Article

The Influence of Developmental Age on the Early Transcriptomic Response of Children with Septic Shock

Department of Pediatrics, Duke University School of Medicine, Durham, North Carolina, United States of America.
Molecular Medicine (Impact Factor: 4.51). 07/2011; 17(11-12):1146-56. DOI: 10.2119/molmed.2011.00169
Source: PubMed

ABSTRACT

Septic shock is a frequent and costly problem among patients in the pediatric intensive care unit (PICU) and is associated with high mortality and devastating survivor morbidity. Genome-wide expression patterns can provide molecular granularity of the host response and offer insight into why large variations in outcomes exist. We derived whole-blood genome-wide expression patterns within 24 h of PICU admission from children with septic shock. We compared the transcriptome between septic shock developmental-age groups defined as neonates (≤ 28 d, n = 17), infants (1 month to 1 year, n = 62), toddlers (2-5 years, n = 54) and school-age (≥ 6 years, n = 47) and age-matched controls. Direct intergroup comparisons demonstrated profound changes in neonates, relative to older children. Neonates with septic shock demonstrated reduced expression of genes representing key pathways of innate and adaptive immunity. In contrast to the largely upregulated transcriptome in all other groups, neonates exhibited a predominantly downregulated transcriptome when compared with controls. Neonates and school-age subjects had the most uniquely regulated genes relative to controls. Age-specific studies of the host response are necessary to identify developmentally relevant translational opportunities that may lead to improved sepsis outcomes.

Full-text

Available from: Thomas P Shanley, Mar 10, 2014
1146 | WYNN ET AL. | MOL MED 17(11-12)1146-1156, NOVEMBER-DECEMBER 2011
INTRODUCTION
Sepsis is a common and deadly condi-
tion that occurs in all patient age groups
requiring intensive care. Survival and
outcomes among children that develop
septic shock are poor (1). Bacterial sepsis
of the newborn is the seventh leading
cause of infant death in the United States
(2), and infection kills >1 million new-
borns worldwide annually (3).
Multiple developmental alterations in
the host innate and adaptive immune re-
sponses highlight the age-related differ-
ences in the capacity to effectively re-
spond to a sepsis challenge (4,5).
Consequently, adjunctive sepsis thera-
pies that prove useful in adults and older
children may have little effect, or even
completely lack biological plausibility, in
less immunologically mature popula-
tions. Thus, clarification of the age-spe-
cific host response to sepsis is critically
important to identify age-appropriate
therapeutic strategies.
Unbiased genome-wide expression
patterns are increasingly used to im-
prove understanding of complex, hetero-
geneous diseases that have large varia-
tions in host response and outcomes. We
and others have used this approach in
children with septic shock to successfully
identify mRNA expression patterns that
enhance diagnostic accuracy, predict sep-
sis severity, stratify disease and identify
novel signaling pathways (6–10).
We now show for the first time that
significant differences in gene expression
exist between developmental-age groups
The Influence of Developmental Age on the Early
Transcriptomic Response of Children with Septic Shock
James L Wynn,
1
Natalie Z Cvijanovich,
2
Geoffrey L Allen,
3
Neal J Thomas,
4
Robert J Freishtat,
5
Nick Anas,
6
Keith Meyer,
7
Paul A Checchia,
8
Richard Lin,
9
Thomas P Shanley,
10
Michael T Bigham,
11
Sharon Banschbach,
12
Eileen Beckman,
12
and Hector R Wong
12
1
Department of Pediatrics, Duke University School of Medicine, Durham, North Carolina, United States of America;
2
Children’s
Hospital and Research Center Oakland, Oakland, California, United States of America;
3
Children’s Mercy Hospital, Kansas City,
Missouri, United States of America;
4
Penn State Children’s Hospital, Hershey, Pennsylvania, United States of America;
5
Children’s
National Medical Center, Washington, DC, United States of America;
6
Children’s Hospital of Orange County, Orange, California,
United States of America;
7
Miami Children’s Hospital, Miami, Florida, United States of America;
8
St. Louis Children’s Hospital, St.
Louis, Missouri, United States of America;
9
The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, United States of
America;
10
CS Mott Children’s Hospital at the University of Michigan, Ann Arbor, Michigan, United States of America;
11
Akron
Children’s Hospital, Akron, Ohio, United States of America; and
12
Cincinnati Children’s Hospital Medical Center and Cincinnati
Children’s Research Foundation, Department of Pediatrics, University of Cincinnati College of Medicine, Cincinnati, Ohio, United
States of America
Septic shock is a frequent and costly problem among patients in the pediatric intensive care unit (PICU) and is associated with
high mortality and devastating survivor morbidity. Genome-wide expression patterns can provide molecular granularity of the host
response and offer insight into why large variations in outcomes exist. We derived whole-blood genome-wide expression patterns
within 24 h of PICU admission from children with septic shock. We compared the transcriptome between septic shock develop-
mental-age groups defined as neonates (28 d, n = 17), infants (1 month to 1 year, n = 62), toddlers (2–5 years, n = 54) and school-
age (6 years, n = 47) and age-matched controls. Direct intergroup comparisons demonstrated profound changes in neonates,
relative to older children. Neonates with septic shock demonstrated reduced expression of genes representing key pathways of
innate and adaptive immunity. In contrast to the largely upregulated transcriptome in all other groups, neonates exhibited a pre-
dominantly downregulated transcriptome when compared with controls. Neonates and school-age subjects had the most
uniquely regulated genes relative to controls. Age-specific studies of the host response are necessary to identify developmentally
relevant translational opportunities that may lead to improved sepsis outcomes.
© 2011 The Feinstein Institute for Medical Research, www.feinsteininstitute.org
Online address: http://www.molmed.org
doi: 10.2119/molmed.2011.00169
Address correspondence and reprint requests to Hector R. Wong, Division of Critical
Care Medicine-MLC 2005, Cincinnati Children’s Hospital Medical Center, 3333 Burnet
Avenue, Cincinnati, OH, 45229. Phone: 513-636-4259; Fax: 513-636-4267; E-mail:
hector.wong@cchmc.org.
