Expression of Toll-like receptors and their association with cytokine responses in peripheral blood mononuclear cells of children with acute rotavirus diarrhoea

Abstract

To understand virus and host interactions and host responses to rotavirus infection in children, we analysed by real-time polymerase chain reaction (PCR) the expression of mRNA for five Toll-like receptors (TLRs) (TLR2, TLR3, TLR4, TLR7 and TLR8) and four T helper (Th)1 and Th2 cytokines [interleukin (IL)-2, IL-12, interferon (IFN)-γ and IL-4) in peripheral blood mononuclear cells (PBMC) of children with acute rotavirus diarrhoea. We observed significantly higher expression of genes encoding TLR2, TLR3, TLR4, TLR7 and TLR8 in PBMC of 41% (31/75) patients within 3 days of illness onset than those in healthy children. After 3 days of illness onset, only TLR3 and TLR8 mRNA expressions were still significantly (P< 0·05) increased in 59% (44/75) children with diarrhoea. We also observed significantly (P< 0·05) elevated expression of IL-12p40 and IFN-γ in PBMC of patients during the entire period of illness and the first 3 days of illness, respectively. We further demonstrated a weak but significant association between elevated levels of gene expression of four TLRs (TLR2, TLR3, TLR4 and TLR8) and IFN-γ. Our results suggest that multiple TLRs may modulate the immune response in the acute phase of rotavirus infection and play a role in the activation of IFN-γ.

Introduction

Toll-like receptors (TLRs) are primary transmembrane proteins of immune cells that play a critical role in innate and adaptive immunity. Currently, 13 TLRs (TLR1–TLR13) have been identified in mammalian species, including 11 in humans [1]. TLRs recognize small conserved molecular sequences, termed pathogen-associated molecular patterns (PAMPs) in membranes, cell walls, proteins, DNA and RNA of microorganisms, including viruses, trigger innate immune response and mediate the adaptive immune response [2]. TLRs recognize viruses and respiratory syncytial virus (RSV) and mouse mammary tumour virus (MMTV) induce immune responses through TLR4 [3–5]. Measles virus and human cytomegalovirus (HCMV) activate inflammatory responses via TLR2 [6,7]. TLR-3-deficient macrophages have greatly reduced responses to polyinosine–polycytidylic acid (poly I : C), which structurally mimics double-stranded viral RNA [8]. Heil and colleagues reported that murine TLR7 and human TLR8 recognized analogues of ssRNA, suggesting that the ssRNA virus may initiate immunity via TLR7 and TLR8 [9,10]. There have been no studies on the identification and function of TLRs in response to rotavirus, the single most important pathogen that causes severe dehydrating diarrhoea in young animals and children.

TLRs have been described to be involved in cytokine production and cellular activation in response to a pathogen. When T helper (Th) cells are activated, they differentiate mainly into two functionally distinct subsets, Th1 and Th2 cells. Interleukin (IL)-2, IL-12 and interferon (IFN)-γ are representatives of Th1 cytokines, whereas IL-4, IL-10 and IL-13 are typical cytokines produced by Th2 cells [11]. Neonatal mice immunized with rotavirus develop a mixed Th1 and Th2 response, while immunization of adult mice seems to lead to a predominant Th1 response [12]. Several recent studies have demonstrated that both Th1 and Th2 cytokines − IL-6, IL-10, tumour necrosis factor (TNF)-α, and IFN-γ − are released in serum or plasma of children with rotavirus diarrhoea [13,14]. Another study showed that in children and adults with rotavirus infection the frequencies of CD4⁺ cells secreting IFN-γ were higher than those of cells secreting IL-13, suggesting that children with rotavirus diarrhoea primarily mount a Th1 response [15]. IFN-γ was also described to play a role in protection against rotavirus disease [13].

Despite various studies in animals with TLR deficiency or in cell cultures that have examined the role of single TLR in viral infection and immunity, it is not clear whether the findings in animals or in vitro can reliably reflect complex mechanisms of pathogenesis and immunity in humans. In addition, equivalents of certain TLRs found in humans are not present in animals. In this study, we sought to investigate directly the mRNA expression of five TLRs (TLR2, TLR3, TLR4, TLR7 and TLR8) and four Th1 and Th2 cytokines (IL-2, IL-12, IFN-γ and IL-4) in peripheral blood mononuclear cells (PBMC) of children with acute rotavirus diarrhoea. We further examined the correlation of expression levels of TLRs and cytokines. The findings in this study may increase our understanding of virus and host interactions and host responses to rotavirus infection in children.

