Early life infection alters adult BALB/c hippocampal gene expression in a sex specific manner

Abstract

During the perinatal period, the developing brain is sensitive to environmental events. Deleterious programing resulting from infection, dietary restriction, or psychological stress has been observed and affects adult immune and endocrine systems as well as behavior. In this study, we determined whether neonatal infection permanently alters immune and glucocorticoid receptor signaling pathways in the adult hippocampus. A Chlamydia muridarum respiratory infection was induced in male and female mice at birth. Mice were allowed to recover and microarray analysis was conducted on RNA from adult hippocampal tissue. In males, neonatal infection induced an up-regulation of genes associated with cellular development, nervous system development and function, such as cyclin-dependent kinase inhibitor 1A. After neonatal infection, adult females exhibited a T-helper 2 immune bias with genes such as major histocompatibility complex, class II, DQ beta 1 up-regulated. Expression of prolactin, vasopressin, hypocretin, corticotrophin-releasing hormone-binding protein, and oxytocin were confirmed by quantitative real-time polymerase chain reaction. This study shows that neonatal infection differentially alters the gene expression profiles of both female and male mice along immune and neuroendocrine pathways.

Keywords::

• Microarray
• hippocampus
• neonatal stress
• infection
• glucocorticoid receptor
• HPA

Introduction

The impact of early life insults on adult physiology is well established (Strachan 1989; McMillen and Robinson 2005). Events such as infection, dietary restriction, or psychological stress have the capacity to detrimentally affect adult immune and endocrine systems as well as behavior (Strachan 2000; Buitelaar et al. 2003; McGowan et al. 2008). The mechanisms underlying these long-term effects of early life programing have been investigated using a variety of paradigms including neonatal exposure to pathogens. Up to one-third of pregnancies suffer complications involving infections or trauma of the fetus or newborn (Garnier et al. 2003) and an increasing body of evidence indicates that perinatal events involving the immune system may contribute to the development of adult disease (Shanks et al. 1995, 2000; Hall and Peckham 1997). Furthermore, the immune system in the neonatal period is not fully developed and consequently the neonate is more at risk of infection than an adult, with newborns particularly susceptible to bacterial and fungal infections (Garvy 2003). Nevertheless, the long-term impact of early life infection and its effect on adult health remains incompletely understood. Therefore, it is important to examine the impact of pre- and post-natal infection on adult physiology and programing to determine the mechanisms leading to disease and to explore potential interventions.

Chlamydiae are common human pathogens in all age groups (Hammerschlag 2000) and frequently cause trachoma, sexually transmitted diseases, and pneumonia (Stephens 2003). We have established a mouse model of subacute respiratory infection using Chlamydia muridarum, which naturally infects mice and is an appropriate choice for investigating natural host–pathogen interactions. The temporal, histological, and immunological progression of disease is typical of the immunopathological reactions that are representative of human infection and disease caused by chlamydiae (Adkins et al. 1993). We have used this model to show that C. muridarum infection during the neonatal period has long-term effects on the immune and respiratory systems. Infection affects the development of allergic airways disease and is associated with altered T helper 1 (Th1) and T helper 2 (Th2) responses in the adult (Horvat et al. 2007, 2010). C. muridarum infection activates the stress response as seen by excess glucocorticoids (GCs) circulating in the developing neonate (Ruzek et al. 1999). However, whether infection in early life can induce alterations in the stress response and immunological programing in the brain later in life remains unknown. In this study, we have examined the effects of neonatal respiratory C. muridarum infection on hippocampal gene expression. This was achieved by microarray profiling of adult hippocampal tissue and the assessment of infection-induced changes to GC receptor (GR) signaling pathways and immune response indicators.

Methods

Animals

All animals were housed in an accredited animal care facility under specific pathogen-free conditions. Pregnant female BALB/c mice were obtained from the central animal house and used with approval from the animal ethics committee, University of Newcastle (NSW, Australia). Dams were left undisturbed until birth of the young. Offspring were weaned and sexed at 21 days of age. Animals were given access to food and water ad libitum, and housed either per litter (mothers with young) or in groups not larger than six (adult and weaned young) under controlled environmental conditions with a 12:12 h light:dark cycle. Samples were taken from an experimental pool of 64 mice born to 11 mothers. Litter sizes were between 2 and 10 mice; resulting in 31 male and 33 female mice from five infected litters and six uninfected litters. Individual litters were not kept separate. The experimental pool consisted of a mix of mice from different litters of different sizes. Non-infected mice were not litter mates of infected mice. The microarray study was conducted using pools of six mice from each treatment condition and sex (i.e. male control, male infected, female control, and female infected). The follow-up reverse transcription (RT) of complementary DNA (cDNA) and quantitative real-time polymerase chain reaction (qRT PCR) was completed using RNA from 10 mice in each treatment condition and sex.

