. Proc Natl Acad
Sci 2000, 97(16):8829-35.



Page 11 of 12
(page number not for citation purposes)


BMC Infectious Diseases 2009, 9:9








BMC Infectious Diseases 2009, 9:9


http://www.biomedcentral.com/1471-2334/9/9


21. Narayanan S, Glasser A, Hu YS, McDermott AM: The effect of
interleukin-I on cytokine gene expression by human corneal
epithelial cells. Experimental eye research 2005, 80(2): 175-183.
22. Son DS, Parl AK, Rice VM, Khabele D: Keratinocyte chemoat-
tractant (KC)/human growth-regulated oncogene (GRO)
chemokines and pro-inflammatory chemokine networks in
mouse and human ovarian epithelial cancer cells. Cancer biol-
ogy & therapy 2007, 6(8): 1302-13 12.
23. Warhurst AC, Hopkins SJ, Warhurst G: Interferon gamma
induces differential upregulation of alpha and beta chemok-
ine secretion in colonic epithelial cell lines. Gut 1998,
42(2):208-213.

Pre-publication history
The pre-publication history for this paper can be accessed
here:

httn-//nAwwwn hinmPdcrpntral mm / 1471 -)9 4 /)/)/nrPnnih


Page 12 of 12
(page number not for citation purposes)


Publish with BioMed Central and every
scientist can read your work free of charge
"BioMed Central will be the most significant development for
disseminating the results of biomedical research in our lifetime."
Sir Paul Nurse, Cancer Research UK
Your research papers will be:
available free of charge to the entire biomedical community
peer reviewed and published immediately upon acceptance
cited in PubMed and archived on PubMed Central
yours you keep the copyright
Submit your manuscript here: BioMedcentral
http://www.biomedcentral.com/info/publishing adv.asp




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs