C57BL/6 and Kunming mice were purchased from the Center for Disease Control of Hubei Province, China. The animals were housed five per cage at room temperature, with a 12-h light/dark cycle and free access to water and food. This study used 5- to 6-week-old mice. The experiments were performed in accordance with the Guide for the Care and Use of Laboratory Animals (14), and the protocols received approval from the Hubei Province Committee on Laboratory Animal Care (HZAUMU2013-0005).

Purified LPS extract from S. typhimurium (Sigma, L6511) was dissolved in sterile phosphate-buffered saline (PBS) at 1 mg/ml and frozen in aliquots at −20°C. Animals were administered LPS via intraperitoneal injection at 10 mg/kg according to previous studies (15–17). Animals were infected with 5 × 10⁶ colony-forming units (CFU) of S. typhimurium via intraperitoneal injection.

For identification of polymorphism in human miR-155, the data of 1,000 genome sequence were examined (18, 19). Four candidate SNPs in the pre-miR-155 were selected for functional evaluation. In mice, the 256-bp DNA fragment containing the mature miR-155 sequence was amplified using DNA samples from C57BL/6 and Kunming mice. Polymerase chain reaction (PCR) was performed in 10-μl reactions containing 10 × PCR buffer, 0.3 μM of each primer, 75 μM of dNTPs, 1 U of Taq DNA polymerase (Takara Biotechnology), and 50 ng of mouse genomic DNA. The PCR cycling conditions were 5 min at 94°C, followed by 36 cycles of 30 s at 94°C, 30 s at 60°C, and 20 s at 72°C and a final extension step of 5 min at 72°C. The polymorphisms were identified based on the sequencing results. Ultimately, four SNPs were identified in mice, and these SNPs were closely linked and formed two haplotypes.

A total of 552 mice from 53 crosses (AB × AB) were used for correlation analysis. Because one of the SNPs in the haplotype could be detected using the restriction enzyme StuI, the genotype of the population was determined using the PCR-restriction fragment length polymorphism (RFLP) method. For phenotype detection, peripheral blood from each mouse at 5–6 weeks of age was collected into an anticoagulant vacuum blood tube. A total of 24 blood parameters were measured using a hematological analysis system (Sysmex XT-2000i, Sysmex, Japan); the tested parameters included white blood cell count (WBC), absolute neutrophil count (NEUT), absolute lymphocyte count (LYMPH), absolute monocyte count (MONO), absolute eosinophil count (EO), absolute basophil count (BASO), neutrophil percentage (NEUTp), lymphocyte percentage (LYMPHp), monocyte percentage (MONOp), eosinophil percentage (EOp), basophil percentage (BASOp), red blood cell count (RBC), hemoglobin concentration (HGB), hematocrit (HCT), mean cell volume (MCV), mean corpuscular hemoglobin (MCH), mean corpuscular hemoglobin concentration (MCHC), red cell distribution width-SD (RDW-SD), red cell distribution width-CV (RDW-CV), platelet count (PLT), plateletcrit (PCT), mean platelet volume (MPV), platelet distribution width (PDW), and platelet-large cell ratio (PLCR). Moreover, 52 and 43 mice harboring the AA and BB genotype, respectively, were selected for analysis of the variations in blood parameters in response to LPS treatment. The mice were intraperitoneally injected with 10 mg/kg LPS. After 8 h of LPS stimulation, hematological parameters were measured. Analysis of the associations between genotypes and blood parameters was performed using the following GLM model in SAS (SAS Institute, Inc., Cary, NC, USA): Y = genotype + sex + e; sex was considered as a fix effect in this model.

For northern blotting analysis, total RNA was extracted using TRIzol reagent according to the manufacturer’s protocol (Invitrogen, USA). From 15 to 30 μg of total RNA was electrophoretically separated on a 15% polyacrylamide denaturing gel. The total RNA was then transferred to a Hybond-N+ membrane (Amersham Biosciences, UK) using a semidry Transblot electrophoresis apparatus (Bio-Rad, USA). The RNA was fixed to the membrane by heating the membrane in the oven at 80°C for 30 min and then cross-linked to the membrane via UV irradiation. Hybridization was performed using PerfectHyb™ Hybridization Solution according to the manufacturer’s protocol (Toyobo, Japan). The hybridization probe sequence was complementary to the mature form of miR-155 (Table S1 in Supplementary Material) and was labeled with γ-³²P. After washing, the membranes were imaged using a phosphor imager (Bio-Rad, USA). U6 was used as a control and was detected in a manner similar to miR-155. The signals of the northern blotting bands were quantified using Quantity One software (Bio-Rad, USA).

