Main

The CD45 (leukocyte common antigen) is a family of membrane glycoproteins that are densely expressed by cells of hematopoietic origin(1, 2). At least five human isoforms, ranging in molecular mass from 180 to 220 kD, have been identified(3–11). These isoforms are produced by alternative mRNA splicing of three exons encoded by a single leukocyte common antigen gene(12). Each isoform shares a common intracellular domain having protein tyrosine phosphatase activity, which is involved in T cell activation(13). However, differences in the amino-terminal region of their extracellular domains create unique antigenic determinants. For example, the CD45RA antigen, encoded by exon A, is present in the two largest isoforms with a relative molecular mass of 220 and 205 kD. In contrast, the CD45RO antigen, which is 180 kD in size, lacks sequences encoded by exons A, B, and C(7, 9, 10). Although individual lymphocytes simultaneously express more than one CD45 isoform, the CD45RA and CD45RO antigens are largely reciprocally expressed by subsets of T cells with distinct functions and activation requirements(4, 13–16).

When stimulated with mitogens such as phytohemagglutinin or concanavalin A in vitro, CD45RA+ T cells lose CD45RA expression and gain CD45RO expression(5, 8, 17, 18). In vivo, the vast majority (>90%) of T lymphocytes present in umbilical cord blood express CD45RA antigen, but not CD45RO antigen. During childhood, the relative frequency of circulating CD45RA+ cells gradually declines, and the frequency of CD45RO+ cells increases, reaching 40-60% in adults(19, 20). Based on these in vitro and in vivo data, it has been proposed that CD45RA expression is characteristic of unprimed or naive T lymphocytes, whereas the CD45RO expression is characteristic of primed or memory T lymphocytes(1–22). However, several works(8, 17, 21, 23) have demonstrated that CD45 isoform expression does not directly correlate with maturational status of T lymphocytes, both in vitro and in vivo, and that individual T lymphocytes alter their isoform expression in a highly regulated manner during thymic selection and upon antigen exposure in the periphery.

Regardless, it has been clinically shown that the proportion of CD45RO+ T lymphocytes markedly increases in association with several specific viral infections, autoimmune diseases, or granulomatosis processes(24–30). Furthermore, it was recently suggested that up-regulation of CD45RO expression on umbilical cord lymphocytes might be a good indicator of congenital viral infections(31).

Infections are a significant cause of mortality and long-term mobility of newborns, which occur during intrauterine development, in association with the birth process, or in the first postnatal days or weeks. The prognosis of these diseases can be improved if recognized early and if appropriate therapy is promptly instituted, particularly in the case of bacterial infections. However, successful therapy, at least in part, must depend upon the physicians' clinical experience, because the early diagnosis of neonatal infections is often difficult(32–37). Thus, a novel marker for the early diagnosis of neonatal infections is needed.

In the present study, we examined the CD45RA and CD45RO expression on total, CD45+, and CD8+ lymphocytes from newborns, and found a significant increase in CD45RO expression, particularly on CD4+ T lymphocytes, in neonatal infections. This change was transient and returned to normal levels after recovery from the disease. Clinical relevance of transient changes in CD45 isoform expression to the early diagnosis of neonatal infection is discussed.

METHODS

Patients and samples. Studies were made of 55 newborns who were hospitalized in Kiyose Metropolitan Children's Hospital or Yamanashi Medical University Hospital. None of the newborns showed severe fetal distress or neonatal asphyxia. These newborns were classified into three groups according to the possibility of bacterial or viral infection, based on the following criteria: I, “no” infection group: showed neither clinical symptoms suspicious of infection such as fever, hypothermia, dyspnea, tachypnea, cyanosis, vomiting, abdominal distention, poor sucking, and petechiae, nor laboratory data suspicious of infection such as CRP ≥ 0.3 and/or IgM ≥ 20 mg/dL; II, “possible” infection group: showed symptoms and/or laboratory and radiographic findings suspicious of infection, but no causative agent was isolated from blood, CSF, urine, intratracheal or gastric aspirate, or stool; and III, “definite” infection group: showed symptoms and/or laboratory and radiographic findings suspicious of infection, and a causative agent was isolated. The case in which CSF showed pleocytosis but no pathogen was isolated was also included in group III(aseptic meningitis). The bacteria isolated were considered as the causative pathogen when colonies were formed of more than 105/mL in urine, and when a single agent was exclusively isolated from intratracheal or gastric aspirate, or stool. A total 63 peripheral blood samples from 55 newborns were collected according to the informed consent guidelines of the medical ethics committee of both hospitals. First samples were obtained 7 d or less after birth. In some patients, samples were obtained at various intervals.