Submitted May 9, 2011; Accepted for publication June 9, 2011; Epub (www.molmed.org)
ahead of print July 5, 2011.
Page 1
RESEARCH ARTICLE
MOL MED 17(11-12)1146-1156, NOVEMBER-DECEMBER 2011 | WYNN ET AL. | 1147
of children with septic shock, particu-
larly within the neonatal group. Further-
more, the unique neonatal response we
describe herein raises the question of
whether adjunctive sepsis therapies that
may be successful in older populations
will retain utility or even pose increased
risks in neonates.
MATERIALS AND METHODS
Patients and Data Collection
The study protocol was approved by
the Institutional Review Boards of each
participating institution (n = 11). Chil-
dren ≤10 years of age admitted to the pe-
diatric intensive care unit (PICU) and
meeting pediatric-specific criteria for
septic shock were eligible for enrollment
(11). Age-matched controls were re-
cruited from the ambulatory depart-
ments of participating institutions using
published inclusion and exclusion crite-
ria (10). All patients and controls were
previously reported in microarray-based
studies addressing hypotheses entirely
different from those of the current report
(6,8–10,12). All microarray data were de-
posited in the National Center for
Biotechnology Information (NCBI) Gene
Expression Omnibus (accession numbers
GSE26440 and GSE26378). The patients
in this study cohort were recruited be-
tween March 2003 and June 2010.
After informed consent from parents
or legal guardians, blood samples were
obtained within 24 h of initial presenta-
tion to the PICU with septic shock.
Clinical and laboratory data were col-
lected daily while in the PICU and
stored using a Web-based database.
Organ failure was defined using pedi-
atric-specific criteria and tracked up to
the first 7 d of PICU admission (11).
Mortality was tracked for 28 d after en-
rollment. The developmental-age cate-
gories used in this analysis are as fol-
lows: neonate (≤28 d of age), infant
(1 month through 1 year of age), tod-
dler (2–5 years of age), and school-age
(≥6 years of age) (11). All patients in the
neonate group were products of full-
term gestations.
RNA Extraction and Microarray
Hybridization
Total RNA was isolated from whole
blood using the PaxGene™ Blood RNA
System (PreAnalytiX; Qiagen/Becton
Dickson, Valencia, CA, USA). Microarray
hybridization was performed as previ-
ously described using the Human Ge-
nome U133 Plus 2.0 GeneChip
(Affymetrix, Santa Clara, CA, USA)
(6,8–10,12).
Data Analysis
Analyses were performed using one
patient sample per chip, and CEL files
were preprocessed using Robust Multiple-
Array Average (RMA) normalization and
GeneSpring GX 7.3 software (Agilent
Technologies, Palo Alto, CA, USA). All
signal intensity–based data were used
after RMA normalization, which specifi-
cally suppresses all but significant varia-
tion among lower-intensity probe sets
(13). All chips representing patient sam-
ples were then normalized to the respec-
tive median values of controls.
Differences in mRNA abundance be-
tween the developmental age categories
were measured by sequential expression
and statistical filters using GeneSpring
GX 7.3. For direct comparisons across the
four developmental-age groups of pa-
tients with septic shock, we used a two-
stage approach. In stage one, we applied
an expression filter to determine the
number of gene probes on the array
(>80,000 gene probes) having ≥two-fold
expression on the basis of all possible in-
tergroup comparisons. In the second
stage, we conducted a four-group analy-
sis of variance (ANOVA) with a
Benjamini-Hochberg false discovery rate
of 1% to determine how many of the
gene probes identified in stage one were
differentially regulated among the four
groups.
We also compared gene expression be-
tween patients with septic shock from
each of the four respective developmental-
age groups and normal age-matched
controls. This analysis also occurred in
two stages: more than twofold expres-
sion filter followed by ANOVA with a
Benjamini-Hochberg false discovery rate
of 1% for each of the four developmental-
age groups and normal controls.
Gene lists of differentially expressed
genes were analyzed using the ingenu-
ity pathways analysis (IPA) application
(Ingenuity Systems, Redwood City, CA,
USA), which provides a tool for discov-
ery of signaling pathways within the
uploaded gene lists as previously de-
scribed (12,14). Gene expression mosaics
representing the expression patterns of
differentially regulated genes were gen-
erated using the Gene Expression Dy-
namics Inspector (GEDI) (15–17). The
signature graphic outputs of GEDI are
expression mosaics that give microarray
data a “face” that is intuitively recogniz-
able via human pattern recognition. Ad-
ditional technical details regarding
GEDI can be found at http://www.
childrenshospital.org/research/ingber/
GEDI/gedihome.htm.
Ordinal and continuous clinical vari-
ables not normally distributed were ana-
lyzed via ANOVA on ranks. Dichoto-
mous clinical variables were analyzed
using a χ
2
test (SigmaStat Software; Sys-
tat Software, San Jose, CA, USA).
All supplementary materials are available
online at www.molmed.org.
RESULTS
Demographics and Clinical
Characteristics of the Developmental
Age Groups
Table 1 provides the demographic
and clinical characteristics of the four
developmental-age groups. The neonate
group had a higher mortality rate and
Pediatric Risk of Mortality (PRISM) score
than the other three groups. The neonate
group also had a higher number of maxi-
mal organ failures than the toddler and
school-age groups. In contrast, the
neonate group had a lower proportion of
subjects with comorbidities than the tod-
dler and school-age groups. The neonate
group had a higher proportion of infec-
tions with gram- positive bacteria than
the school-age group and a lower pro-
Page 2
1148 | WYNN ET AL. | MOL MED 17(11-12)1146-1156, NOVEMBER-DECEMBER 2011
DEVELOPMENTAL DIFFERENCES IN THE RESPONSE TO SEPTIC SHOCK
portion of infections with gram-negative
organisms than the infant and toddler
groups. There were a variety of signifi-
cant differences between the develop-
mental-age groups with respect to pe-
ripheral white blood cell counts.
