Main

The limited functional capacity of the neonatal immune system increases the neonatal risk for infections (1), rendering the neonate at risk for certain complications (2,3,4,5). Preterm infants are especially susceptible to infections and morbidities that are linked to an unusual state of immune activation (2,6,7).

Although the determinants of immune activation in the premature infant are not well understood, in some cases, it may be related to in utero exposure to inflammation from chorioamnionitis. Microbial products (e.g., endotoxin and bacterial DNA) in the amniotic fluid have been detected in the presence of chorioamnionitis (8,9) and are related to a systemic inflammatory response in the fetus (10). Boggess et al. demonstrated in an animal model that chronic maternal exposure to Porphyromonas gingivalis resulted in maternal systemic dissemination and transplacental passage of bacterial products; fetal exposure was confirmed by detection of 16s DNA by PCR in both maternal and fetal compartments (11).

Antigen exposure is necessary to induce maturation of the adaptive immune system; thus, neonatal lymphocytes are predominantly naive and have a “resting” phenotype, as exposure to foreign antigen in utero is uncommon in uncomplicated pregnancies (1,12). Among preterm infants, however, T cells are often activated as reflected by increased proportions of cells expressing the activation markers CD25, CD69, and human leukocyte antigen D-related (HLA-DR) as compared with T cells of term infants (13).

In this study, we characterized the maturation phenotype and expression of homing receptors in T cells from the cord blood (CB) of term and preterm infants. T cells from preterm infants showed evidence of in vivo activation and evidence of maturation towards a memory phenotype. In addition, T cells from term infants (composed largely of naive T cells) but not naive T cells obtained from adult subjects can be activated in vitro by exposure to microbial Toll-like receptor (TLR) ligands to develop a more mature phenotype and to alter homing receptor expression. We suspect, therefore, that microbial products can play an important role in phenotypic maturation in the neonate but not in adults.

Results

The Proportion of CD45RO+ T Cells Is Higher in Preterm vs. Term CB

T cells were identified by size, granularity, and expression of CD45 and CD3 ( Figure 1a,b ). The proportion of CD3+CD45RO+ T cells ( Figure 2a ) was higher in preterm (median = 23.1%) vs. term CB (median = 2.7%) ( Figure 2b ). This difference was statistically significant (P = 0.0004).

Figure 1
figure 1

Whole-blood flow cytometric analysis of lymphocytes. CB was stained with fluorochrome-labeled antibodies followed by red blood cell lysis and fixation. (a) Box-plot with lymphocytes gated based on size (forward scatter, FSC) and granularity (side scatter, SSC). (b) Neonatal T cells are selected based on CD45 and CD3 coexpression (box-plot). CB, cord blood.

Full size image
Figure 2
figure 2

Increased proportion of CD3+CD45RO+ T cells in CB of preterm infants. (a) Whole-blood staining from term (upper histogram, 3.5%) and preterm CB (lower histogram, 35.9%). CD45+CD3+ cells were examined for expression of CD45RO. (b) Box and whiskers plot (median, minimum, and maximum values) of proportions of CD3+CD45+ cells expressing CD45RO in CB from term (n = 25) and preterm infants (n = 22). Statistical significance was determined by Mann–Whitney rank test (*P = 0.0004). CB, cord blood.

Full size image

Microbial TLR Ligands Increase the Expression of CD45RO in Neonatal T Cells but Not in Adult Naive T Cells

TLR ligands activate cells of the innate immune system in neonates and adults (14,15,16). As we previously described, TLR ligands also can drive T-cell activation. This effect that we described in adult cells was seen uniquely among memory (CD45RO+) T cells. No effect was observed in naive (CD45RA+) T cells (17). As we found an increased proportion of CD45RO+ T cells in CB of premature infants who are often exposed to microbial elements in the setting of chorioamnionitis, we hypothesized that bacterial TLR ligands might be activating neonatal CD45RA+ T cells to upregulate expression of CD45RO. TLR ligands were selected based on their expression by bacteria that commonly cause intrauterine infection and neonatal sepsis (18,19). Among CD45+ CD3+ lymphocytes, there was a significant increase in the proportion of CD3+ cells expressing CD45RO ( Figure 3a ). Incubation with flagellin resulted in the highest median percentage of CD45RO (14%), followed by macrophage-activating lipopeptide-2 (11%) and lipopolysaccharide (9%) ( Figure 3b ). By contrast, these TLR ligands did not induce CD45RO expression when incubated with naive T cells (CD45RO-depleted) from healthy adults ( Figure 4a,b ).