MAbs. PE-conjugated MAb reactive with CD3 (T3), CD4 (T4), CD8(T8), and FITC-conjugated mAbs reactive with CD45RA (2H4) and CD45RO (UCHL-1) were used. All mAbs were purchased from Coulter Immunology (Hialah, FL), except for UCHL-1, which was from DAKO A/S (Glostrup, Denmark).

Phenotypic analysis. PBMC were separated by Ficoll-Hypaque density gradient sedimentation from 1 mL of heparinized blood, obtained by routine venipuncture, and were incubated with FITC- and PE-conjugated MAb for 30 min on ice. After washing with PBS, reactivity of MAb was analyzed on the gated lymphocyte population using a flow cytometer (EPICS Profile Analyzer, Coulter Electronics) as shown in Figure 1A. The method of two-color analysis (PE on the vertical axis versus FITC on the horizontal axis) have been described previously(21).

Figure 1
figure 1

Flow cytometric analysis. (A) The forward× side plot on the gated lymphocyte population. PBMC obtained from peripheral blood by Ficoll-Hypaque density gradient sedimentation were applied to a flow cytometer. Reactivity of mAbs was analyzed on the gated lymphocyte population as shown in the forward scatter (FS) × the lateral side scatter (LSS) plot. (B) Two-color fluorescence histogram of CD4 and CD45RO. PBMC of case 20 from group I (left) and case 52 from group III (right) were stained with anti-CD4-PE and anti-CD45RO-FITC, and analyzed on a flow cytometer. (CD4 expression on the vertical axis vs CD45RO expression on the horizontal axis.)

Full size image

Western blot analysis. The procedure has been described previously(38, 39). Briefly, PBMC obtained from several newborns were lysed in buffer (50 mM Tris-HCl, pH 8.0, containing 1% Nonidet P-40, 150 mM NaCl, 0.05% NaN3, 5 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 0.2 trypsin-inhibiting unit/mL aprotinin, 1μg/mL pepstatin A, and 10 mM iodoacetamide) for 30 min on ice. The lysates were run on a 5.5% SDS-polyacrylamide gel under reducing conditions and transferred to nitrocellulose membranes. Membranes were incubated overnight at room temperature with anti-CD45RO MAb (UCHL-1, 1 μg/mL), and then incubated with alkaline phosphatase-labeled rabbit anti-mouse IgG (Promega, Madison, WI) for 1 h. The bands were developed with nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate (Promega).

Statistical analysis. Comparison between two of three groups was performed using a χ2 test or ANOVA in Stat-View programs(Abacus Concepts, Inc., Berkeley, CA). A value (Fisher's exact test or Scheffe's F) of p < 0.05 was considered as significant.