Direct Comparison of Gene
Expression Across the Four
Developmental Age Groups
In this analysis, we directly compared
gene expression across the four
developmental-age groups of patients
with septic shock as described in Materi-
als and Methods. Table 2 provides the
number of gene probes having ≥two-fold
expression on the basis of all possible in-
tergroup comparisons. The number of
gene probes meeting the expression crite-
ria increased in proportion to age differ-
ences. For example, the comparison be-
tween the neonate group and the school-
age group yielded more than three times
the number of gene probes relative to the
comparison between the neonate group
and the infant group. In contrast, the
comparison between the infant group
and the toddler group yielded no gene
probes meeting the expression criteria.
The statistical test applied in stage II of
this analysis yielded 1,638 gene probes
differentially regulated between the four
developmental-age groups. The top 100
(on the basis of P values) differentially
expressed unique and well-annotated
genes from these 1,638 gene probes are
listed in Supplementary Table 1.
To derive a global view of the respec-
tive gene expression patterns, we up-
loaded the expression data for the 1,638
gene probes to the GEDI platform and
generated gene expression mosaics for
each developmental-age group. Figure 1A
provides the average mosaic patterns for
each group and demonstrates that the
neonatal group had the most distinct
global gene expression pattern.
To extract more specific biological in-
formation from the 1,638 gene probes, we
uploaded the entire list of gene probes to
the IPA platform and focused the data
Table 2. The number of gene probes
having two-fold expression difference
between all possible developmental-age
group comparisons among patients with
septic shock.
Infant Toddler School-age
Neonate 410 1,057 1,498
Infant 0 40
Toddler 9
Table 1. Demographic and clinical characteristics of the four developmental-age groups.
Neonate Infant Toddler School-age
Number of patients 17 62 54 47
Age (years) 0.1 (0.0–0.1) 1.0 (0.7–1.5) 3.0 (2.4–4.4) 8.6 (7.3–9.3)
Males/females (n) 13/4 37/25 29/25 30/17
Deaths [n (%)] 8 (47)
a
8 (13) 6 (11) 7 (15)
PRISM score 24 (14–35)
b
16 (8–22) 15 (10–19) 11 (10–17)
Maximum number of organ failure
c
3 (3–4)
d
2 (2–3) 2 (2–3) 2 (2–3)
Days to death from presentation 2 (1.5–5) 7 (5–14.5) 2 (1–9) 2 (1–10)
White blood cell count × 10
3
/mm
3
3.5 (2.3–9.9)
b
16.1 (7.2–22.8) 13.4 (7.1–17.6) 13.2 (7.8–18.2)
Neutrophil count × 10
3
/mm
3
0.9 (0.4–4.9)
b
11.4 (4.7–18.5) 8.1 (3.9–14.2) 9.4 (3.4–13.8)
% Neutrophils 45 (30–55)
b
70 (58–79) 73 (58–83) 77 (69–87)
Lymphocyte count × 10
3
/mm
3
1.3 (0.8–3.0) 2.3 (1.1–4.5)
e
2.3 (0.9–3.8) 1.3 (0.6–2.3)
% Lymphocytes 46 (34–59)
b
19 (8–31) 18 (10–31) 11 (6–18)
Monocyte count × 10
3
/mm
3
0.2 (0.0–0.3)
f
0.8 (0.4–1.5) 0.6 (0.3–1.4) 0.3 (0.1–0.9)
% Monocytes 6 (2–9) 6 (3–8) 7 (3–10) 4 (2–8)
Number with gram-positive bacteria (%) 8 (47)
g
18 (29) 15 (28) 6 (13)
Number with gram-negative bacteria (%) 0 (0)
h
18 (29) 17 (31) 11 (23)
Number with other organism (%)
i
2 (12) 5 (8) 4 (7) 5 (11)
Number with negative cultures (%) 7 (41) 21 (34) 18 (33) 25 (53)
Number with bacteremia (%) 8 (47) 23 (37) 20 (37) 9 (19)
Number with comorbidities (%) 2 (12)
d
25 (40) 26 (48) 23 (49)
All data are median (interquartile range) unless otherwise noted.
a
P
< 0.05 versus infant, toddler and school-age (χ
2
).
b
P
< 0.05 versus infant, toddler and school-age (ANOVA on ranks).
c
Refers to maximal number of organ failures over the first 7 d of admission.
d
P
< 0.05 versus toddler and school-age.
e
P
< 0.05 versus school-age (ANOVA on ranks).
f
P
< 0.05 versus infant and toddler (ANOVA on ranks).
g
P
< 0.05 versus school-age (χ
2
).
h
P
< 0.05 versus infant and toddler (χ
2
).
i
Refers to viral or fungal pathogens.
Page 3
RESEARCH ARTICLE
MOL MED 17(11-12)1146-1156, NOVEMBER-DECEMBER 2011 | WYNN ET AL. | 1149
output on enrichment for genes corre-
sponding to signaling pathways directly
related to inflammation and immune
function. Table 3 provides the results of
this analysis and demonstrates enrich-
ment for genes corresponding to signal-
ing pathways highly relevant to sepsis bi-
ology. We next extracted the normalized
expression values for each of the genes
corresponding to the signaling pathways
in Table 3 and calculated the respective
median relative gene expression values
for each of the four developmental-age
groups (Figure 1B). The neonatal group
had significantly lower expression of the
genes corresponding to all six signaling
pathways compared with the other three
developmental-age groups.
To provide more detail regarding the
signaling pathway data provided in Fig-
ure 1B, we generated pathway- specific
diagrams that depict gene expression by
red (upregulation) and green (downreg-
ulation) node coloring. We limited these
comparisons to the neonate and school-
age groups because they had the largest
degree of variation. Figure 1C shows
gene expression corresponding to the
nuclear factor (NF)-κB pathway and
demonstrates generalized downregula-
tion of NF-κB pathway–related genes
in the neonate group. Similar diagrams
for the pathogen recognition receptor
and triggering receptor expressed on
myeloid cells-1 (TREM-1) pathways
are shown in Supplementary Figures 1
and 2, respectively.