Figure 3
figure 3

TLR ligands increased the proportion of neonatal T cells expressing CD45RO. (a) CBMCs from term infants were cultured for 7 days in the presence of anti-CD3 antibody or TLR ligands (LPS, flagellin, MALP-2, and CpG). Cells were then analyzed by flow cytometry, gated for the expression of CD45 and CD3, and further evaluated for the expression of CD45RO (histograms). (a) Representative histograms of expression of CD45RO in T cells in the presence of medium alone (1.2%) and anti-CD3 antibody (15.6%) (first row), LPS (35.2%) and flagellin (39.1%) (second row), and MALP-2 (27.8%) and CpG (8.9%) (third row). (b) Summary data representing median and interquartile ranges of CD3+CD45RO+ cells in CB of term infants (n = 18) after in vitro incubation in medium or medium supplemented with anti-CD3 or selected TLR ligands. Statistical significance for difference between medium alone and TLR groups (*P < 0.05) was determined by Mann–Whitney rank test. CB, cord blood; CBMC, cord blood mononuclear cell; CpG, –C–phosphate–G– site; LPS, lipopolysaccharide; MALP-2, macrophage-activating lipopeptide-2; TLR, Toll-like receptor.

Full size image
Figure 4
figure 4

TLR ligand exposure does not increase the expression of CD45RO in naive T-cell–enriched adult PBMCs. (a) CD45RO-depleted PBMCs from healthy adults were cultured for 7 days in the presence of anti-CD3 or selected TLR ligands and then examined for expression of CD45RO (histograms). First row: medium alone (0.6%), anti-CD3 (14.1%), and LPS (0.1%). Second row: flagellin (0.7%), MALP-2 (0.5%), and CpG (0.2%). (b) Summary data representing three experiments. CpG, –C–phosphate–G– site; LPS, lipopolysaccharide; MALP-2, macrophage-activating lipopeptide-2; PBMC, peripheral blood mononuclear cell; TLR, Toll-like receptor.

Full size image

Distinct Homing Receptor Patterns Manifested by Preterm, Term, and Adult Naive (CD45RA+) T Cells

The traffic of T lymphocytes from blood to tissues occurs via a series of interactions between lymphocytes and endothelial cells (20). Naive T cells typically traffic through secondary lymphoid organs in search of peptides bound to major histocompatibility complex molecules expressed by antigen-presenting cells. Upon encounter with antigen and T-cell-receptor triggering, these naive T cells gain distinct homing characteristics (21). Following priming, clonal expansion, and differentiation, these matured T cells leaving the lymph node acquire new migration capabilities (22). T cells activated and matured in Peyer’s patches or mesenteric lymph nodes typically express C-C chemokine receptor 9 (CCR9) and the integrin α4β7, whereas activation in peripheral lymph nodes draining nongastrointestinal tissues, such as the skin and the lung, typically results in upregulation of CCR4 (23).

Cells were assessed for surface expression of CD3 and homing receptors α4β7 and CCR4 ( Figure 5a ). Among adult T cells (CD45+CD3+), 55% were α4+β7+, 27% were α4+β7−, 12% were α4−β7−, and only 4% were α4−β7+. In contrast, 92% of T cells from term infants were α4+β7+, 5% were α4+β7−, 3% were α4−β7−, and 0% were α4−β7+. Of note, the homing receptor pattern differed between preterm and term infant T cells ( Figure 5b ). Preterm infant T cells had a lower proportion of cells that were α4+β7+ (median values of 80% and 92% for preterm and term, respectively; P = 0.0034) and a higher proportion of cells that did not express the β7 homing receptor (median values of 16% and 8% in preterm and term infants, respectively; P = 0.0034). There was no difference in the expression of CCR4 on T cells between term and preterm infant cells.