RESULTS

Clinical and laboratory features. Twenty-four newborns (male 13, female 11) were classified into group I (no infection; cases 1-24), 17 newborns (male 9, female 8) into group II (possible infection; cases 25-41), and 14 newborns (male 5, female 9) into group III (definite infection; cases 42-55) according to the criteria described in “Methods.” In group III, 12 newborns were diagnosed as having bacterial infection (sepsis 4, colitis 3, meningitis 2, pneumonia 2, urinary tract infection 1) and two as having viral infection (Coxsackie virus B3 1, unknown 1). Etiologic pathogens in bacterial infection were as follows: Staphylococcus aureus 4,γ-streptococcus 2, Escherichia coli 2, Pseudomonas aeruginosa 1, Staphylococcus epidermidis 1, Staphylococcus faecalis 1, and Listeria monocytogenes 1. As summarized in Table 1, there was no significant difference between two of three groups in sex, gestational age, body weight, APGAR score, incidence of small for dates, and sampling day. However, incidence of premature rupture of the membranes was significantly higher in group II than in group I(p = 0.04). Regarding hematologic data, Hb and platelets were significantly decreased in group III (Hb, I versus III: p= 0.04; platelets, I versus III: p < 0.01), whereas there was no significant difference in white blood cells, neutrophil (%), lymphocytes (%), IgG, IgA, and IgM between two of three groups. CRP was significantly increased in group III (versus I or II, p< 0.01). These results suggest that low Hb and platelet in addition to high CRP might be useful for laboratory evaluation of neonatal infection. However, it should be noted that decrease in Hb was very small in almost all cases, and an apparent decrease in platelet count (<100 × 109/L) was observed only in half of the cases in group III, and that elevated CRP levels were not always observed in group III (cases 44 and 49).

Table 1 Summary of clinical and laboratory data
Full size table

The expression of surface antigens. First, CD3, CD4, and CD8 expression were compared between two of three groups as shown in Figure 2. The CD3 expression (I: 74.6 ± 13.3%, II: 69.1 ± 14.1%, III: 74.5 ± 13.9%), the CD4 expression (I: 62.1± 15.7%, II: 56.6 ± 22.8%, III: 65.2 ± 15.2%), and the CD8 expression (I: 20.9 ± 8.4%, II: 24.1 ± 12.7%, III: 27.2± 9.2%) showed no significant difference. There was also no significant difference in the CD4/CD8 ratio between two of three groups (I: 4.0 ± 2.9%, II: 3.5 ± 3.5%, III: 2.8 ± 1.3%).

Figure 2
figure 2

The expression of CD3, CD4, and CD8 on lymphocytes. The CD3, CD4, and CD8 expression on lymphocytes in groups I, II, and III was analyzed 7 d or less after birth. Each bar represents the mean. There was no significant difference in the CD3, CD4, and CD8 expression between two of three groups.

Full size image

Next, the CD45RA expression on total, CD4+, and CD8+ lymphocytes was compared. As shown in Figure 3, the CD45RA expression on total lymphocytes in group III (65 ± 8.8%) was significantly lower than that in group I (80.6 ± 8.6%) or group II(75.0 ± 13.9%). However, the CD45RA expression on CD4+ or CD8+ lymphocytes showed no statistical difference between two of three groups (CD4, I: 50.4 ± 16.5%; II: 42.7 ± 20.3%; III: 41.1± 16.5%; CD8, I: 17.4 ± 6.9%; II: 18.1 ± 8.8%; III: 18.5± 6.2%). These results suggest that low CD45RA expression on total lymphocytes in group III may reflect a decrease in CD45RA expression on CD4-CD8- T lymphocytes, natural killer cells, or B cells.

Figure 3
figure 3

The expression of CD45RA on total, CD4+, and CD8+ lymphocytes. The CD45RA expression in groups I, II, and III was analyzed 7 d or less after birth. Each bar represents the mean. The CD45RA expression on total lymphocytes was significantly decreased in group III(vs group I).

Full size image

Third, the CD45RO expression on total, CD4+, and CD8+ lymphocytes was evaluated. As shown in Figure 4, the CD45RO expression on total lymphocytes in group III (27.1 ± 14.0%) was significantly higher than that in group I (7.1 ± 2.7%) or group II(13.5 ± 8.9%). The CD45RO expression on CD4+ or CD8+ lymphocytes was also significantly highest in group III (CD4, I: 4.5 ± 2.2%, II: 9.1 ± 6.5%, III: 18.6 ± 9.6%; CD8, I: 1.1 ± 0.7%, II: 3.1 ± 4.4%, III: 7.9 ± 10.1%). It should be noted that the CD45RO expression on CD4+ lymphocytes was always greater than 10% in group III, and showed no overlap with the respective value in group I. Representative two-color analysis on CD45RO versus CD4 expression in groups I and III is shown in Figure 1B. Of clinical importance, among newborns belonging to group II, the CD45RO expression on CD4+ T lymphocytes from six cases was markedly increased and was almost similar to that of the mean in group III as shown in Figure 4, suggesting “definite” infection in these cases.