Infection protocol and tissue collection

The C. muridarum strain ATCC VR-123 (formerly known as the mouse pneumonitis biovar of C. trachomatis) was obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). Stocks were stored at − 80°C in sucrose–phosphate–glutamate buffer (SPG). Within 24 h of birth, neonatal mice were infected intranasally with C. muridarum without anesthesia. Mice were held in the upright position and the inoculum (400 inclusion-forming units; ifu) in 5 μl of SPG was gently pipetted onto the nares until the whole inoculum was inhaled (Horvat et al. 2007, 2010). Control groups were sham-inoculated with 5 μl of SPG. We have previously shown that infection peaks between 10 and 15 days after inoculation, bacterial clearance occurs by day 21, and pulmonary inflammation is largely resolved by day 45 (Horvat et al. 2007). Mice were allowed to recover from the infection and mature into adults. At 63 days of age, mice were euthanized by sodium pentobarbital overdose (Abbott Australasia, Kurnell, Australia) and the brain was removed and placed in RNAlater (Applied Biosystems/Ambion, Austin, TX, USA) solution and stored at − 80°C.

Parallel RNA/DNA extractions and purification

Parallel RNA/DNA extractions were carried out on the whole hippocampal tissue. The hippocampus was identified by referring to the Allen Mouse Brain Atlas (www.brain-map.org, Lein et al. 2007). For the microarray portion of the study, RNA from six mice was pooled per treatment group (infection and control) and sex (male and female). For the follow-up qRT PCR, 10 mice per treatment group and sex were used. RNA was extracted using RNA/DNA mini kit according to the manufacturer's instruction. Purification of RNA was carried out using an RNeasy mini kit (Qiagen, Hilden, Germany) with DNase treatment using an RNase-free DNase set (Qiagen). RNA concentration was determined by spectrophotometer (NanoDrop 3300, Thermo Scientific, Wilmington, DE, USA) and the quality was assessed by agarose gel electrophoresis.

Microarray profiling

For the microarray portion of the study, the RNA was sent for gene profiling at an external facility [Australian Genome Research Facility (AGRF), St Lucia, QLD, Australia]. Samples were pooled by sex and treatment with six mice in each treatment/sex condition. Microarray profiling was performed using the Illumina MouseWG-6 v1.1 Expression BeadChip, one chip per pooled sample, which then was analyzed using an Illumina BeadArray Reader. The multi-sample format of the chip allows profiling of more than 45,200 transcripts and six samples simultaneously on a single BeadChip. The resulting data were normalized by quantile normalization after variance-stabilizing transformation, using Lumi package (http://www.basic.northwestern.edu/projects/lumi/) in the R statistical software (Du et al. 2008) and were subsequently imported into GeneSpring GX 7 (Agilent Technologies, VIC, Australia) for further analysis. Comparisons were of infected vs. control groups for the same sex; male infected vs. male control and female infected vs. female control. Differentially expressed genes were identified as those showing a ≥ 50% difference in mean expression levels between groups. There were no replicates; therefore, no statistical test (such as t-test) was possible.

To identify gene networks and signaling pathways containing differentially expressed genes between conditions, we used ingenuity pathway analysis (IPA; Ingenuity^® Systems, Redwood City, CA, USA, www.ingenuity.com/). IPA is a web-based software which organizes differentially expressed genes into molecular pathways and functional groupings based on the published literature, and the software enables examination of affected canonical signaling pathways. Functional analysis in IPA identifies the published biological functions that were most significantly associated with the genes in the network. Genes are represented as nodes, and the biological relationship between two nodes is represented by as a line. All lines are supported by at least one reference in the literature, textbook, or from canonical information stored in the ingenuity pathway's knowledge database. IPA also allows the overlay of specific signaling pathways on identified networks. In this study, we overlaid the GR signaling pathway onto the top gene networks identified by IPA.

Rt Of Cdna And Qrt-pcr

RT was carried out following the manufacturer's instructions using Superscript First-Strand Synthesis System III for RT-PCR kit (Invitrogen, Carlsbad, CA, USA). qRT-PCR was used to determine mRNA abundance of several genes, which were compared to the levels of the constitutively expressed gene, β-actin. Primer sequences were designed with the Primer Express program (Applied Biosystems, Warrington, UK: Table I) and products were sequenced at an external facility (AGRF). qRT PCR was performed using the supplied program on a PCR 7700 sequence detector (Applied biosystems) in duplicate or triplicate under the following reaction conditions: 2 min at 50°C, 10 min at 95°C, and 40 cycles of 15 s at 95°C and 1 min at 60°C. A single or duplicate sample that had not been reverse transcribed was included as a control for each sample, along with single or duplicate controls without template for each primer. Data were analyzed using the ABI Prism Sequence Detection Systems software version 1.9 (Applied Biosystems). The comparative CT method (where CT is the threshold cycle) was used to derive a relative quantitative measure of the target gene expression compared with β-actin mRNA.

Table I.  Primer details for target and reference genes.