Figure 5
figure 5

Differential homing receptor distribution in adult, term, and preterm T cells. (a) Adult peripheral blood (first row), term infant CB (second row), and preterm infant CB (third row). CD45+CD3+ cells were examined for expression of the homing receptors α4β7 and CCR4 (box-plots). (b) Stacked bars graph comparing the distribution of the α4β7 receptor among CD3+ T cells in adults (n = 10) and term (n = 22), and preterm infants (n = 26). White bars represent α4+β7+ T cells, black bars α4+β7− T cells, gray bars α4−β7− T cells, and dotted bars α4−β7+ T cells. Statistical significance (P < 0.0001) was determined by ANOVA. CB, cord blood; CCR, C-C chemokine receptor.

Full size image

In Vitro Exposure of Neonatal T Cells to Microbial TLR Ligands Alters the Expression of Homing Receptors

After incubation with selected TLR ligands, the percentage of CD3+α4+β7+ cells in term infant T cells decreased as compared with cells incubated with medium ( Figure 6a,b ); this decrease was primarily due to decreased expression of the β7 integrin (lipopolysaccharide 93% α4+β7−, flagellin 62% α4+β7−, macrophage-activating lipopeptide-2 80% α4+β7, –C–phosphate–G– site (CpG) 42% α4+β7−). We analyzed the expression of CCR4 in the CD3+α4+ subpopulation that did or did not express β7 ( Figure 7a,b ) and found that decreased expression of the β7 integrin was accompanied by increased expression of CCR4 after exposure to TLR ligands. This effect was especially striking in lipopolysaccharide-exposed cells ( Figure 7b ).

Figure 6
figure 6

TLR ligands change the proportion of neonatal T cells expressing the α4β7 homing receptor. (a) CBMCs from term infants were cultured for 7 days in the presence of medium alone, anti-CD3 antibody, and LPS (first row), and flagellin, MALP-2, and CpG (second row). Cells were analyzed by flow cytometry, gated for the expression of CD45 and CD3, and further evaluated for the expression of the α4β7 receptor (box-plots). (b) Column bar graph comparing cumulative data of the distribution of the α4β7 receptor in CD3+CD45RO+ cells in CB of term infants (n = 16) after exposure to TLR ligands. White bars represent percentage of cells positive for α4+β7+ and black bars represent percentage of cells positive for α4+β7−. Statistical significance (P < 0.0001) was determined by ANOVA. Median values are shown. CB, cord blood; CBMCs, cord blood mononuclear cells; CpG, –C–phosphate–G– site; LPS, lipopolysaccharide; MALP-2, macrophage-activating lipopeptide-2; TLR, Toll-like receptor.

Full size image
Figure 7
figure 7

Increased proportion of neonatal CD3+ T cells expressing CCR4 after exposure to TLR ligands. (a) CBMCs from term infants were cultured for 7 days in the presence of medium alone, anti-CD3 antibody, and LPS (first row), and flagellin, MALP-2, and CpG (second row). Cells were analyzed by flow cytometry; gated for the expression of CD45, CD3, α4; and further evaluated for the expression of β7 and CCR4 (box-plots). (b) Stacked bars graph comparing the distribution of the β7 subunit and CCR4 receptor among CD3+α4+ T cells in term infants (n = 16). White bars represent CCR4-negative cells and black bars represent CCR4-positive cells. Statistical significance (P < 0.0001) was determined by ANOVA. CBMC, cord blood mononuclear cell; CCR, C-C chemokine receptor; CpG, –C–phosphate–G– site; LPS, lipopolysaccharide; MALP-2, macrophage-activating lipopeptide-2.

Full size image

Chorioamnionitis Affects the Expression of CD45RO and the α4β7 Receptors in Neonatal T Cells

The proportion of T cells expressing CD45RO was compared between infants exposed to clinical chorioamnionitis (n = 5) and those without clinical chorioamnionitis (n = 41) ( Figure 8a ). The proportion of CD3+CD45RO+ was higher in the group with clinical chorioamnionitis (median = 25%) vs. no clinical chorioamnionitis (median = 6.9%). This difference was statistically significant (P = 0.008). There was no significant difference (P = 0.37) in the proportion of CD3+CD45RO+ T cells of preterm infants born to mothers without clinical chorioamnionitis (median = 23.1%), those born to mothers with clinical chorioamnionitis (median = 28.1%), and those born to mothers with histological chorioamnionitis (median = 23.2%) ( Figure 8b ).The proportion of CD3+α4+β7− was also significantly increased in infants exposed to clinical chorioamnionitis (median = 14.3%) and histological chorioamnionitis (median = 12.6%) as compared with those without exposure (median = 6.5%) (P ≤ 0.0001) ( Figure 9 ).