Figure 4
figure 4

The expression of CD45RO on total, CD4+, and CD8+ lymphocytes. The CD45RO expression in groups I, II, and III was analyzed 7 d or less after birth. Each bar represents the mean. The CD45RO expression on total and CD4+ lymphocytes was significantly increased in group III (vs group I or II).

Full size image

Finally, there was no difference in each antigen expression between small for date and appropriate for date newborns, which was statistically confirmed in newborns belonging to group I (data not shown). No significant changes were also demonstrated in group I when the CD45 isoform expression was compared between newborns of early sampling <3 d and those of later sampling ≥3 d(data not shown).

The temporal changes in CD45RA and CD45RO expression on total lymphocytes along the clinical course was evaluated in three and five newborns from groups II and III, respectively. Second samples were obtained 14 d or more after birth, at a time they were recovered and showed few clinical symptoms. As shown in Figure 5, the decreased CD45RA expression seen 7 d or less after birth in the three newborns belonging to group III (cases 43, 50, and 51) increased back toward normal levels after recovery from infection. Conversely, the increased CD45RO expression in five newborns from group III returned toward normal levels 14 d or more after birth. There were no significant changes in either CD45RA or CD45RO expression in three newborns from group II over the same time period (although a small increase in CD45RA expression was observed in case 26). These results suggested that a decrease in CD45RA expression and an increase in CD45RO expression in group III are transient.

Figure 5
figure 5

Temporal changes in CD45RA and CD45RO expression on total lymphocytes. CD45RA and CD45RO expression on total lymphocytes 7 d or less and 14 d or more after birth was compared in three newborns from groups II (open symbols) and five newborns from group III (filled symbols). Thre were no significant changes in the CD45RA and CD45RO expression over time in group II except for a small increase in CD45RA expression in case 26.

Full size image

To further evaluate changes in CD45RO expression, lysates from PBMC obtained from several newborns were run on a 5.5% SDS-polyacrylamide gel and blotted to nitrocellulose. CD45RO expression was analyzed using anti-CD45RO MAb as described in “Methods.” As shown in Figure 6, the CD45RO antigen (180 kD) was not visualized in the lysates from newborns belonging to group I (cases 19, 22, and 24), whereas the 180-kD band was clearly detected in the lysates from those belonging to group III (cases 50 and 51) on d 2 and 6 after birth, but almost disappeared in lysates on d 18 and 14, respectively. It appears that this Western blot analysis can detect the population expressing CD45RO only at high levels. Similar experiments examining CD45RA expression could not be performed due to the inability of anti-CD45RA MAb to recognize its antigen on Western blots.

Figure 6
figure 6

Analysis of CD45RO expression on Western blot. The CD45RO expression in cell lysates from PBMC of three neonates from group I(lanes 1-3) and two newborns from group III (lanes 4-7) was analyzed as described in “Methods.” Samples were obtained 7 d or less (lanes 1-4 and 6) and 14 d or more (lanes 5 and 7) after birth.

Full size image

As a representative example, the clinical course and changes in surface antigen expression of a female newborn (case 43) from group III is shown in Figure 7. She developed frequent apnea within a few hours of birth. Laboratory data revealed high white blood cells (32.7 × 109/L) and CRP (2.5 mg/dL), but no elevation of IgM. The leukocyte count in CSF was 8.0 × 109/L, and Listeria monocytogenes was subsequently demonstrated in both blood and CSF. Intravenous administration of ampicillin, gentamicin, and vancomycin was started, and she recovered completely at 1 mo from the onset of symptoms. Serial examination of surface antigens expression revealed that the CD45RA expression on total lymphocytes was only 24.0% on d 2, but gradually increased throughout the clinical course, reaching 82.5% on d 50. In contrast, CD45RO expression was markedly up-regulated (48.6%) on d 2, but decreased to 12.0% on d 50.