Target gene	Forward	Reverse
Oxytocin	TTGCTGCCTGCTTGGCTTA	GGAGACACTTGCGCATATCCA
Prolactin	TGGTTCTCTCAGGCCATCTTG	CAGATTGGCAGAGGCTGAACA
Vasopressin	TCGCCAGGATGCTCAACAC	CACTGTCTCAGCTCCATGTCAGA
CRH BP	AGCTGGAGCCATTCGAGCTA	TGCTGCCCATGAAGATCACA
Hypocretin	GCGCAGAGCTAGAGCCACAT	TGCTAAAGCGGTGGTAGTTACG
β-actin	AGAGGGAAATCGTGCGTGAC	CAATAGTGATGACCTGGCCGT

Data analysis

qRT PCR data were analyzed by between subjects two-factor full factorial ANOVA using the PASW software (SPSS Inc., Chicago, IL, USA), with the factors being treatment (control and post-natal infection) and sex (male and female), the significance level was set at p ≤ 0.05. Data were log transformed before the analysis. Post hoc analysis was performed within each sex if the two-way ANOVA showed a significant interaction between sex and treatment. Individual data were removed from the analysis if individual samples returned gene expression below detectable levels after qRT PCR.

Results

Overall gene changes

In male hippocampus, 392 (100%) genes were differentially expressed after neonatal infection with C. muridarum; of those genes, 134 (34%) were increased in expression by more than 1.5-fold and 258 (66%) decreased in expression by 0.5-fold. In female adult hippocampus, 541 genes were differentially expressed, of which 330 (61%) were increased in expression by more than 1.5-fold and 211 (39%) were decreased in expression by 0.5-fold. Of the genes differentially expressed, 90 were commonly expressed in males and females; 276 were specific to males and 409 were specific to females.

Genes differentially expressed in male adult mice after neonatal infection

The IPA analysis was conducted on all genes that were differentially expressed in the adult hippocampus after neonatal C. muridarum infection. The five most substantially altered gene networks in the adult male hippocampus are shown in Table II. The specific networks identified were associated with cellular development, nervous system development and function, neurological disease, cellular compromise, and tissue morphology. In network 1, 14 genes were differentially expressed and associated with tissue morphology, cellular development, and connective tissue development and function (Table II). For example, hippocampal 11 β-hydroxysteroid dehydrogenase 1 (11β HSD1) gene expression was increased by 1.63-fold following male neonatal infection and was identified as being indirectly regulated by Interleukin 4 (IL-4; Figure 1).

Table II.  Networks identified of genes differentially expressed in male adult mice after neonatal infection (bold indicates up-regulation).

ID	Molecules in network	Score	Focus molecules	Top functions
1	ACP5, AHR, ATF4, B2M, C5, C5AR1, CDKN1A, DCN, DUSP1, FAS, FCER1G, FCGR2A, FHL2, FOS, HSD11B1, IFITM2, IFNG, IL3, IL4, JUN, JUNB, KITLG, KLF2, LGALS1, MEF2C, MYH7, PLAU, PRDM1, PRKCB, RUNX2, SELP, SELPLG, STAT3, TNF, TNFSF11	21	14	Tissue morphology, cellular development, connective tissue development, and function
2	ARC, CREB1, CREM, DUSP14, FMR1, GPRIN1, SERTAD1, SIK1	6	4	Nervous system development and function, cellular compromise, and neurological disease
3	MITF, MYL4	2	1	Cellular development, embryonic development, and gene expression
4	KAP, PPIB	2	1	Drug metabolism, small molecule biochemistry, DNA replication, recombination, and repair
5	PLXNA1, TREM2	2	1	Cell-to-cell signaling and interaction, cellular function and maintenance, and genetic disorder

Figure 1.  Network 1 identified by IPA showing differentially expressed genes in male adult hippocampus after neonatal respiratory C. muridarum infection. Red shading indicates up-regulation of the gene. No color indicates genes that form the network and have no change in expression in the present study. Fourteen genes were differentially expressed and associated with tissue morphology, cellular development, and connective tissue development and function. For example, 11β-HSD1 gene expression was altered following male neonatal infection and was indirectly regulated by IL-4. For full names of genes involved in the network, see Table IV.

Genes differentially expressed in female adult mice after neonatal infection

The five most substantially altered gene networks in females are shown in Table III. The female-specific networks identified were associated with cell-to-cell signaling, immune cell trafficking, cell death, cellular movement, cancer, as well as cellular growth and proliferation. In network 1, 13 genes were differentially expressed after infection and were identified as indirectly involved with cytokine regulation as shown in Figure 2. For example, major histocompatibility complex, class II, DQ beta 2 (HLA-DQβ2) was increased by 1.68-fold following infection and identified to indirectly regulate IL-6 and IL-10 (Figure 2). In network 4 (Table III), 10 genes were differentially expressed including tyrosine hydroxylase (1.60-fold increase), a rate-limiting enzyme in catecholamine synthesis.

Table III.  Networks identified of genes differentially expressed in female adult mice after neonatal infection (bold indicates up-regulation; italic indicates down-regulation).