Figure 8
figure 8

Increased proportion of CD3+CD45RO+ T cells in infants exposed to chorioamnionitis. (a) Column bar graph comparing proportion of CD3+CD45RO+ T cells in infants exposed to clinical chorioamnionitis (n = 5) with those not exposed (n = 41). This figure includes both term and preterm infants. (b) Column bar graph comparing proportion of CD3+CD45RO+ T cells in preterm infants exposed to clinical chorioamnionitis (n = 6) or histological chorioamnionitis (n = 12) with those not exposed (n = 11). Statistical significance was determined by Mann–Whitney rank test (*P = 0.003).

Full size image
Figure 9
figure 9

Increased proportion of CD3+α4+β7− T cells in infants exposed to chorioamnionitis. Column bar graph comparing proportion of CD3+α4+β7− T cells in infants exposed to clinical chorioamnionitis (n = 5) or histological chorioamnionitis (n = 12) with those not exposed (n = 30). This figure includes both term and preterm infants. Statistical significance (*P < 0.0001) was determined by ANOVA.

Full size image

Discussion

Although the neonatal immune system lacks memory of prior antigen exposure, it provides defense, albeit limited, against microbial pathogens while it is maturing and developing responses to newly encountered antigens. In this study, we report a novel characteristic of neonatal T cells that distinguishes them from adult T cells. We previously demonstrated that preterm labor is associated with neonatal T-cell activation, as evidenced by an increased proportion of cells expressing the activation markers CD25, CD69, and HLA-DR (13). In this study, we also report an increased proportion of T lymphocytes expressing the maturation antigen CD45RO in preterm CB and a distinct homing receptor pattern that distinguishes preterm infant T cells from term infant CB and adult T cells. Finally and most strikingly, we report that, in response to TLR ligand exposure, neonatal T cells have the capacity to acquire a maturation phenotype and that adult naive T cells lack this capacity. These observations provide novel insights into immune responsiveness of the neonate and the potential role of innate responses in the maturation of the adaptive immune system.

In adults, the principal stimulus for naive T-cell activation and maturation is considered to be T-cell-receptor triggering by neoantigenic peptides. In this study, however, we demonstrate that neonatal naive T cells can upregulate expression of the maturation antigen CD45RO through TLR-dependent mechanisms. Using selected TLR ligands, we found consistent upregulation of CD45RO by neonatal T cells. Of note, this response is lost in adult naive T cells. In the murine model, neonatal and adult naive T cells exhibit phenotypical and functional differences whereby neonatal naive T cells exhibit a higher level of effector Th1/Th2 cytokine production and an early proliferation response to stimulation with IL-7 (24). Our study indicates there is an additional difference between neonatal and adult cells. We speculate that the apparent maturation phenotype of neonatal T cells that we found in preterm CB may be driven by in vivo exposure to bacterial products. In utero exposure to plasmodial products that cross the placenta have been linked to altered neonatal immune responses to immunizations (25,26). We found a significant difference in the pattern of maturation and homing receptors in infants exposed to chorioamnionitis. This in utero maturation process could be related to placental inflammation. Of note, the analysis restricted to only preterm infants showed no differences between infants exposed to chorioamnionitis (either clinical or histological) and those not exposed to chorioamnionitis. These results should be interpreted very carefully because of the relatively small preterm group. Other factors should be taken into consideration that may increase the passage of bacterial products to the fetus, such as the duration of the rupture of membranes, presence of bacteria in maternal urine and/or blood, maternal periodontal disease, and presence of bacteria that are difficult to grow in cultures and whose presence could be detected only by PCR. Future studies should include measurement of microbial products in CB samples. We are currently investigating whether the maturation effect is mediated directly through TLR receptors in T cells or if antigen-presenting cells are needed.