Collectively, this analysis based on
direct comparison of gene expression
across the four developmental-age
groups demonstrates that developmen-
tal age strongly influences the early
whole blood transcriptomic response
during pediatric septic shock. The de-
gree of differential gene expression in-
creases in proportion to the difference
between developmental ages. In addi-
tion, neonatal responses are character-
ized by decreased expression of genes
corresponding to key inflammation-
and immunity-related signaling path-
ways, relative to the responses of older
children.
Gene Expression Patterns Relative to
Controls
A direct comparison across the four
developmental-age groups with septic
shock has the potential to overlook im-
portant gene expression profiles that
more directly reflect perturbations from a
normal state. Accordingly, we also con-
ducted an analysis in which we com-
pared gene expression between patients
with septic shock from each of the four
respective developmental-age groups
and normal age-matched controls, as de-
scribed in Materials and Methods.
Relative to controls, neonates had a sig-
nificantly greater proportion of downreg-
ulated gene probes than the three other
groups, and the school-age group had the
largest total number of differentially ex-
pressed genes compared with controls
(Table 4). To broadly compare gene ex-
pression patterns for each developmental-
age group, relative to controls, we con-
structed Venn diagrams of all possible
group comparisons (Figure 2). The Venn
diagrams demonstrate that between 805
and 1,408 gene probes were common
across any one of the four possible group
comparisons. In addition, the Venn dia-
grams demonstrate that in all of the com-
parisons, either the neonate group or the
school-age group had the largest number
of uniquely regulated genes.
We next uploaded the individual lists
of upregulated and downregulated genes
in Table 4 to the IPA application and
again focused the analytical output on
enrichment for genes corresponding to
inflammation- and immunity-related sig-
Figure 1. (A) GEDI-generated gene expression mosaics for each of the developmental-
age groups. Each mosaic represents the average expression pattern of the same 1,638
gene probes from patients with septic shock in each of the four developmental-age
groups. The degree of red intensity correlates with increased gene expression and the de-
gree of blue intensity correlates with decreased gene expression. (B) Relative expression
of genes corresponding to the indicated signaling pathways among patients with septic
shock from each developmental-age group. Data are expressed as medians with in-
terquartile ranges and were analyzed using ANOVA on ranks. *
P
< 0.05 versus infant, tod-
dler and school-age groups.
P
< 0.05 versus school-age group. (C) Relative expression of
genes corresponding to the NF-κB pathway in neonates and school-age children with
septic shock. An illustrative example of expression data is shown in B. Greater intensity of
the gene node coloring represents greater change in expression (green: downregulated;
red: upregulated).
Page 4
1150 | WYNN ET AL. | MOL MED 17(11-12)1146-1156, NOVEMBER-DECEMBER 2011
DEVELOPMENTAL DIFFERENCES IN THE RESPONSE TO SEPTIC SHOCK
Figure 1.
Continued.
Page 5
RESEARCH ARTICLE
MOL MED 17(11-12)1146-1156, NOVEMBER-DECEMBER 2011 | WYNN ET AL. | 1151
naling pathways. Figure 3 provides the
top signaling pathways represented by
the upregulated and downregulated
genes for each of the four developmental-
age groups. For the upregulated genes,
the level of significance generally in-
creased for each signaling pathway in
proportion to increasing developmental-
age group. In addition, for all of the up-
regulated signaling pathways, except in-
terleukin (IL)-8, the level of significance
was lowest for the neonate group. To fur-
ther illustrate these differences in gene
expression, we generated a pathway-
specific diagram that depicts the degree
and number of upregulated genes corre-
sponding to the IL-10 pathway for the
neonate and school-age groups, respec-
tively (Figure 4). Figure 4 further illus-
trates that the neonate group had a sub-
stantially lower number of upregulated
genes corresponding to the IL-10 path-
way, compared with that of the school-
age group.
The downregulated genes were highly
enriched for signaling pathways corre-
sponding to adaptive immunity (Figure
3). The level of significance for each sig-
naling pathway was generally higher for
the neonate group, indicating that the
neonate group had a proportionally
larger number of downregulated genes
corresponding to adaptive immunity sig-
naling. To further illustrate these differ-
ences in gene expression, we generated a
pathway-specific diagram that depicts
the degree and number of downregu-
lated genes corresponding to the antigen
presentation pathway for the neonate
and school-age groups, respectively (Fig-
ure 5). Figure 5 further illustrates that the
neonate group had a substantially higher
number of downregulated genes corre-
sponding to the antigen presentation
pathway, compared with that of the
school-age group.
Collectively, this analysis comparing
each individual developmental-age
group to controls further demonstrates
that developmental age strongly influ-
ences the early whole-blood transcrip-
tomic response during pediatric septic
shock. While the groups shared some
common patterns of gene expression, the
two extremes of developmental-age
groups in this cohort (that is, neonate
and school-age) had a relatively large
number of uniquely regulated gene sets.
Among the upregulated genes that corre-
spond to inflammation- and immunity-
related signaling pathways, the propor-
tion of genes that were upregulated for a
given pathway increased in proportion
to developmental age. Notably, the
downregulated genes corresponded to
adaptive immunity-related signaling
pathways, and the neonate group tended
to have the highest proportion of down-
regulated genes corresponding to adap-
tive immunity.
DISCUSSION
This study represents the first
developmental-age group comparison of
the transcriptomic response of children
with septic shock. We show that devel-
opmental age strongly influences the
early whole blood transcriptomic re-
sponse. This assertion is supported by
direct comparisons of patients with sep-
tic shock across four developmental-age
groups and by comparisons between the
respective developmental-age groups
and age-matched controls. The direct
comparisons demonstrated minimal dif-
ferences between the infant, toddler and
school-age groups with septic shock. In
contrast, age-specific alterations in host
Table 3. Top inflammation- and immunity-related signaling pathways represented in the
1,638 gene probes differentially regulated between patients with septic shock in the four
developmental-age groups.