Network ID	Molecules in network	Score	Focus molecules	Top functions
1	AHCYL1, ATP6V0E2, B2M, C5, C5AR1, CD74, CSF2, DCTN1, FCER1G, FCGR2A, FMR1, FOXP3, HCN2, HLA-C, HLA-DQB2, HSPD1, HTT, IFNG, IL4, IL6, IL10, IL13, ITPR1, MBP, MYC, MYT1L, POU3F1, RGS14, RGS16, RHOA, STAT3, TGFB1, TLR4, TNF, TP53	13	13	Cell-to-cell signaling and interaction, hematological system development and function, and immune cell trafficking
2	AKAP9, BCAR1, CIT, CTNNB1, EGFR, FGF8, FNTA, FYN, GBX2, GRIK2, GRIN1, GRM1, MMP2, MMP14, MSX1, MTCH2, NR3C1, OTX2, PDIA3, PTP4A2, SEP15, SLC6A3, SRC, SUMOl, TGFB1I1, TRH, UGCG, WNT5A	11	11	Cell death, cellular movement, and cancer
3	AGRP, APBB1, ASS1, CEBPB, CYTB, EGFR, GPX1, GSK3B, HNRNPK, LEP, LEPR, MYH7, ND4, NF2, PPARGC1A, PRKCB, PRL, PSMB8, RORA, SOCS3, SRC, STAT1, STAT3, TNFSF11	9	9	Cellular growth and proliferation, cellular development, and genetic disorder
4	ADIPOQ, CEBPB, CIITA, CSF2, DDC, DLK1, FOS, FOXA1, FOXA2, GH1, GPC3, GRB10, IFNG, IL6, IL10, IL13, INSR, MMP2, MMP12, NNAT, PDPK1, PRKCB, PRKCD, PRKCE, SCNN1B, SELP, SELPLG, SERPINE1, STAT3, TGFB1, TH, THRB (includes EG:7068), TNF, TOP2A, YY1	9	10	Cellular movement, lipid metabolism, and molecular transport
5	SNRPN, UBE3A	1	1	Cellular function and maintenance, reproductive system development and function, and genetic disorder

Table IV.  Definitions of abbreviations used in figures and tables.