Spleen, lung, and liver tissues were collected from Kunming mice harboring the AA or BB genotype following LPS treatment or no treatment. After 24 h of fixation, the spleen, lung, and liver tissues were dehydrated and embedded in paraffin wax. Then, 4-μm tissue sections were sliced using a microtome (Leica Microsystems Nussloch GmbH, Nussloch, Germany), mounted on poly-lysine-coated slides (Boster, Wuhan, China), and stored at 4°C until staining. Afterward, the sections were deparaffinized in xylene, rehydrated in a graded ethanol series, and stained with hematoxylin and eosin. Finally, the sections were mounted using neutral gum. Staining was examined via light microscopy (Olympus BX51, Tokyo, Japan).

To compare the cytokine levels between the two genotypes, we first collected serum from peripheral blood. In brief, whole blood from the eye socket of each mouse was collected into clean tubes. The tubes were then placed at an angle of 45° to 60° and incubated for 1 h at 37°C. Then, the tubes were centrifuged at 3,000 rpm for 5 min at room temperature. The supernatant was transferred to a new tube and centrifuged further at 12,000 rpm for 2 min at room temperature. Then, the purified supernatant was stored at −80°C. For cytokine (IL1β, TNFα, IL6, and IL8) detection, mouse ELISA kits were used, and all of the experimental steps were performed strictly according to the manufacturer’s protocol (NeoBioscience, Guangzhou, PR China). The fluorescence at 450 nm was measured using a microplate reader (iMark, Bio-Rad, USA). The concentration of each cytokine present in the samples was calculated in reference to a standard curve that was constructed using recombinant cytokines provided with each kit.

Total RNA was obtained from spleen tissues after 0, 4, and 8 h of LPS treatment using TRIzol reagent (Invitrogen, USA). For each genotype, equal amounts of RNA from three individuals at each time point were pooled together. Ultimately, six RNA libraries were generated, and RNA-seq was performed by a commercial service (Genergy Biotechnology, Shanghai, PR China). The mRNA sequencing data were analyzed using Bowtie, TopHat, and Cufflinks (20). The reference genome version was GRCm38/mm10.

After determining the RNA concentrations, reverse transcription was performed using a RevertAid First Strand cDNA Synthesis Kit (MBI Fermentas). For quantitative PCR (Q-PCR) analysis of miR-155, stem–loop RT primers were designed according to a previous report (21). Q-PCR was performed using a standard SYBR Green PCR kit (Toyobo, Japan) in a LightCycler 480 thermal cycler (Roche, Switzerland). The PCR protocol was 2 min at 95°C followed by 40 cycles of amplification (30 s at 95°C, 30 s at the annealing temperature, and 30 s at 72°C). Melting curves were obtained by increasing the temperature from 58 to 95°C at 0.5°C/s and then holding for 10 s. The U6 or tubulin gene was used as an internal control for the genes expression of miR-155, SHIP1, and PU.1. PCR was performed at least in triplicate, and relative gene expression levels were calculated using an optimized comparative Ct (ΔΔCt) method. The primers used for Q-PCR detection are listed in Table S1 in Supplementary Material. The significance of the differences in gene expression was analyzed using a t-test.

To confirm the differential protein expression of the two important hematopoietic-related miR-155 target genes between the two genotypes of mice, spleen tissue extracts were fractionated according to molecular weight via sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred to a polyvinylidene difluoride (PVDF) membrane using a semidry transfer apparatus (Bio-Rad, USA). Western blotting was performed using the following antibodies: anti-PU.1 (9G7), anti-SHIP1 (D1163), anti-GAPDH (14C10), and horseradish peroxidase (HRP)-conjugated anti-rabbit IgG (Cell Signaling Technology, USA). All of the antibodies were diluted 1,000-fold. The signal of the western blotting bands was quantified using ImageJ software (National Institutes of Health, USA).

