Multiple leucocyte activation markers to detect neonatal infection

Abstract

Diagnosis of congenital or neonatal infection is often based on clinical signs. However, clinical symptoms of infections may not be specific, and for this reason early diagnosis is often determined on results of laboratory tests, which may not currently be adequate. A more reliable method of detection of infection may be the demonstration of activated lymphocytes, which can be conducted rapidly and before the isolation of the infected organism. We have shown that detection of up-regulation of CD45RO, an activated/memory isoform of CD45 present on T cells, provides a reasonably sensitive screening test for neonatal infection. We also showed that dual expression of CD45RA/CD45RO was up-regulated early during the infective process in neonates with documented infection. However, other leucocytes are also activated during the infective process. To improve the sensitivity of the neonatal infection screening test and to identify the types of leucocytes involved in the immune response to the infective organism, we studied further the up-regulation of a comprehensive range of surface activation markers on T cells, monocytes and natural killer (NK) cells from a group of 17 newborn patients with positive culture, a group of 40 possibly infected patients based on clinical signs and a control group. ‘Normal’ ranges were established for each activation marker for each leucocyte subset from 1 to 7 and 7-14-day-old newborns <35 weeks’ gestation and 35-40 weeks’ gestation. There was a significant increase in the percentage of T cells expressing CD25 in the peripheral blood from infants at 2 weeks of age. Expression of HLA-DR on T cells, CD25 and CD69 on monocytes and HLA-DR on NK cells was also increased significantly in the peripheral blood from infants at 2 weeks of age and may reflect a maturation of these functional surface molecules. Up-regulation of CD69 on NK cells was the most sensitive marker for neonatal sepsis (positive in 13/16 patients). CD69 and CD25 expression was increased significantly on T cells in 11/17 and 10/17 patients, respectively. A combination of CD45RA/CD45RO and CD45RO identified 11/16 infected patients. Measurement of CD69 expression on NK cells with CD45RA, CD45RO, CD25 and CD69 expression on T cells resulted in a significant increase in at least two leucocyte activation markers from infected patients. In conclusion, this is the first report of the up-regulation of CD69 on NK cells as a sensitive marker of neonatal infection. A combination of this marker with CD45RA, CD45RO, CD25 and CD69 expression on peripheral blood derived T cells is the most sensitive and specific for neonatal infection.

INTRODUCTION

Infection is still an important cause of neonatal morbidity and mortality, despite development of broad spectrum antibiotics and technical advances in life support therapy. Diagnosis of congenital or neonatal infection is often based on clinical signs. However, clinical symptoms of infection may not be specific, and for this reason early diagnosis is often based on the results of laboratory tests, which currently may not be adequate. Serological tests and isolation of microorganisms do not give immediate results and haematological tests that are used currently to screen for the presence of infection in the newborn can be difficult to interpret. A leucocytosis is difficult to interpret due to the large ‘normal’ range in the newborn (8-32.5 × 10³/µl) and increase in neutrophil band forms is subjective, with different morphologists giving variable results. The expression of surface functional antigens on neutrophils, and their up-regulation in infection is difficult to interpret due to the effects on these surface antigens of variables such as anticoagulant, type of fixative used, cell separation procedures. In addition, such tests require immediate processing of specimens and time-consuming processing techniques [1].

A more reliable method of detection of infection may be the demonstration of activated lymphocytes, which can be conducted rapidly and before the isolation of the infective organism. We have shown that detection of up-regulation of CD45RO, an activated/memory isoform of CD45 present on T cells, is a sensitive and reliable marker of neonatal infection [2]. We also showed that dual expression of CD45RA/CD45RO was up-regulated early during the infective process in neonates with documented sepsis. Furthermore, we found that other confounding influences on lymphocyte activation; type of delivery, respiratory distress or isoimmunization were not associated with up-regulation of lymphocyte CD45RO expression. However, other leucocytes are also activated during the infective process. Monocytes process and present foreign antigen to naïve T cells in the newborn patient and natural killer cells (NK) cells become lymphokine-activated killer cells in the presence of IL-2 and respond directly to virally infected and malignant cells [3]. Therefore, to improve the sensitivity of the neonatal sepsis screening test and to identify the types of leucocytes involved in the immune response to the infective organism, we studied the up-regulation of a comprehensive range of surface activation markers on peripheral blood derived T cells, monocytes and NK cells. We included CD69 in the investigative panel, as we have found previously that CD69 is up-regulated within 2 h of stimulation by mitogen [4] and hence detection of this molecule on the surface of these leucocyte subsets may indicate a very recent infection. It is inducibly expressed in a wide variety of haemopoietic cells and contributes to signal transduction, Ca⁺⁺ influx, cytokine production and cytokine receptor synthesis [5]. Other early markers of leucocyte activation included in the panel were CD25 (IL-2Rα) and CD71 on T cells and monocytes and tissue factor (TF) on monocytes. Late markers of activation were CD122 (IL2Rβ) on T cells and HLA-DR on T cells and NK cells, were also determined using multiparameter flow cytometry as reported previously [4].