Abbreviation	Gene name
ACP5	Acid phosphatase 5, tartrate resistant
ADIPOQ	Adenylate cyclase 9
AGRP	Agouti-related protein homolog (mouse)
AHCYL1	Adenosylhomocysteinase-like 1
AHR	Aryl hydrocarbon receptor
AKAP9	A kinase (PRKA) anchor protein (yotiao) 9
APBB1	Amyloid beta precursor protein (cytoplasmic tail) binding protein 1
ASS1	Argininosuccinate synthetase 1
ATF4	Activating transcription factor 4 (tax-responsive enhancer element B67)
ATP6V0E2	ATPase, H^+ transporting V0 subunit e2
ARC	Activity-regulated cytoskeleton-associated protein
B2M	Beta-2-microglobulin
BCAR1	Breast cancer anti-estrogen resistance protein 1
C5	Complement component 5
C5AR1	Complement component 5a receptor 1
CD74	CD74 molecule, major histocompatibility complex, class II invariant chain
CDKN1A	Cyclin-dependent kinase inhibitor 1A (p21, Cip1)
CIT	Citron (rho-interacting serine/threonine kinase 21)
CIITA	Class II, major histocompatibility complex, transactivator
CREB1	cAMP responsive element binding protein 1
CREM	cAMP responsive element modulator
CSF2	Colony stimulating factor 2 (granulocyte-macrophage)
CTNNB1	catenin (catherin-associated protein), beta 1, 88 kDa
CYTB	Mitochondria encoded cytochrome B
DCN	Decorin
DCTN1	Dynactin 1 (p150, glued homolog, drosophila)
DDC	Dopa decarboxylase (aromatic l-amino acid decarboxylase)
DLK1	Delta-like 1 homolog (drosophila)
DUSP1	Dual specificity phosphatase 1
EGFR	Epidermal growth factor receptor
FAS	FAS (TNF receptor superfamily, member 6)
FCER1G	Fc fragment of lgE, high affinity I, receptor for, gamma polypeptide
FCGR2A	Fc fragment of IgG, low affinity IIa, receptor (CD32)
FGF8	Fibroblast growth factor 8 (acidic)
FHL2	Four-and-a-half LIM domains 2
FOS	v-fos FBJ murine osteosarcoma viral oncogene homolog
FOXA1	Forkhead box Al
FOXA2	Forkhead box A2
FOXP3	Forkhead box P3
FMR1	Fragile × mental retardation 1
FNTA	Famesyltransferase, CAAX box, alpha
FYN	Proto-oncogene tyrosine-protein kinase
GBX2	Gastrulation brain homeobox 2
GH1	Growth hormone 1
GPC3	Glypican 3
GPRIN1	G protein regulated inducer of neurite outgrowth 1
GPX1	Glutathione peroxidase 1
GRB10	Growth factor receptor-bound protein 10
GRIK2	Glutamate receptor, ionotropic, kainate 2
GRIN1	Glutamate [NMDA] receptor subunit zeta-1
GRM1	Glutamate receptor, metabotropic 1
GSK3B	Glycogen, synthase kinase 3 beta
HCN2	Hyperpolarization-activated cyclic nucleotide-gated potassium channel 2
HLA-C	Major histocompatibility complex, class I, C
HLA-DBQ2	Major histocompatability complex, class II, DQ beta 1
HSD11B1	Hydroxysteroid (11-beta) dehydrogenase 1
HSPD1	Heat shock 60 kDa protein 1 (chaperonin)
HTT	Huntingtin gene
IFITM2	Interferon (alpha, beta, and omega) receptor 1
IFNG	Interferon, gamma
IL	Interleukin
INSR	Insulin receptor
ITPR1	Inositol 1,4,5-trisphosphate receptor type 1
JUN	Jun oncogene
JUNB	JunB proto-oncogene
KITLG	KIT ligand
KLF2	Kruppel-like factor 2 (lung)
LEP	Leptin
LEPR	Leptin receptor
LGALS1	Lectin, galactoside-binding, soluble, 1
MBP	Myelin basic protein
MEF2C	Myocyte enhancer factor 2C
MITF	Microphthalmia-associated transcription factor
MMP2	Matrix metalloproteinase 2
MMP12	Matrix metallopeptidase 12
MMP14	Matrix metallopeptidase 14
MSX1	MSH homeobox 1
MTCH2	Mitochondrial carrier homolog 2 (C. elegans)
MYC	Myc gene
MYL4	Myosin, light chain 4, alkali, atrial, embryonic
MYH7	Myosin, heavy chain 7, cardiac muscle, beta
MYT1L	Myelin transcription factor 1-like
NR3C1	Glucocorticoid receptor
ND4	NDH dehydrogenase subunit 4
NF2	Neurofibromin 2
NNAT	Neuronatin
OTX2	Orthodenticle homeobox 2
PDIA3	Protein disulfide isomerase associated 3
PDPK1	3-Phosphoinositide dependant protein kinase-1
PLAU	Plasminogen activator, urokinase
PLXNA1	Plexin A1
POU3F1	POU class 3 homeobox l
PPARGC1A	Peroxisome proliferative activated receptor, gamma, coactivator 1 alpha
PPIB	Peptidylprolyl isomerase B (cyclophilin B)
PRDM1	PR domain containing 1, with ZNF domain
PRKCB	Protein kinase C, beta
PRKCD	Protein kinase C, delta
PRKCE	Protein kinase C, epsilon
PRL	Prolactin
PSMB8	Proteasome (prosome, macropain) subunit, beta type, 8
PTP4A2	Protein tyrosine phosphatase type IV A, member 2
RGS14	Regulator of G-protein signaling 14
RGS16	Regulator of G-protein signaling 16
RHOA	RAS homolog gene family, member A
RORA	RAR-related orphan receptor A
RUNX2	Runt-related transcription factor 2
SCNN1B	Sodium channel, non-voltage-gated 1, beta
SELP	Selectin P (granule membrane protein 140 kDa, antigen CD62)
SELPLG	Selectin P ligand
SEP_15	15 kDa selenoprotein
SERPINE1	Serine (or cysteine) peptidase inhibitor, clade E, member 1
SERTAD1	SERTA domain containing 1
SIK1	Salt-inducible kinase 1
SLC6A3	Solute carrier family 6 (neurotransmitter transporter, dopamine), member 3
SOCS3	Suppressor of cytokine signaling 3
SNRPN	Small nuclear ribonucleoprotein N
SRC	Rous sarcoma oncogene
STAT	Signal transducer and activator of transcription
SUMOl	SMT3 suppressor of mif two 3 homolog 1 (S. cerevisiae)
TGFB1	Transforming growth factor beta 1
TGFB1I1	Transforming growth factor beta 1 induced transcript 1
TH	Tyrosine hydroxylase
THRB	Thyroid hormon receptor beta
TLR4	Toll-like receptor 4
TNF	Tumor necrosis factor (TNF superfamily, member 2)
TNFSF11	Tumor necrosis factor (ligand) superfamily, member 11
TOP2A	Topoisomerase (DNA) II alpha
TP53	Tumor protein 53
TREM2	Triggering receptor expressed on myeloid cells 2
TRH	Thyrotropin-releasing hormone
UBE3A	Ubiquitin protein ligase E3A
UGCG	UDP-glucose ceramide glicosyltransferase
WNT5A	Wingless-related MMTV integration site 5A.
YY1	YY1 transcription factor

Figure 2.  Network 1 identified by IPA showing differentially expressed genes in female adult hippocampus after neonatal respiratory C. muridarum infection. Red shading indicates up-regulation of the gene. Gray shading indicates down-regulation. No color indicates genes that form the network and have no change in expression in the present study. Thirteen genes were differentially expressed after infection and were indirectly involved with cytokine regulation. For example, HLA DQB2 was increased following infection and identified to indirectly regulate IL-6 and IL-10. For full names of genes involved in the network, see Table IV.

Differential expression of genes related to GR signaling pathway

Using IPA, the relationship between the GR signaling pathway and the networks listed for male adult hippocampus after neonatal C. muridarum infection was examined and two networks were related to the GR pathway. Network 1 (Table II and Figure 1) showed eight genes regulated by the GR pathway. The genes were related to inflammatory response and cell regulation. Two implicated genes were up-regulated after infection, FOS (v-fos FBJ murine osteosarcoma viral oncogene homolog; 1.60-fold) and cyclin-dependent kinase inhibitor 1A (CDKN1A; 1.605-fold).