Materials and methods

Study population and specimen collection

From September 2004 to January 2005, we collected blood and faecal specimens from 75 children under 3 years of age who were treated for acute rotavirus gastroenteritis at Children’s Hospital, Fudan University, Shanghai, China. All children were otherwise in generally good health and had not been vaccinated against rotavirus. Detailed symptom data, including fever, vomiting, diarrhoea and dehydration, were documented during the entire period of illness and on the day of blood collection. Thirty-eight healthy children without diarrhoea or other infectious disease were also enrolled as controls. Blood and faecal samples were collected from each child. Stools were stored at −20°C before being tested for rotavirus by enzyme immunoassay (EIA). Faecal specimens from patients with diarrhoea were tested positive for rotavirus by a commercial EIA kit (Anqun, Shenzhen, China) and those from control subjects were negative. PBMC were obtained from freshly collected blood in heparinized tubes, separated on Ficoll gradients (Huamei, Shangahi, China), and washed free of platelets and Ficoll. Cells were mixed with 1 ml Trizol reagent (Invitrogen, CA, USA) for RNA purification and stored at −70°C. The study was approved by the Ethics Committee of the Children’s Hospital of Fudan University.

Expression of mRNA for TLRs and cytokines

The concentration of total RNA extracted from PBMC was determined by measuring optical density at 260 nm. RNA (2 µg) was reverse-transcribed using oligo(dT), random hexamers and Moloney murine leukemia virus (MMLV) reverse transcriptase (Invitrogen) in a final volume of 20 µl. The resulting cDNA was diluted 1 : 5 and 5 µl was used for polymerase chain reaction (PCR) amplification. Expression analysis was performed by quantitative real-time PCR with SYBR Green I (Roche, Mannheim, Germany) on ABI 7000 (Applied Biosystem Inc., CA, USA). Primers for IL-2, IL-4, IL-12, IFN-γ, TLR2, TLR3, TLR4, TLR7 and TLR8 were designed by software primer express (ABI) and their specificity was examined with blast software on National Center for Biotechnology Information (NCBI) (Table 1). Thermal cycler parameters included heating at 95°C (20 min) and 40 cycles at 95°C (30 s) and 60°C (1 min). Gene-specific standards in serial dilutions were co-amplified and all gene-specific mRNA expression values were normalized against the housekeeping gene β-actin. After amplification, the melting curves were generated automatically by ABI 7000, only those that have single high production peak were considered to be valid amplification. Negative controls without sample templates had only a low peak for primer dimmers. This method helped to determine if the PCR products are specific.

Table 1.

Primers used for real-time polymerase chain reaction.

	Reverse primer	Forward primer
TLR2	ctgcaagctgcggaagataat	aggactttatcgcagctctcaga
TLR3	gccaggaatggagaggtctaga	gttacgaagaggctggaatggt
TLR4	gattgctcagacctggcagtt	tgtcctcccactccaggtaagt
TLR7	gtggaaattgccctcgttgt	tgtcagcgcatcaaaagcat
TLR8	tgcgctaccaccttgaagaga	tctgcatgaggttgtcgatga
IL-2	ctgctggatttacagatgattttga	gtggccttcttgggcatgt
IL-12p40	gcaaaaccctgaccatccaa	gtgaagcagcaggagcgaat
IFN-γ	tgtccaacgcaaagcaatac	tcgacctcgaaacagcatct
IL-4	cagcagttccacaggcacaa	ctctggttggcttccttcaca

IFN: interferon; IL: interleukin; TLR: Toll-like receptor.

Statistical analysis

Two-tailed Student’s t-test was performed to compare the expressions of TLRs and cytokines between different groups. Pearson’s correlation was applied to analyse the association of expressions of TLRs and cytokines. The level of significance was set at P≤ 0·05. Data analysis was carried out with the Statistical Package for Social Sciences (spss version 11·5 for Windows).

Results

Characteristics of subjects

Of the 75 patients enrolled in this study, 54 were boys and 21 were girls (Table 2). The age of these children ranged from 3 to 31 months (12·1 ± 5·8 months). The 38 healthy controls (26 boys and 12 girls) ranged in age from 1 month to 60 months (15·7 ± 5·0 months). No significant difference was found in age or gender ratio between patients and healthy controls. All 75 patients had typical clinical symptoms of acute rotavirus gastroenteritis, including fever, vomiting and diarrhoea. The number of stools per day was 1–30 (10·0 ± 5·3) and the duration of diarrhoea was 1–23 days (8·9 ± 4·3 days). The duration of vomiting was 0–6 days (1·7 ± 1·6 days); the percentage of patients with dehydration and fever was 80% and 72%, respectively. Onset of illness at the time of blood collection was 1–14 days (4·7 ± 2·7 days). Thirty-one patients were enrolled within 3 days of illness onset, 36 were between 4 and 8 days of onset and eight were between 9 and 14 days. Patients within 3 days of illness onset had shorter duration days of diarrhoea than those with more than 4 days of illness onset, P = 0·004. No significance was found with other clinical symptoms between these two groups.