Signaling pathway
P
Number of genes
B-cell receptor signaling 2.1x10
–8
27
TREM1 signaling 1.5x10
–7
15
Pattern recognition receptor signaling 3.6x10
–7
17
NF-κB signaling 1.8x10
–6
25
Dendritic cell maturation 1.6x10
–5
22
Communication between innate and adaptive immunity 1.7x10
–4
13
Table 4. Differential gene expression between controls and patients with septic shock from
each of the respective developmental-age groups.
Upregulated Downregulated
gene probes gene probes
(relative to controls) (relative to controls) Total gene probes
Neonate 599 1,224 1,823
Infant 930 636 1,566
Toddler 1,184 745 1,929
School-age 1,632 1,286 2,918
Figure 2. Total differentially regulated
genes between patients with septic shock
in each developmental-age category and
age-matched controls from across all pos-
sible group comparisons. Venn diagrams
represent differential gene expression be-
tween age-matched controls and patients
with septic shock from each of the respec-
tive developmental-age groups.
Page 6
1152 | WYNN ET AL. | MOL MED 17(11-12)1146-1156, NOVEMBER-DECEMBER 2011
DEVELOPMENTAL DIFFERENCES IN THE RESPONSE TO SEPTIC SHOCK
response were most profound in the
neonate group, as demonstrated by re-
duced expression of genes representative
of several key pathways of the innate
and adaptive immune systems.
In the comparisons to age-matched
controls, the neonate and school-age
groups had the largest number of
uniquely regulated genes. The upregu-
lated genes corresponded to several key
inflammatory/immunity pathways. Im-
portantly, the number of upregulated
genes corresponding to these pathways
increased in proportion to developmen-
tal age. In contrast, the downregulated
genes derived from these comparisons
corresponded to adaptive immunity-re-
lated pathways, and the number of
downregulated genes in each pathway
was greatest in the neonate group.
The innate immune system plays a
critical role in a successful host response
to sepsis, particularly in the neonate (18).
Multiple developmental alterations of in-
nate host response capabilities are pres-
ent in neonates compared with older age
groups, including pathogen recognition
receptors, inflammatory signaling path-
ways and overall innate immune cellular
function (19–23). Consistent with these
observations, our current data demon-
strate reduced expression of genes corre-
sponding to the pathogen recognition re-
ceptor and TREM-1 pathways, as well as
their relevant downstream signaling
molecules (for example, Janus kinase 2
[JAK2], signal transducer and activator of
transcription 5 [STAT5] and extracellular
signal regulated kinase 1/2 [ERK1/2]; Sup-
plementary Figure 2) in the neonate
group.
TREM-1 signaling is critical for ampli-
fication of the inflammatory responses to
microbial products in adults. Inhibition
of TREM-1 signaling through antibody-
mediated blockade reduced mortality in
septic adult animals and has been pro-
posed as a potential therapeutic target
for septic shock (24). Our current data in-
dicate that TREM-1 pathway-related
genes are not substantially expressed in
neonates with septic shock. Thus, block-
ade of TREM signaling may not be bio-
logically warranted in neonates. The no-
tion of a TREM-1–limited reduced
neonatal capacity to produce an intense
innate response to a septic challenge is
also supported by the attenuated inflam-
matory response seen in septic murine
neonates compared with septic young
adult mice (25). Taken together, these
data suggest that the neonate has a rela-
tively reduced capacity to generate as ro-
bust an innate immune response to sep-
tic shock as seen in older age groups,
which may be in part related to alter-
ations in TREM-1 signaling.
In stark contrast to the largely upregu-
lated transcriptomic responses from all
Figure 3. Top signaling pathways represented by the differentially regulated (upregulated
and downregulated) genes between patients with septic shock in each developmental-
age category versus age-matched controls. The
y
-axis is depicted as the –log(
P
value) and
provides an indication of how likely a gene list is enriched for a given pathway by chance
alone. The –log for a
P
value of 0.05 is ~1.3 and is indicated by the horizontal dashed line. In
contrast, the –log for a
P
value of 1.0x10
–8
is ~10. The level of significance for a given path-
way is directly proportional to the number of genes in a given gene list that correspond to
the pathway and indirectly proportional to the total number of genes in the list. PRR, pat-
tern recognition receptor; MAPK, mitogen-activated protein kinase; iCOS, inducible costim-
ulator; NFAT, nuclear factor of activated T cells; PKCθ, protein kinase C theta; Nur77, NR4A
nuclear receptor family member Nur77; CTLA-4, cytotoxic T lymphocyte antigen-4.
Page 7
RESEARCH ARTICLE
MOL MED 17(11-12)1146-1156, NOVEMBER-DECEMBER 2011 | WYNN ET AL. | 1153
Figure 4. Differential regulation of genes in the IL-10 pathway in neonates and school-age children with septic shock versus age-
matched controls. An illustrative example of the upregulated pathway shown in Figure 3 is demonstrated. The greater intensity of red
color represents a greater degree of upregulation in gene expression.
Page 8
1154 | WYNN ET AL. | MOL MED 17(11-12)1146-1156, NOVEMBER-DECEMBER 2011
DEVELOPMENTAL DIFFERENCES IN THE RESPONSE TO SEPTIC SHOCK
Figure 5. Differential regulation of genes in the antigen presentation pathway in neonates and school-age children with septic shock ver-
sus age-matched controls. An illustrative example of the downregulated pathway shown in Figure 3 is demonstrated. Greater intensity of
green color represents greater degree of downregulation in gene expression.
Page 9
RESEARCH ARTICLE
MOL MED 17(11-12)1146-1156, NOVEMBER-DECEMBER 2011 | WYNN ET AL. | 1155
three other developmental-age groups,
neonates exhibited predominantly down-
regulated responses when compared
with age-matched controls. These alter-
ations represented downregulated path-
ways related to adaptive immunity. It is
well known that baseline neonatal adap-
tive immune responses are distinct from
those seen in more mature populations.