The relationship between the GR signaling pathway and the networks listed for female adult hippocampus after neonatal C. muridarum infection was examined. In female adult hippocampus after infection, within network 1 (Table III and Figure 2), the GR signaling pathway interacted with nine genes including IL-4, IL-6, IL-10, tumor necrosis factor α (TNFα), interferon (IFN)-γ, and signal transducer and activator of transcription (STAT)-3 which are immune-related genes. In network 2 (Table III), small ubiquitin-like modifier-1 was up-regulated (1.61-fold) and this gene interacts directly with the GR signaling pathway as defined by the IPA. In network 3, prolactin was increased (7.51-fold) and identified as directly regulated by GR signaling pathway (Table III). Other components of this network; STAT-1, STAT-3, and CCAAT/enhancer-binding protein (C/EBP) were also identified as part of the GR signaling pathway.

Confirmation of gene changes in individual hippocampal mRNA

Confirmation of the microarray results by qRT PCR was performed on a number of genes. Candidate genes were chosen a priori according to the following specifications, if the microarray data analysis showed high fold changes; if the genes were altered in each treatment/sex condition; and if the genes were linked to GR/HPA functioning. The genes selected were prolactin, vasopressin, hypocretin, corticotrophin-releasing hormone-binding protein (CRH BP), and oxytocin. After neonatal infection, males had an increase in CRH BP (1.78-fold); and decreases in vasopressin (0.44-fold), hypocretin (0.30-fold), and oxytocin (0.24-fold). Females had increases in prolactin (7.51-fold), oxytocin (4.92-fold), hypocretin (9.51-fold), and vasopressin (13.07-fold) after neonatal infection. Table V compares gene expression changes found by microarray to gene expression changes found by qRT PCR and shows the results of the pairwise comparison by treatment conducted after ANOVA.

Table V.  Direction of change in mRNA in microarray (fold change) and qRT PCR (relative abundance) with pairwise comparison across treatment for qRT PCR.

		Infection vs. control		Pairwise comparison
Gene	Sex	Fold change	qRT PCR	F	P
Prolactin	Male	–	–	5.29	0.03
	Female	↑	–	0.03	0.87
Vasopressin	Male	↓	↓	136.61	< 0.01
	Female	↑	↑	4.11	0.05
Hypocretin	Male	↓	↓	14.61	< 0.01
	Female	↑	↑	19.08	< 0.01
CRHBP	Male	↑	↓	1227.25	< 0.01
	Female	–	–	0.10	0.75
Oxytocin	Male	↓	↓	197.51	< 0.01
	Female	↑	–	3.1	0.87

The two-way ANOVA showed that prolactin mRNA abundance was unchanged by neonatal infection for both sexes (Figure 3a). For vasopressin mRNA, the ANOVA revealed main effects of sex, F(1,35) = 90.1, p < 0.01, and treatment, F(1,36) = 48.89, p < 0.01, with a significant interaction, F(1,35) = 96.24, p < 0.01. Males had significantly less vasopressin mRNA after infection when compared to same sex controls, p < 0.01 (Figure 3b). Females had significantly more vasopressin mRNA after infection when compared to same sex controls, p = 0.05.

Figure 3.  Expression of target mRNA relative to β-actin (log) in adult male and female hippocampus after neonatal respiratory infection with C. muridarum compared to sham-infected controls. The expression of (a) prolactin, (b) vasopressin, (c) hypocretin, (d) CRH-BP, and (e) oxytocin.

A two-way ANOVA revealed a significant interaction of sex and treatment for hypocretin mRNA expression after neonatal infection, F(1,32) = 33.53, p < 0.01. Infected males had a significant reduction in hypocretin, p < 0.01, and females had significant increase compared with uninfected controls, p < 0.01 (Figure 3c). There was significant main effects of sex, F(1,36) = 562.14, p < 0.01 and treatment, F(1,36) = 602.45, p < 0.01 and the interaction was significant, F(1,36) = 624.91, p < 0.01 for CRH BP. In males, infection resulted in significantly less CRH BP than controls, p < 0.01. However, there was no significant difference in CRH BP abundance due to treatment in females, p = 0.75 (Figure 3d). The between-subjects ANOVA for oxytocin mRNA revealed a main effect of sex, F(1,35) = 72.74, p < 0.01, and treatment, F(1,35) = 73.44, p < 0.01 with the interaction also significant, F(1,35) = 122.88, p < 0.01. In males, oxytocin mRNA was significantly reduced following neonatal infection relative to the controls, p < 0.01, whereas expression was increased in females, but not significantly, p = 0.09 (Figure 3e).

Discussion

The rate of infection is high in newborns due to their relatively immature, naive immune system (Garvy 2003) and our results demonstrate that early life infection significantly alters adult hippocampal gene expression in a sex-specific manner. The data indicate that adult gene expression profiles were altered with males that were exposed to infection early in life, exhibiting gene changes associated with tissue morphology or cellular development such as CDKN1A, a regulator of cell cycle progression, when compared to male controls. In contrast, females exposed to infection in early life exhibited a Th2 cytokine bias with an up-regulation of genes related to humoral immunity such as IL-6 and IL-10, when compared to females in the control condition.