Table 2.

Clinical characteristics of children with rotavirus diarrhoea.

	Onset day of illness (day)
Variables	< 4 (n = 31)	≥ 4 (n = 44)	1–14 (n = 75)
Age (month)	12·1 ± 5·1	12·0 ± 6·3	12·1 ± 5·8
Sex (male/female)	22/9	31/13	53/22
Maximum temp (°C)	38·5 ± 1·3	38·7 ± 1·2	38·6 ± 1·3
Duration of diarrhoea (day)	7·2 ± 3·3*	10·1 ± 4·5	8·9 ± 4·3
No. of stools per day	9·2 ± 5·2	10·5 ± 5·3	10·0 ± 5·3
Duration of vomiting (day)	1·5 ± 1·5	1·8 ± 1·6	1·7 ± 1·6
Onset day of illness	2·5 ± 0·6	6·3 ± 2·6	4·7 ± 2·7
No. of patients with vomiting/total (%)	20/31 (65%)	31/44 (70%)	51/75 (68%)
No. of patients with dehydration/total (%)	26/31 (84%)	34/44 (77%)	60/75 (80%)

Data are presented as means ± standard deviations. Patients within 3 days of illness onset had shorter diarrhoea days than those with more than 4 days of illness onset.

^*P< 0·05.

Characterization of rotavirus strains from faecal specimens

All faecal specimens from patients were tested positive for rotavirus by EIA. The G and P genotypes of all rotavirus strains in faecal specimens were determined by reverse transcription–polymerase chain reaction (RT–PCR) [16,17]. Of the 75 stool specimens examined for rotavirus strains by RT–PCR, 50 (67%) were P(8),G3, five (6%) were P(8),G1, four (5%) were P(8),G2 and one (1%) each was P(4),G3 and P(9),G3. Three (4%) were P8 but had mixed G2 and G3 strains. One specimen could not be G-typed and nine were not P-typeable. One specimen was not G- and P-typeable.

Expressions of TLR mRNA in PBMC of patients with rotavirus diarrhoea

We first examined the kinetics of expression of genes coding for several TLRs in PBMC of children with acute rotavirus diarrhoea and compared these with those of healthy controls. As the sample size was not large enough to examine the kinetics on a daily basis, we analysed mRNA expressions of TLRs in PBMC collected in the first 3 days of illness (n = 31) and 3 days after illness onset (n = 44). The 31 patients had significantly higher levels of mRNA expression for TLR2, TLR3, TLR4, TLR7 and TLR8 than that in 38 healthy controls. In contrast, only TLR3 and TLR8 mRNA expressions were still significantly (P< 0·05) increased after 3 days of illness onset in 44 children with diarrhoea (Fig. 1).

Fig. 1.

Expression of genes encoding five Toll-like receptors (TLRs) in peripheral blood mononuclear cells of children with acute rotavirus diarrhoea. Of the 75 blood specimens from patients, 31 were collected in the first 3 days of illness and 44 were collected 4 days after illness onset. Levels (copy number/100 ng RNA) of TLR gene expression are presented in logarithmic scale. Expressions of all five TLRs were significantly elevated in the first 3 days of illness, whereas expressions of TLR3 and TLR8 remained significantly elevated after 3 days of illness onset compared with those of 38 healthy controls. * and #: P < 0.05.

Expressions of cytokine mRNA in PBMC of patients with rotavirus diarrhoea

We also demonstrated higher levels of expression for cytokine genes in PBMC of patients with rotavirus diarrhoea than in those of healthy controls. Expression of mRNA for IL-12p40 was significantly (P< 0·05) higher in the entire period of illness. Levels of mRNA expression of IFN-γ were also significantly higher in PBMC collected in the first 3 days of illness but returned to levels close to controls after 3 days (Fig. 2). Expression levels for IL-2 and IL-4 were little changed during the entire period of illness.

Fig. 2.

Expression of genes encoding four Toll-like receptors (TLRs) in peripheral blood mononuclear cells of children with acute rotavirus diarrhoea. Of the 72 blood specimens examined, 30 were collected in the first 3 days of illness and 42 were collected 4 days after illness onset. Levels (copy number/100 ng RNA) of Toll-like receptor gene expression are presented in logarithmic scale. Interleukin (IL)-12p40 and interferon-γ expression levels were stimulated in the first 3 days of illness and only IL-p40 remained significantly elevated after 3 days of illness onset when compared with those of 38 healthy controls. * and #: P < 0.05.

Correlation of mRNA expressions of TLRs and cytokines

We last examined whether elevated levels of TLR expression were associated positively with cytokine responses in PBMC of patients with acute rotavirus diarrhoea. Of the five TLRs examined, four (TLR2, TLR3, TLR4 and TLR8) with elevated levels of expression had a weak but significant association with increased IFN-γ response during the entire period of illness (Fig. 3). No positive associations were observed between elevated expression of TLR7 and IFN-γ responses.