To our knowledge, this is the first demonstration of a distinct homing receptor pattern in preterm infant CB T cells that differs from those of term infant and adult T cells. Lymphocyte migration is highly regulated, which ensures that distinct cell populations are delivered to intended site(s) in the proper physiological or pathological context (27). Homing receptors that are important for T-cell homeostasis include α4β7 and CCR4. The homing molecule α4β7 promotes T-cell entry into intestinal sites and CCR4 supports T-cell chemotaxis and entry into nongastrointestinal sites through interaction with its ligand induced in the setting of inflammation (28). The high proportion of α4β7 cells in CB may reflect the critical role of intestinal homing—and perhaps the importance of exposure to microbial elements in the gut—in neonatal T-cell development (29). In this study, we showed that in vitro exposure of CB cells to microbial TLR ligands alters the expression of these homing receptors, decreasing expression of α4β7 and increasing the expression CCR4. Our finding of fewer α4+β7+ T cells in preterm CB may reflect an earlier (premature) maturation of these cells in response to premature exposure to microbial elements. A consequence of this exposure may be increased migration of these cells to nongastrointestinal sites (e.g., lungs, skin, or brain) that may underlie the systemic inflammatory organ dysfunction often seen in premature infants. Studies are under way to better characterize the effector function and cytokine potential of these cells.

What are the possible clinical implications of T-cell activation/maturation and alteration in homing receptor patterns in the preterm infant? Premature infant morbidities (e.g., bronchopulmonary dysplasia, necrotizing enterocolitis, respiratory distress syndrome, periventricular leukomalacia, and intraventricular hemorrhage) appear to be related to a state of immune activation and inflammation (6,7,30,31,32). Evidence for this relationship includes direct associations between placental pathology and these morbidities in preterm infants (33) and an increased proportion of T lymphocytes migrating to lungs and brain in addition to the presence of CD45RO+ T cells (6,7,32). Here, we demonstrate alterations in T-cell maturation and homing receptor expression in premature infants that may contribute to these morbid events.

Most striking in this study was our finding of a novel and qualitative difference in the way neonatal and adult naive T cells respond to TLR ligand exposure. The apparent maturation response of neonatal T cells to these microbial products suggests either that microbial elements are necessary for physiologic maturation of the adaptive immune defenses (34) or that neonates can respond to microbial challenges using mechanisms that are not used by adults with more mature adaptive immune systems. These differences between neonatal and adult cells may relate to several factors including the developmental age of the cells, a history of previous exposure to antigen, differences in the nature of TLR interactions with their ligands, or function of antigen-presenting cells. These possibilities will be important to explore in further investigations.

Methods

Subjects

CB samples were obtained from subjects delivered at Case Medical Center/MacDonald’s Women’s Hospital and Tampa General Hospital after obtaining informed consent from parent(s). Healthy adult donor samples were obtained from the Center for AIDS Research at Case Western Reserve University after obtaining informed consent. These studies were performed in accordance with the policies of and approved by the institutional review boards at Case Western Reserve University/University Hospitals of Cleveland and the University of South Florida/Tampa General Hospital. CB samples were obtained from healthy term infants (gestational age ≥37 weeks, n = 26) and preterm infants (gestational age ≤36 weeks, n = 22); and peripheral blood was obtained from healthy adults (n = 10). Demographic and clinical details for the infants were obtained from the medical record ( Table 1 ).

Table 1 Subject characteristics
Full size table

Cell Preparation and Flow Cytometry

CB from term and preterm pregnancies was collected according to the National Cord Blood Program protocol (35). Peripheral blood was collected from adults by venipuncture and transferred into heparin-coated tubes. Cell-surface-molecule expression was assessed in whole blood within 6 h. This rapid analysis permits accurate assessment of the in vivo phenotype state.

Whole blood (200 µl) samples were stained for 15 min in the dark at room temperature with fluorochrome-labeled antibodies ( Table 2 ) and then incubated for 15 min with FACS Lysing Solution (BD Bioscience, San Diego, CA). Cells were washed in wash buffer (phosphate-buffered saline with 1% bovine serum albumin and 0.1% sodium azide). Cells were then placed in wash buffer and fixed in 1% formaldehyde. Cells were analyzed using an LSRII Flow Cytometer (Becton Dickinson, San Jose, CA); FACSDIVA software (version 6.1.1 Bioscience, San Diego, CA) was used to organize the data, and FlowJo software (version 7, Tree Star, Ashland, OR) and Prism 5.0 Graphpad software (La Jolla, CA) were used to analyze the data. Nucleated red blood cells were excluded from the analysis by gating on cells expressing CD45 (36) ( Figure 1a ). T cells were identified by size, granularity, and expression of CD3 and CD45 ( Figure 1b ). An average of 6,000 CD3+ events were collected for analysis.