The 256-bp mouse miR-155 expression cassette containing the entire pre-miR-155 sequence and partial flanking regions was inserted into the 3′ untranslated region (UTR) of the green fluorescent protein (GFP) gene in the pEGFP-C1 vector (BD Biosciences Clontech, USA). The two haplotypes (A and B) and seven mutant constructs (M1–M7) were produced and were labeled as pEGFP-C1-A/B/M1–M7. The 255-bp human miR-155 fragments were inserted into the 3′UTR of the GFP gene as performed on mouse miR-155. The normal and four mutant constructs of human miR-155 gene were generated and were labeled as pEGFP-C1-h or pEGFP-C1-h1–4, respectively. For the luciferase assay, the 3′UTR fragments of the human and mouse Tab2, Bach1, Ikbke, and Map3k14 genes were inserted into the psi-check2 vector. The inserted fragments contained the validated miR-155 binding sites located at the 3′UTR of the respective target genes.

BHK-21 cells were cultured in complete Dulbecco’s modified eagle medium (DMEM) containing 10% fetal bovine serum (FBS) in a humidified incubator in 5% CO₂ at 37°C. The different constructs were transfected at equal plasmid DNA concentrations into BHK-21 cells at 80–90% confluence. For analysis of miR-155 expression in vitro, the pEGFP-C1-A, pEGFP-C1-B, pEGFP-C1-M1–M7, and pEGFP-C1 empty constructs were transfected into BHK-21 cells at equal plasmid DNA concentrations. For the luciferase activity assay, equal amounts of the pEGFP-C1-A, pEGFP-C1-B, pEGFP-C1-h, and pEGFP-C1-h1–4 constructs were co-transfected with one of the psi-check2-Tab2/Bach1/Ikbke/Map3k14 vectors into BHK-21 cells.

All data are presented as the means ± SEM. Significant differences between groups were determined using a t-test or analysis of variance (ANOVA). A P-value <0.05 was considered to indicate a statistically significant difference, and a P-value <0.01 was considered to indicate an extremely significant difference.

In mice, a 256-bp genomic fragment of miR-155 was amplified using DNA samples from non-biologically related Kunming (n = 25) and C57BL/6 mice (n = 20). The sequencing results for the different strains of mice revealed four SNPs in the 256-bp fragment, which formed two haplotypes (Figure 1A, Supplementary Figure 1). Furthermore, Kunming mice harbored the AA/AB and BB genotypes, but C57BL/6 mice harbored only the BB genotype. The secondary structures of miR-155 containing different SNPs were predicted using mfold (http://unafold.rna.albany.edu/?q=mfold). According to these results, two SNPs in miR-155, one in the loop region of miR-155 and the other near the 3′ end of pre-miR-155, influenced the secondary structure of pre-miR-155 in mice (Figure 1B). In humans, according to results of 1,000 human genomic sequence, four SNPs located in the pre-miR-155 regions were identified (Figure 2A, Supplementary Figure 2). The first SNP was located in the mature miR-155 region, the second in the stem–loop regions of miR-155, and the third and the fourth in the miR-155* region. Of these four SNPs, only the third SNP changed the secondary structure of pre-miR-155 (Figure 2B).

To explore the differences in the functions of the two haplotypes, a trait association study was first performed. Because the second mutation site of the haplotype could be detected using the restriction enzyme StuI (GGCC), the genotypes of 552 Kunming mice were identified using the PCR–RFLP method. The results of the trait association analysis revealed that the A/B haplotype was significantly associated with 10 different blood parameters. Among them, the WBC, NEUT, LYMPH, and HGB were significantly increased in the AA genotype compared with the BB genotype (P < 0.05) (Figure 3A, Table S2 in Supplementary Material). Following LPS treatment, 14 blood parameters were significantly different between the A and B haplotypes. The mean BASO, MONO, EO, BASOp, MONOp, and EOp values were increased in BB genotype mice (P < 0.05). The mean values of HGB, HCT, MCV, MCH, PLT, PCT, MPV, and PLCR were significantly increased in AA genotype mice (P < 0.05) (Figure 3B, Table S3 in Supplementary Material). In both AA and BB genotype mice, 18 different blood parameters were significantly altered in response to LPS treatment compared with the corresponding untreated control mice. Among them, BASOp, NEUTp, MONOp, BASO, MCV, PDW, MPV, and PLCR were significantly increased in response to LPS stimulation compared with the control treatment among both AA and BB genotype mice (P < 0.05). In addition, the mean values of WBC, LYMPH, LYMPHp, RBC, HGB, HCT, MCHC, RDW-CV, PLT, and PCT were significantly reduced following LPS treatment in both genotypes of mice (P < 0.05) (Figure 3C).