Fig. 3.

Association of mRNA expressions of Toll-like receptors (TLRs) with interferon-γ response in children with acute rotavirus diarrhoea. Levels (copy number/100 ng RNA) of TLR gene expression are presented in logarithmic scale. Gene expression and statistical analysis are described in the text. Pearson’s correlation and coefficient (r) and P-values are presented.

Discussion

This is the first study to examine mRNA expression of TLRs and their relationship with cytokine responses in patients with acute rotavirus diarrhoea. We observed increased expression of five TLRs and demonstrated positive associations between four of these and IFN-γ response. Expression of TLRs was time-dependent; all five TLRs (TLR2, TLR3, TLR4, TLR7 and TLR8) were expressed at high levels during the first 3 days after the initiation of disease, but only TLR3 and TLR8 remained elevated 3 days after the onset of illness. Of interest, the expression levels of TLR2, TLR3 and TLR8 were positively correlated with each other (data not shown).

TLRs recognize conserved microbial and viral signals to instigate immune responses. TLRs are polymorphic; one TLR detects not only one specific ligand, but could also be the common receptors of various ligands. On the other hand, any given pathogen may express several different PAMPs, resulting in a combined recognition by multiple TLRs [2,18]. For example, TLR3, TLR7 and TLR8 have been described to mediate the immune response to dsRNA or ssRNA viruses [8–10,19–21]. TLR2 can be activated by various PAMPs, including glycolipids, outer membrane protein, phenol-soluble modulin of bacterial origin, cytomegalovirus (CMV) virions and haemagglutinin (H) protein of wild-type measles [6,7,22,23]. TLR4 is critical for host response to Gram-negative bacterial lipopolysaccharide (LPS) and several other ligands, including flavolipin produced by flavobacteria, mouse mammary tumour virus, RSV and even fibronectin and heat-shock protein produced by the host during immune responses [3–5,24,25].

Previous studies have demonstrated that combinations of TLRs are required for the recognition of components from bacteria [26,27]. Differential regulations of the expression of various TLR mRNAs were observed in mice with bacterial and viral infections [28]. Influenza virus has been described previously to mediate effects via TLR7, but a recent study established that TLR3 may also participate in the immune response [20]. RSV was shown to up-regulate TLR4 and sensitize the response to endotoxin in human epithelial cells [4]. TLR3 has been described to possibly contribute to the exacerbation of response to RSV in infected epithelial cells [29]. However, a more recent study indicated that absence of TLR3 neither altered viral pathogenesis nor impaired host adaptive responses to some viruses, suggesting that other TLRs may play a role in recognizing virus when TLR3 was not present [30]. Both TLR7 and TLR8 were described to be involved in the innate immune response to human parechovirus 1, an ssRNA virus [31]. Rotavirus is a typical dsRNA virus and when it infects intestinal epithelial cells, ssRNA, dsRNA, structural and non-structural proteins are produced and they may all lead to activation of multi-TLRs in immune cells. Elevated expression of TLR3 and TLR8 in PBMC of children with acute rotavirus diarrhoea suggest that these two TLRs may work together and play a major role in rotavirus pathogenesis and immunity.

We detected elevated expressions of genes for IL-12p40 and IFN-γ, but not for IL-4 in PBMC of children with acute rotavirus diarrhoea. IL-12 is an important proinflammatory cytokine and a key mediator for Th1 response as it induces expression of IFN-γ in natural killer (NK) cells and T cells [32]. IL-12 is composed of a heterodimer of p40 and p35; p40 is a highly inducible subunit that regulates the overall production of biologically active IL-12p70 [33]. IFN-γ, produced mainly by Th1, Tc cells and NK cells, plays an essential role in the development of Th1 response and host response to infection in adaptive immunity. IFN-γ is known to have anti-viral activities and has been described to help patients recover from viral infection [34–36]. Elevated levels of IFN-γ in serum of children with acute rotavirus diarrhoea were correlated with less severe disease [13]. IFN-γ secreted by rotavirus-stimulated PBMC of healthy adults inhibited rotavirus entry into human intestinal epithelial cells [37].

A positive correlation between elevated expression of several TLRs and IFN-γ responses suggests that TLRs may have contributed to the initiation of Th1 response, including the activation of IFN-γ, and modulated the immune response to rotavirus in children with acute diarrhoea. TLRs have been described to control multiple dendritic cell functions and activate signals that are critically involved in the initiation of adaptive immune response. They influence and induce Th1 adaptive immune responses to bacterial and viral components [38–41]. Future studies will need to investigate and understand further the immune responses to rotavirus infection, and especially the role of TLRs, in the mechanisms of disease and in the induction and regulation of innate and adaptive immunity in children.