These alterations in cellular response
have been suggested to permit avoidance
of perpetual hyperinflammation through
regulation of T-cell responses and in-
creased T-cell apoptosis (26).
This is the first report describing
downregulation of adaptive immunity-
related genes during septic shock in
neonates compared with age-matched
controls. The predominance of downreg-
ulated adaptive immune pathways in
neonates could be interpreted to support
why adaptive immune responses were
not critical for survival in an animal
model of neonatal polymicrobial sepsis
(27). In distinct contrast, the absence or
dysfunction of the adaptive immune sys-
tem has a profound negative impact on
adult survival in preclinical models (28)
and in humans (29,30). As these data il-
lustrate, the contribution of adaptive im-
munity for protection and response
against septic shock, and in particular
which components may be protective, is
unclear in neonates and requires further
investigation.
Many attempts have been made to im-
prove immune function in neonates and
reduce the incidence and burden of in-
fection (31). The failure of these interven-
tions in large randomized trials likely re-
flects underappreciated differences in the
functional capacity of the neonatal host
response (32). To successfully modify im-
mune function and improve infection
outcomes in human neonates, as has
been done in neonatal animal models
(27,33,34), consideration of the unique
immuno-developmental stage of the
neonate must be taken into account.
We acknowledge there are several po-
tential confounding factors in this study.
First, there were fewer neonatal patients
compared with the other age groups. To
address this potential confounder and
the associated risk of over-fitting the
data, we set the primary filter to require
a ≥two-fold increase in gene expression.
We then used a stringent statistical test
by setting a false discovery rate at 1%
(equivalent to a P value of 0.01).
Second, the neonatal group had a
higher mortality rate and a higher
PRISM score, thus raising the possibility
that the differences in gene expression
reflect a poorer physiologic state, rather
than differences reflecting developmental
age. To address this potential con-
founder, we calculated the median time
(d) to death for all nonsurvivors. There
was no significant difference in median
days to death (interquartile range) across
the four developmental-age groups:
neonate = 2 (1.5–5); infant = 7 (5–14); tod-
dler = 2 (1–9); and school-age = 2 (1–10).
We also extracted the 1,823 gene probes
differentially regulated between the
neonate group and controls (Table 4) and
compared these genes between the
neonate survivors and nonsurvivors.
None of the 1,823 gene probes were dif-
ferentially regulated between the sur-
vivors and nonsurvivors (ANOVA with a
false discovery rate of 1%).
Third, neonates had a significantly
higher proportion of infections due to
gram-positive bacteria compared with
the school-age group. This observation
raises the possibility that the differences
in gene expression described above re-
flect a pathogen class effect rather than
an effect of developmental age. An anal-
ysis (same sequential expression and
statistical filters as described for the pre-
vious analyses) was performed to com-
pare expression data from all patients
with gram-negative infection (n = 46) to
all patients with gram-positive infec-
tions (n = 47) and revealed only 11 dif-
ferentially regulated probes (Supple-
mentary Table 2).
Fourth, there were a variety of signifi-
cant differences between the four
developmental-age groups with respect
to peripheral differential white blood
cell counts. Because we used whole
blood–derived RNA, it is possible that
the differential gene expression patterns
described above reflect differences in pe-
ripheral white blood cell counts rather
than an effect of developmental age. To
address this, we analyzed our data for
the presence of previously published
“signature probe sets” for neutrophils
(38 probes), lymphocytes (50 probes)
and monocytes (28 probes), respectively
(35,36). We used the following criteria to
assess the presence of the signature
probe sets: ≥300 raw expression values
in a least one-half of the subjects in each
developmental age category. Table 5
demonstrates that the signature probe
sets were present to a similar degree
across the four developmental-age
groups. These data indicate that the rela-
tive contributions of the three major
leukocyte subsets to the whole blood
transcriptome expression patterns were
not substantially different across the
four developmental-age groups.
Although we cannot fully correct for
all potential confounders, the above
analyses indicate that the differences in
gene expression reported in this study
reflect, at least in part, a direct influence
of developmental age. We recognize that
whole-blood transcriptome alterations
corresponding to specific immune path-
Table 5. Expression of leukocyte subset signature probes across the four developmental-
age groups of patients with septic shock.
Neutrophil probes Lymphocyte probes Monocyte probes
n385028
Neonate 28 24 14
Infant 30 21 15
Toddler 30 21 15
School-age 32 17 16
Data are number of signature probes present (see text for presence criteria).
Page 10
1156 | WYNN ET AL. | MOL MED 17(11-12)1146-1156, NOVEMBER-DECEMBER 2011
DEVELOPMENTAL DIFFERENCES IN THE RESPONSE TO SEPTIC SHOCK
response syndrome, sepsis, and septic shock
spectrum. Crit. Care Med. 37:1558–66.
9. Wong HR, et al. (2009) Identification of pediatric
septic shock subclasses based on genome-wide
expression profiling. BMC Med. 7:34.
10. Wong HR, et al. (2007) Genome-level expression
profiles in pediatric septic shock indicate a role
for altered zinc homeostasis in poor outcome.
Physiol. Genomics 30:146–55.
11. Goldstein B, Giroir B, Randolph A. (2005) Inter-
national pediatric sepsis consensus conference:
definitions for sepsis and organ dysfunction in
pediatrics. Pediatr. Crit. Care Med. 6:2–8.
12. Shanley TP, et al. (2007) Genome-level longitudi-
nal expression of signaling pathways and gene
networks in pediatric septic shock. Mol. Med.
13:495–508.
13. Irizarry RA, et al. (2003) Exploration, normaliza-
tion, and summaries of high density oligonu-
cleotide array probe level data. Biostatistics.
4:249–64.
14. Calvano SE, et al. (2005) A network-based analy-
sis of systemic inflammation in humans. Nature.
437:1032–7.