Table 2 Cell-surface markers, fluorochrome labeling, and function
Full size table

T-Cell Exposure to TLR Ligands

CB mononuclear cells from term infants were isolated over a Ficoll-Hypaque cushion. Density sedimentation was performed twice to improve the purity of the mononuclear cells (Supplementary Figure S1 online) (37). Cell viability was examined using trypan blue. Cell viability after Ficoll-Hypaque was always higher than 95%. Cells were cultured in RPMI 1640 (Cambrex BioWhitaker; Walkersville, MD) supplemented with 10% fetal bovine serum (Hyclone; Logan, Utah), 1% 2 mmol/l L-glutamine (Cambrex BioWhitaker), and 1% streptomycin (Cambrex BioWhitaker). Cells were cultured for 7 days with anti-CD3 monoclonal antibody alone (BD Pharmigen, San Diego, CA; 100 ng/ml) or individual TLR, as indicated. Lipopolysaccharide (E. coli, lipopolysaccharide 5 µg/ml, TLR 4 ligand), flagellin (S. tyhimurium, 100 ng/ml, TLR 5 ligand), macrophage-activating lipopeptide-2 (MALP-2 isolated from Mycoplasma fermentans, 100 ng/ml, TLR 2–6 dimer), and CpG oligodeoxynucleotide 2006 (5 µg/ml, TLR 9 ligand) were purchased from Imgenex (San Diego, CA). After 7 days, cells were stained with fluorochrome-labeled antibodies to characterize T-cell populations (naive CD45RA+, memory CD45RO+, α4β7+ T cells, and CCR4+ T cells) ( Table 2 ) and analyzed by flow cytometry, as described above. In addition, cells were stained with propidium iodide for viability at day 4 and day 6 (Supplementary Figure S2 online). Viability remained higher than 92%. Adult peripheral blood mononuclear cells were depleted of CD45RO+ cells by magnetic bead separation and examined for the presence of antigen-presenting cells (Supplementary Figure S3 online). Peripheral blood mononuclear cells were suspended in 80 µl of magnetic-activated cell sorting depletion buffer (phosphate-buffered saline, 0.5% bovine serum albumin, and 2 mmol/l EDTA) per 107 total cells. Then 20 µl of CD45RO microbeads (Miltenyi Biotec no. 130-046-001; Miltenyi Biotec, Auburn, CA) were added per 107 total cells. Cells were mixed and refrigerated for 15 min (4–8 °C). Cells were washed by adding 1–2 ml of depletion buffer per 107 cells and centrifuged for 10 min. Cells were then resuspended in 500 µl of buffer and run in the AutoMACS Magnetic Separator (Miltenyi Biotec) using the “Depletes” separation program. Depletion of CD45RO+ cells was confirmed by flow cytometry and the resulting T-cell populations were determined to be >95% CD45RA+CD45RO−. Neonatal T-cell populations were examined for the presence of CD45RO+ T cells before culture. CD45RO+ T cells in CB samples were <5%.

Placental Pathology

A histological placental diagnosis was also used because chorioamnionitis could still be present even in the absence of clinical symptoms. This diagnosis is routinely performed in placentas of preterm infants born at Tampa General Hospital and MacDonald’s Women’s Hospital. Results are summarized in Table 1 .

Statistical Methods

Group medians were compared for statistically significant differences using a Mann–Whitney rank test or ANOVA; a P value of 0.05 was considered statistically significant. We analyzed the samples and generated graphs using GraphPad Prism 5 (La Jolla, CA).

Statement of Financial Support

This work has been supported by grants from the Rainbow Babies and Children’s Foundation Fellowship Research Award, the NICHD Research Career Development in Child Health (K12 HD057581), the Case Western Reserve University Center for AIDS Research (AI-36219), the National Institutes of Health (AI 076174), and the Tampa General Hospital Office of Clinical Research Award.