Perinatal stress can modify the hippocampal gene expression. Previous studies have examined the effects of pre- and post-natal stress, including psychological (Vallee et al. 1999; Szuran et al. 2000; Tuchscherer et al. 2004; Kapoor et al. 2008) and physiological stress (Sutanto et al. 1996; Liu et al. 1997; Dodic et al. 2002; Walker et al. 2004b, 2009; Shoener et al. 2006); and have found that early life stressors permanently alter hippocampal corticosteroid receptor gene expression. These studies report that early life stress leads to a reduction in GR and mineralocorticoid receptor (MR) expression (Sutanto et al. 1996; Dodic et al. 2002; Kapoor et al. 2008), resulting in anxiety-like behavior (Walker et al. 2004b, 2009), learning deficits (Vallee et al. 1999; Szuran et al. 2000), and altered stress responsivity in adulthood (Liu et al. 2001; Walker et al. 2009). This suggests that the GR/MR signaling pathway may be permanently altered by early life stress. The present study aimed to expand on previous research by examining potential global gene changes and exploring the relationship of these gene alterations to the GR signaling pathway.

The IPA analysis of the adult male hippocampus data indicated that the genes differentially expressed compared to male controls had limited interaction with the GR signaling pathway with only 10 genes in all the networks identified as related to GR signaling, suggesting that in the males, GR-related genes were relatively unaffected by neonatal infection. One exception was 11β-HSD1 which was shown to be up-regulated after neonatal infection. In the hippocampus, 11β-HSD1 converts the inactive 11-dehydrocorticosterone to the active corticosterone (Rajan et al. 1996). 11β-HSD1 was up-regulated after infection which would increase the amount of corticosterone in the hippocampus, thereby decreasing the amount of GR and MR. This suggests that in the males, hippocampal function may be altered through genes not directly related to the GR signaling pathway.

Of the networks altered in female adult hippocampi after infection, the GR signaling pathway interacted with many genes including IL-6, IL-10, TNFα, IFN-γ, and STAT-3, with those Th2 factors associated with up-regulated genes. Prolactin was up-regulated and identified as being directly regulated by GRs. Other genes such as STAT-1, STAT-3, and C/EBP were also identified as part of the GR signaling pathway by regulating or being regulated by GRs (Zhang et al. 1997; Boruk et al. 1998). This suggests a pervasive up-regulation of genes linked to the GR signaling pathway after neonatal infection in the female rodent.

Generally, with age comes a shift from an IL-10 immune dominance to an IL-4 immune dominance (Pettiford et al. 2002), and in the present study, the males show evidence of this shift, whereas the females show evidence of IL-10 and IL-6 dominance. This suggests that for male mice, immune organization was not changed by neonatal infection, whereas females had their immune profile altered. Clinically, many diseases and disorders occur in sex-specific patterns, for example women are two to three times more likely than men to present with affective disorders, autism affects four times as many boys as girls, and multiple sclerosis and lupus erythematosus are more prevalent in women (Bale 2009). Although it is well established that early life events program the functioning of physiological systems (Hales et al. 1991; Meaney et al. 1996; Hodgson et al. 2001; Welberg et al. 2001; Walker et al. 2009; Horvat et al. 2010), there is a gap in the knowledge of the role sex plays in determining adult outcomes after perinatal immune activation.

The present study shows that it is necessary to examine both males and females, due to their physiological differences it is unsuitable to extrapolate the results of one to the other. Primary targets that may account for such differences are, most obviously, the sex hormones. Human studies have shown that in response to stress, (i.e. the trier social stress test), men have a higher adrenocorticotropic hormone (ACTH) response to the stressor compared with women, but the plasma cortisol response is the same in both men and women (Kirschbaum et al. 1999; Young and Korszun 2010). However, the female cortisol response is dependent on the menstrual cycle phase (Young et al. 1991; Young and Korszun 2010). This response appears to be due to more than just the influence of estrogens. The differences are related to levels of progesterone, with decreases in progesterone resulting in cortisol suppression (Young et al. 1991). Estradiol influences hypothalamic–pituitary–adrenal (HPA) axis functioning and has been shown to decrease stress responsiveness in rodents (Young et al. 2001), sheep (Komesaroff et al. 1998), and humans (Komesaroff et al. 1999). Human studies suggest that estradiol acts as a ‘brake’ on the HPA axis response and that progesterone impairs GC negative feedback (Young and Korszun 2010).

In the present study, confirmation by qRT-PCR of target genes found that for the males, the microarray data were supported for all genes tested except CRH-BP and prolactin. For the females, the findings of the microarray data were supported with a significant increases in vasopressin, hypocretin, and oxytocin mRNA. Overall, this study provides evidence of the effects of early life infection on adult gene expression and provides important information of potential gene and network targets, as seen in Figures 1 and 2, that may be investigated to further understand the impact of perinatal infection on adult immune functioning and stress responsivity.