15. Eichler GS, Huang S, Ingber DE. (2003) Gene Ex-
pression Dynamics Inspector (GEDI): for integra-
tive analysis of expression profiles. Bioinformatics.
19:2321–2.
16. Guo Y, Eichler GS, Feng Y, Ingber DE, Huang S.
(2006) Towards a holistic, yet gene-centered anal-
ysis of gene expression profiles: a case study of
human lung cancers. J. Biomed. Biotechnol.
2006:69141.
17. Wong HR, et al. (2010) Toward a clinically feasi-
ble gene expression-based subclassification strat-
egy for septic shock: proof of concept. Crit. Care
Med. 38:1955–61.
18. Wynn JL, Wong HR. (2010) Pathophysiology and
treatment of septic shock in neonates. Clin. Peri-
natol. 37:439–79.
19. Belderbos ME, et al. (2009) Skewed pattern of
Toll-like receptor 4-mediated cytokine produc-
tion in human neonatal blood: low LPS-induced
IL-12p70 and high IL-10 persist throughout the
first month of life. Clin. Immunol. 133:228–37.
20. Kollmann TR, et al. (2009) Neonatal innate TLR-
mediated responses are distinct from those of
adults. J. Immunol. 183:7150–60.
21. Levy O. (2007) Innate immunity of the newborn:
basic mechanisms and clinical correlates. Nat.
Rev. Immunol. 7:379–90.
22. Strunk T, Currie A, Richmond P, Simmer K,
Burgner D. (2011) Innate immunity in human
newborn infants: prematurity means more than
immaturity. J. Matern. Fetal. Neonatal. Med.
24:25–31.
23. Wynn JL, Levy O. (2010) Role of innate host de-
fenses in susceptibility to early-onset neonatal
sepsis. Clin. Perinatol. 37:307–37.
24. Bouchon A, Facchetti F, Weigand MA, Colonna M.
(2001) TREM-1 amplifies inflammation and is a
crucial mediator of septic shock. Nature. 410:1103–7.
25. Wynn JL, et al. (2007) Increased mortality and al-
tered immunity in neonatal sepsis produced by
generalized peritonitis. Shock. 28:675–83.
26. Adkins B, Leclerc C, Marshall-Clarke S. (2004)
Neonatal adaptive immunity comes of age. Nat.
Rev. Immunol. 4:553–64.
27. Wynn JL, et al. (2008) Defective innate immunity
predisposes murine neonates to poor sepsis out-
come but is reversed by TLR agonists. Blood.
112:1750–8.
28. Hotchkiss RS, et al. (2000) Caspase inhibitors im-
prove survival in sepsis: a critical role of the lym-
phocyte. Nat. Immunol. 1:496–501.
29. Hotchkiss RS, Coopersmith CM, Karl IE. (2005)
Prevention of lymphocyte apoptosis: a potential
treatment of sepsis? Clin. Infect. Dis. 41 Suppl
7:S465–9.
30. Kalman L, et al. (2004) Mutations in genes re-
quired for T-cell development: IL7R, CD45,
IL2RG, JAK3, RAG1, RAG2, ARTEMIS, and ADA
and severe combined immunodeficiency: HuGE
review. Genet. Med. 6:16–26.
31. Wynn JL, Neu J, Moldawer LL, Levy O. (2009)
Potential of immunomodulatory agents for pre-
vention and treatment of neonatal sepsis. J. Peri-
natol. 29:79–88.
32. Wynn JL, Seed PC, Cotten CM. (2010) Does IVIg
administration yield improved immune function
in very premature neonates? J. Perinatol. 30:635–42.
33. Ito S, et al. (2005) CpG oligodeoxynucleotides
enhance neonatal resistance to Listeria infection.
J. Immunol. 174:777–82.
34. Pedras-Vasconcelos JA, et al. (2006) CpG
oligodeoxynucleotides protect newborn mice
from a lethal challenge with the neurotropic
Tacaribe arenavirus. J. Immunol. 176:4940–9.
35. Sun L, Gorospe JR, Hoffman EP, Rao AK. (2007)
Decreased platelet expression of myosin regula-
tory light chain polypeptide (MYL9) and other
genes with platelet dysfunction and CBFA2/
RUNX1 mutation: insights from platelet expres-
sion profiling. J. Thromb. Haemost. 5:146–54.
36. Wong HR, Freishtat RJ, Monaco M, Odoms K,
Shanley TP. (2010) Leukocyte subset-derived
genomewide expression profiles in pediatric sep-
tic shock. Pediatr. Crit. Care Med. 11:349–55.
ways do not yield specific pathophysiol-
ogy. However, these data do offer insight
into the complex, multifactorial heteroge-
neous host response to sepsis and allow
identification of critical differences be-
tween age groups.
Children are not small adults, and, in
the host response to sepsis, we show that
neonates are not small children. Age-spe-
cific neonatal and pediatric studies of the
host response to sepsis are critically nec-
essary to permit identification of novel,
developmentally appropriate transla-
tional opportunities that might lead to
improved sepsis outcomes.
ACKNOWLEDGMENTS
This study was supported by grants
from the National Institutes of Health
(R01GM064619 and RC1HL100474).
The authors wish to thank Dr. Philip
O. Scumpia for his careful review of the
manuscript and comments.
DISCLOSURE
The authors declare that they have no
competing interests as defined by Molecu-
lar Medicine, or other interests that might
be perceived to influence the results and
discussion reported in this paper.
REFERENCES
1. Watson RS, Carcillo JA. (2005) Scope and epide-
miology of pediatric sepsis. Pediatr. Crit. Care
Med. 6:S3–5.
2. Mathews TJ, MacDorman MF. (2008) Infant mor-
tality statistics from the 2005 period linked
birth/infant death data set. Natl. Vital Stat. Rep.
57:1–32.
3. Lawn JE, Cousens S, Zupan J. (2005) 4 million
neonatal deaths: when? where? why? Lancet
365:891–900.
4. Checchia PA, et al. (2008) Myocardial transcrip-
tional profiles in a murine model of sepsis: evi-
dence for the importance of age. Pediatr Crit. Care
Med. 9:530–5.