Mice harboring the two genotypes were administered 10 mg/kg LPS via intraperitoneal injection. Splenomegaly symptoms were noted at 8 h after LPS injection (Figure 4A). The characteristics of the tissues were analyzed using histological analysis. Increased red pulp and reduced white pulp proportions were noted in the spleen tissue of the AA genotype mice compared with the BB genotype mice 8 h after LPS injection (Figure 4B). In the lung tissue, the alveolar wall was thicker in the AA genotype mice than in the BB genotype mice at 8 h of LPS stimulation (Figure 4C). Liver congestion and inflammatory cell infiltration symptoms were more serious in the AA genotype mice than in the BB genotype mice at 8 h of LPS stimulation (Figure 4D). We also performed TUNEL assays to compare the injury to these two genotype mice under LPS treatment. Our results showed that the proportion of apoptotic cells in the spleen and liver tissues of AA genotype mice was significantly higher than that of BB genotype mice after LPS stimulation for 8 h (Supplementary Figures 3A–C). The ELISA results revealed increased IL1β, TNFα, IL6, and IL8 levels upon LPS treatment. The IL1β, TNFα, and IL6 levels in AA genotype mice were significantly increased compared with BB genotype mice at 8 or 4 h of LPS exposure (Figures 4E–H). RNA-seq data from the spleen tissues showed that the transcriptional levels of all the detectable inflammation related cytokines were up-regulated after 4 or 8 h of LPS treatment. Moreover, the IL1β, TNFα, IL6, IL1α, IL18, CCL4, CCL5, and CCL3 genes were relatively higher in spleen tissues from the AA genotype mice than in those from the BB genotype mice at 0 h. While the transcriptional levels of CXCL11, IL10 and IL12b genes were relatively lower in spleen tissues from the AA genotype mice at 0 h. However, for all the genes, not much difference were found at mRNA level between AA and BB genotype mice at 4 or 8 h of LPS exposure (Supplementary Figure 3D). Further, the AA genotype mice exhibited significantly reduced survival time after infection with S. typhimurium compared with the BB genotype mice (P < 0.05) (Figure 4I). All of these results indicated that the inflammatory response was stronger in the AA genotype mice than in the BB genotype mice following LPS treatment or pathologic infection.

The expression levels of miR-155 in the spleen and lung were detected by northern blotting. The results showed that miR-155 was upregulated sharply, reaching its highest level at 8 h of LPS exposure, and then declined to approximately the normal level at 24 h of LPS exposure in the spleen and lung tissues. The trend of miR-155 expression following LPS treatment was similar in between AA and BB genotype mice (Figures 5A,B). Moreover, miR-155 expression in spleen tissue of AA genotype mice was approximately 1.5- to 2-fold higher than that of BB genotype mice at 0 and 8 h of LPS exposure. Additionally, miR-155 expression in the lung tissue was significantly increased in AA genotype mice compared with BB genotype mice at 0 and 8 h of LPS exposure (Figures 5C–H). These results indicated that miR-155 was rapidly upregulated under LPS treatment, and that AA genotype mice exhibited increased miR-155 expression compared with BB genotype mice under both LPS treatment and non-treatment conditions.

To further analyze the difference in miR-155 expression between the AA and BB genotypes, we analyzed the expression profiles of its target genes. Equal volumes of RNA from three individuals harboring each genotype who were exposed to LPS for 0, 4, or 8 h were pooled together, and the expression profiles of miR-155 target genes were predicted using TargetScan software and were determined using RNA-seq. The results showed that 76.99% (174 genes), 73.01% (165 genes), and 58.85% (133 genes) of the target genes were downregulated in AA genotype mice compared with BB genotype mice at 0, 4, and 8 h of LPS exposure, respectively (Figures 6A–C, Table S5 in Supplementary Material). These results indicated that most of the miR-155 target genes were downregulated at 0 and 4 h of LPS exposure in the AA genotype mice compared with the BB genotype mice. We further analyzed the signaling pathways regulated by the 20 downregulated miR-155 target genes [fold-change (FC) ≥ 1.2] at 4 h of LPS exposure using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database (Figure 6E). The results showed that the TCR, BCR, MAPK, insulin, and Wnt pathways as well as various cancer signaling pathways were targeted by miR-155 (Figure 6D, Table S4 in Supplementary Material).