The focus genes for follow-up qRT PCR were chosen due to the high fold changes seen in the microarray and their proposed role in the GR/HPA functioning. Hypocretin is involved in the biostability of arousal states and hippocampal plasticity (Selbach et al. 2010), and neurons expressing hypocretin have been shown to be directly and indirectly regulated by catecholamines (Yamanaka et al. 2006). Vasopressin, along with CRH is one of the principal effectors of the stress system (Charmandari et al. 2003) and has been extensively studied in the hypothalamus, but not in the hippocampus. Females in our study showed an increase in vasopressin mRNA after infection, whereas the males showed a significant decrease. CRH has been shown to be neuroprotective in the hippocampus (Pedersen et al. 2001; Elliott-Hunt et al. 2002) and we found that males had less CRH BP after infection, thereby increasing the availability of CRH in the hippocampus. This suggests that males may be more protected from stress-related excitotoxicity. Oxytocin has been shown to be regulated by GCs with decreases in GCs resulting in decreases of oxytocin in the hippocampus (Liberzon et al. 1994). Also, oxytocin may override the stress response by eliminating the hypothalamic response to stress (Lightman 1992). More recent data support this theory, as oxytocin has been shown to regulate GR and MR in the hippocampus with increases in oxytocin resulting in decreased GR and MR (Petersson and Uvnas-Moberg 2003). Our follow-up qRT PCR showed after infection that females had higher oxytocin mRNA, whereas males had significantly less oxytocin. Overall, the data are further evidence of the complex influence of sex on potential health outcomes. The gene expression profiles of both female and male mice were altered by neonatal infection, but along differing immune and neuroendocrine pathways.

There are a number of possible alternate explanations for our results; one factor not accounted for is the influence of maternal behavior on adult outcomes. Any alterations to endocrine activity may be in part mediated by the impact of infection on maternal behavior. In rodent studies, an attentive dam results in offspring with reduced plasma ACTH and corticosterone, increased hippocampal GR, and enhanced negative feedback sensitivity when compared to less attentive dams (Liu et al. 1997; Francis et al. 1999). Although there has been much research on the level of maternal care and the response to adult stressors, few studies have assessed maternal care after neonatal infection. Dams of pups exposed to endotoxin have been found to display significantly less arched-back nursing with significantly more time spent off nest (Walker et al. 2004a). Others have found endotoxin exposure to have no effect on maternal behavior (Spencer et al. 2006). During the present study, maternal behavior was not assessed, as such we cannot rule out maternal care as a contributing factor. Therefore, future studies should assess maternal behavior in parallel with the offspring's functional and physiological parameters.

The use of microarray in the present study was largely exploratory with a focus on potential changes to GR-related genes; as such the results highlight many areas for further study. For example, given both males and females showed changes to dopamine-related genes after infection, and a closer examination of dopamine-responsive hormones/neuropeptides is warranted. The results of the microarray analysis also showed changes to 11β-HSD1 and genes related to immune function which were not directly measured by qRT PCR. In addition, the analysis of transcription factors that could mediate the impact of infection on hippocampal gene expression may help to elucidate mechanisms behind the changes seen in the present study. Given the limited number of genes confirmed by qRT PCR, the present findings should be regarded as a preliminary indication of adult outcomes after neonatal infection.

Much of the previous work that has investigated immune factors and the hippocampus has focused on the role of the immune response in the development of Alzheimer's disease, specifically the potential preventative effects of a Th2 bias. It has been found that the administration of a Th2-biased Amyloid β-specific T-cell vaccine leads to a significant decrease in plaque-associated microglia and less vascular amyloidosis (Cao et al. 2009). The beneficial effects are only observed with the induction of a Th2 immune response. This is evidenced by the original human trials being halted due to the adverse effects of inflammatory Th1 responses caused by components in the formula of the vaccine (Gilman et al. 2005); subsequent trials with an altered formula for the vaccine did not result in any Th1 immune response (Bayer et al. 2005). In the present study, both sexes show a pervasive up-regulation of immune-related genes with males having more genes related to gene transcription and cellular development, and females showing a higher number of altered immune-related genes with a Th2 bias. The study provides evidence that the altered immune profile after neonatal infection persists and may indicate future directions for preventative intervention of neurodegenerative disease.

Early life infection has been found to result in altered patterns of neurogenesis in the hippocampus (Sharma et al. 2002; Cui et al. 2009), deficits in long-term potentiation (Lante et al. 2007; Lowe et al. 2008), and decreased hippocampal volume (Ikeda et al. 2005; Fatemi et al. 2009). These physiological outcomes result in increased anxiety-like behavior, and decreased learning and memory. Taking into account, the impact of early life infection on adult health predispositions could facilitate health practitioners in prevention, diagnosis, and treatment of adult disease. The present study shows that neonatal infection has effects on adult hippocampal gene expression that differ between males and females. It is clear that early life experiences influence brain development and physiology, which in turn has fundamental implications for an individual throughout their lifespan.

Acknowledgements

This work was funded by grants from the National Health and Medical Research Council of Australia (401238 and 569219).

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.