Figure 7
figure 7

Representative clinical course and changes in surface antigens in a neonate (case 43) from group III. Laboratory examination, CSF analysis, and analysis of surface antigen expression on lymphocytes were performed serially. Abbreviations: VCM, vancomycin; GM, gentamicin; AB-PC, ampicillin; WBC, white blood cell.

Full size image

DISCUSSION

Diagnosis of neonatal infection is generally known to be difficult because newborns suffering from infections show few specific symptoms, and because availability of samples for laboratory diagnosis is limited. Moreover, although peripheral blood cell counts, IgM, CRP, granulocyte/colony-stimulating factor, and so forth have been used as tools for laboratory evaluation of neonatal infection, these are not always reliable and satisfactory as reported previously(32–37). Thus, the transient changes in CD45 isoforms, in particular the increased CD45RO expression on CD4+ T lymphocytes presented in this study, might shed new insights into evaluation of neonatal infection and possibly provide a useful tool for the early diagnosis of neonatal infection.

Recently, there have been a number of reports describing changes in the CD45 isoform expression in association with specific viral infections. It was reported that the percentage of CD45RA+ T lymphocytes decreased and the percentage of CD45RO+ T lymphocytes, particularly CD4+ lymphocytes, increased, when cultured in vitro with human cytomegalovirus(40). Clinically, the CD8+ CD45RO+ population was reported to markedly expand in peripheral blood from patients with infection due to EBV, HIV-1, or hepatitis B virus(24, 26, 29). Thus, it is now evident that the CD45RO expression on T lymphocytes is up-regulated after particular viral infections. We showed that neonatal CD45RO+ T lymphocytes increased by not only viral but also bacterial infections. Concomitantly, CD45RA expression on neonatal lymphocytes was significantly decreased. It is thought that, in adults, a small increase in CD45RO expression after bacterial infection is difficult to detect because the CD45RO antigen is highly expressed on T lymphocytes before the present infection. On the other hand, a small increase in CD45RO expression is expected to more easily be detected in neonatal T lymphocytes because of its low basal levels of expression. Similarly, because the vast majority of T lymphocytes from newborns express the CD45RA antigen, a small decrease in CD45RA expression might be detected after infection. Of clinical importance, there was no overlap of CD45RO on CD4+ T lymphocytes expression between groups I and III. Further examination in additional newborns will be required to confirm its clinical utility.

Our data demonstrated that neonatal T lymphocytes underwent phenotypic conversion from CD45RA+ to CD45RO+ after infectious“stimulation.” Of interest, changes in the CD45RA and CD45RO expression were transient and reversible in infected newborns. Although its precise process is unclear at present, both or either mechanisms,“clonal reconstitution” and “phenotypic reconversion,” can be postulated because it is reported that the life span of CD45RO+ lymphocytes is shorter than that of CD45RA+ lymphocytes, presumably due to the apoptotic process(41, 42). Furthermore, the population originally CD45RA+ can reconvert from the CD45RO phenotype to the CD45RA phenotype after the stage of stimulation(8, 21, 43).

The mechanisms by which the T cell receptor triggering affects the negative regulatory splicing factors that regulate CD45 alternative splicing is unknown(16). It is also unclear what influences the alternative splicing pattern in T lymphocytes during the process of infection. In addition to T cell receptor triggering, changes in levels and repertoire of cytokines and intercellular ligand counter-ligand interactions induced by infection may be involved in this process. Newborn PBMC are known to produce equivalent amounts of IL-2, but less interferon-γ and IL-4, than adult PBMC in response to polyclonal activators, such as staphylococcal enterotoxin B(44, 45). It is also reported that exposure of purified CD45RA+ T lymphocytes to IL-2 promotes their phenotypic conversion to CD45RO+(46), suggesting a crucial role of IL-2 in infection-associated phenotypic conversion of neonatal T lymphocytes. Thus, it is conceivable that the transient expression of particular CD45 isoforms plays an important role in development of the functional repertoire required for appropriate regulatory or effector functions of the immune responses after infections.