To further confirm the differential expression of miR-155 between the two genotypes of mice, we measured the protein expression of two miR-155 target genes, SHIP1 and PU.1, which are closely related to hematopoiesis. The protein levels of these two genes in the spleen tissue were significantly reduced in AA genotype mice compared with BB genotype mice at 0 and 8 h of LPS exposure (P < 0.05) (Figures 7A–D). Additionally, the mRNA levels of the SHIP1 and PU.1 genes were detected using Q-PCR. The results revealed a non-significant pattern of downregulated expression of these two genes in AA genotype mice compared with BB genotype mice (Figures 7E,F).

To further confirm the differential expression of miR-155 between the A and B haplotypes in mice, we cloned the 256-bp DNA fragment that contained the entire pre-miR-155 sequence into the pEGFP-C1 vector. The constructed A and B haplotype vectors were labeled pEGFP-C1-A and pEGFP-C1-B, respectively. The two constructs were transfected into BHK-21 cells at equal plasmid DNA concentrations. Then, the expression levels of pre-miR-155 and mature miR-155 were detected by northern blotting. The expression of pre-miR-155 and mature miR-155 from the A haplotype construct was significantly higher by approximately 1.5-fold than their expression from the B haplotype construct (P < 0.05) (Figures 8A,B). Moreover, the Q-PCR results confirmed that miR-155 expression from the A haplotype construct was significantly higher than that from the B haplotype construct (P < 0.05) (Figure 8C). In addition, the dual-luciferase assay results indicated that both the A and B haplotype constructs inhibited the expression of the miR-155 target genes Tab2, Bach1, Ikbke, and Map3k14. Moreover, the A haplotype construct had a significantly stronger effect than the B haplotype construct (P < 0.05) (Figure 8D). In vitro experiments indicated that the A haplotype construct generated more mature miR-155 than the B haplotype construct.

The 255-bp human DNA fragments, including the SNP sites of miR-155, were inserted into the pEGFP-C1 vector; the resulting constructs were labeled as pEGFP-C1-h/h1/h2/h3/h4. The dual-luciferase assay results indicated that the pEGFP-C1-h/h1/h2/h3/h4 constructs inhibited the expression of the human miR-155 target genes Tab2, Bach1, Ikbke, and Map3k14. Notably, the pEGFP-C1-h2 and pEGFP-C1-h3 constructs had significantly weaker effects on these four miR-155 target genes than pEGFP-C1-h (P < 0.05). However, the other two constructs did not have clear effects on the expression of these four miR-155 target genes (Figure 8E).

To further explore whether functional SNPs in the haplotype contributed to the different expression levels of the mouse miR-155 gene, we designed a series of mutant constructs (M1–M7). In these constructs, SNPs in the A haplotype were converted to the corresponding nucleotide in the B haplotype individually in constructs M1–M4. In M5, the two middle SNPs of the A haplotype were converted to the corresponding nucleotide in the B haplotype synchronously. In M6 and M7, the two middle SNPs of the B haplotype were individually converted to the corresponding nucleotide in A haplotype (Figure 9A). All of the mutant and normal A and B haplotype constructs were transfected into BHK-21 cells at equal plasmid DNA amounts (Figure 9B). The expression level of miR-155 was detected by northern blotting 24 h after transfection. The expression of miR-155 from the M1 and M4 constructs was similar to that from the A haplotype construct and was significantly increased compared with that from the B haplotype construct (Figures 9C,D). The expression of miR-155 from the M2, M3, and M5 constructs was significantly reduced compared with that from the A haplotype construct (Figures 9E,F). The expression of miR-155 from the M6 and M7 constructs was significantly increased compared with that from the B haplotype construct but was slightly reduced compared with that from the A haplotype construct (Figures 9G,H). The results of this point mutation analysis confirmed that the two middle SNPs contributed to the differential expression of the mouse miR-155 gene between the A and B haplotypes. The expression level of the mouse miR-155 gene decreased when these two SNPs of the A haplotype were converted to the corresponding nucleotide in the B haplotype and increased when these two SNPs of the B haplotype were converted to the corresponding nucleotide in the A haplotype.