5. Wynn J, Cornell TT, Wong HR, Shanley TP,
Wheeler DS. (2010) The host response to sepsis
and developmental impact. Pediatrics. 125:1031–41.
6. Cvijanovich N, et al. (2008) Validating the ge-
nomic signature of pediatric septic shock. Physiol.
Genomics. 34:127–34.
7. Ramilo O, et al. (2007) Gene expression patterns
in blood leukocytes discriminate patients with
acute infections. Blood. 109:2066–77.
8. Wong HR, et al. (2009) Genomic expression pro-
filing across the pediatric systemic inflammatory
Page 11
  • Source
    • "Genome-level expression patterns in children with septic shock strongly support this concept of immune suppression (2, 11–17). Specifically, pediatric septic shock is characterized by wide spread repression of gene programs corresponding to various major components of the adaptive immune system, including the T cell receptor signaling pathway, T cell function, B cell function, and the MHC antigen presentation pathway. "
    [Show abstract] [Hide abstract] ABSTRACT: For nearly a decade, our research group has had the privilege of developing and mining a multi-center, microarray-based, genome-wide expression database of critically ill children (≤ 10 years of age) with septic shock. Using bioinformatic and systems biology approaches, the expression data generated through this discovery-oriented, exploratory approach have been leveraged for a variety of objectives, which will be reviewed. Fundamental observations include wide spread repression of gene programs corresponding to the adaptive immune system, and biologically significant differential patterns of gene expression across developmental age groups. The data have also identified gene expression-based subclasses of pediatric septic shock having clinically relevant phenotypic differences. The data have also been leveraged for the discovery of novel therapeutic targets, and for the discovery and development of novel stratification and diagnostic biomarkers. Almost a decade of genome-wide expression profiling in pediatric septic shock is now demonstrating tangible results. The studies have progressed from an initial discovery-oriented and exploratory phase, to a new phase where the data are being translated and applied to address several areas of clinical need.Pediatric Research (2013); doi:10.1038/pr.2013.11.
    Preview · Article · Jan 2013 · Pediatric Research
  • Source
    • "Reuse of samples in several experiments: this is for instance the case in the five experiments, used as the core data of five papers, GSE9692 (9), GSE26378 (10), GSE8121 (11), GSE13904 (12), and GSE26440 (13). It appears that over the 101 samples that we have annotated from these experiments, 72 were reused several times: 15 were duplicated in four experiments (60 annotated samples, GSE9692, GSE8121, GSE13904, GSE26440); three were duplicated in two experiments (six annotated samples, GSE13904 and GSE26440), leading these experiments to have a total of 18 samples in common; yet, three others in two experiments (six annotated samples, GSE26378 and GSE26440). "
    [Show abstract] [Hide abstract] ABSTRACT: As part of the development of the database Bgee (a dataBase for Gene Expression Evolution), we annotate and analyse expression data from different types and different sources, notably Affymetrix data from GEO and ArrayExpress, and RNA-Seq data from SRA. During our quality control procedure, we have identified duplicated content in GEO and ArrayExpress, affecting ∼14% of our data: fully or partially duplicated experiments from independent data submissions, Affymetrix chips reused in several experiments, or reused within an experiment. We present here the procedure that we have established to filter such duplicates from Affymetrix data, and our procedure to identify future potential duplicates in RNA-Seq data.Database URL: http://bgee.unil.ch/
    Full-text · Article · Jan 2013 · Database The Journal of Biological Databases and Curation
  • Source
    • "All patients with microarray data in the current study were previously reported in studies addressing hypotheses entirely different from that of the current report [7,9-16,18,20]. For the current study, all patients in the sepsis and septic-shock cohorts had clinical microbiology laboratory confirmation of a bacterial pathogen from blood cultures or other normally sterile body fluids, whereas all patients in the SIRS cohort had negative bacterial cultures. "
    [Show abstract] [Hide abstract] ABSTRACT: Introduction: Differentiating between sterile inflammation and bacterial infection in critically ill patients with fever and other signs of the systemic inflammatory response syndrome (SIRS) remains a clinical challenge. The objective of our study was to mine an existing genome-wide expression database for the discovery of candidate diagnostic biomarkers to predict the presence of bacterial infection in critically ill children. Methods: Genome-wide expression data were compared between patients with SIRS having negative bacterial cultures (n = 21) and patients with sepsis having positive bacterial cultures (n = 60). Differentially expressed genes were subjected to a leave-one-out cross-validation (LOOCV) procedure to predict SIRS or sepsis classes. Serum concentrations of interleukin-27 (IL-27) and procalcitonin (PCT) were compared between 101 patients with SIRS and 130 patients with sepsis. All data represent the first 24 hours of meeting criteria for either SIRS or sepsis. Results: Two hundred twenty one gene probes were differentially regulated between patients with SIRS and patients with sepsis. The LOOCV procedure correctly predicted 86% of the SIRS and sepsis classes, and Epstein-Barr virus-induced gene 3 (EBI3) had the highest predictive strength. Computer-assisted image analyses of gene-expression mosaics were able to predict infection with a specificity of 90% and a positive predictive value of 94%. Because EBI3 is a subunit of the heterodimeric cytokine, IL-27, we tested the ability of serum IL-27 protein concentrations to predict infection. At a cut-point value of ≥5 ng/ml, serum IL-27 protein concentrations predicted infection with a specificity and a positive predictive value of >90%, and the overall performance of IL-27 was generally better than that of PCT. A decision tree combining IL-27 and PCT improved overall predictive capacity compared with that of either biomarker alone. Conclusions: Genome-wide expression analysis has provided the foundation for the identification of IL-27 as a novel candidate diagnostic biomarker for predicting bacterial infection in critically ill children. Additional studies will be required to test further the diagnostic performance of IL-27. The microarray data reported in this article have been deposited in the Gene Expression Omnibus under accession number GSE4607.
    Full-text · Article · Oct 2012 · Critical Care
Show more