PATIENTS AND METHODS

Reference ranges

Following institutional ethics approval and parental consent, blood was collected into EDTA and tested within 6 h, or heparin and tested within 24 h. Full blood counts were determined using a Technicon H1E (Bayer Diagnostics), and blood films examined following May-Grunwald-Giemsa staining. Any sample showing features suggestive of infection was excluded from the ‘normal’ group. Reference ranges were established for each activation marker for each different leucocyte subset from analysis of leucocytes from newborn infants at 0-7 days (n = 55) and 7-14 days (n = 25) of age. Infants aged 0-7 days were categorized further as less than 35 weeks’ gestation (n = 25) and greater than 35 weeks’ gestation (n = 30).

Control samples

These were prepared simultaneously with the patient samples and were included with every batch of tests. Controls included at least one cord blood from an infant in which any possibility of infection had been excluded. ‘Septic screening’ is routinely carried out at birth in most infants admitted to the Neonatal Intensive Care Unit at the Women's and Children's Hospital. This group was selected from such infants in whom no infection was proven subsequently by blood culture or viral serology (to investigate test specificity). One adult ‘normal’ blood was also included with every batch of infant blood for phenotyping as an internal control sample, as we have established previously adult normal ranges for these markers (unpublished results).

Infected and possible infected infants

All infants, born or admitted to the Women's and Children's Hospital intensive care unit or nurseries with suspected infection within a 12-month period, were eligible for inclusion in the study. Newborn infants were selected for inclusion in the ‘proven infected’ group on the basis of the following selection criteria: (i) organism cultured from blood culture or cerebrospinal fluid; (ii) viral serology indicative of infection; (iii) urine bacterial culture or streptococcal antigen-positive; (iv) positive culture from endotracheal tube with radiological evidence of parenchymal lung involvement; and (v) clinical and radiological evidence of necrotizing enterocolitis with supportive evidence of infection. There were 17 infants with proven culture positive bacterial infection in this group. Forty patients were included in the ‘possible infected’ group based on clinical suspicion of infection. These patients did not qualify for any of the selection criteria for infected infants.

Fluorescent staining for lymphocyte surface markers

Blood was incubated for 15 min with optimal dilutions of monoclonal antibodies, as reported previously, with minor modifications [4]. Heparinized whole blood collected within 24 h of testing or blood collected into EDTA within 6 h of testing was used. The panel of monoclonal antibodies used is described in Table 1. Cord blood samples contain significant numbers of nucleated erythrocytes (0-35%), which are present in the lymphoid gate based on forward-scatter (FSC) versus side-scatter (SSC) flow cytometric profiles. Gating of a minimum of 10 000 events based on PC5 versus SSC gating allowed exclusion of all nucleated erythrocytes [6]. Analysis was performed on a FACSCalibur flow cytometer using CellQuest software (Becton Dickinson, San Jose, CA, USA).

Table 1.

Combinations of fluorescently conjugated monoclonal antibodies used to identify cell type and various activation markers

FITC	PE	PC5
CD45RABD	CD45ROBD	CD4IM
CD25D	CD69BD	CD3IM
HLA-DRBD	CD71D	CD3IM
	CD122P	CD3IM
CD25D	CD69BD	CD14IM
TFD	CD71D	CD14IM
CD3BD	CD69BD	CD56IM
CD3BD	HLA-DRBD	CD56IM

^BDBecton Dickinson, CA, USA:

^DDako, Glostrup, Denmark:

^IMCoulter/Immunotech, Florida, USA. Samples were collected using live gating using PC5 versus side-scatter to exclude other leucocytes and nucleated red blood cells.

Serial testing

Serial testing for leucocyte activation markers was performed on blood samples from a group of 17 infants with proven bacterial infection. Samples from these infants were tested at birth (cord blood) and weekly postnatally until patient discharge from hospital.

Statistical analysis

Group comparisons were made using Student's t-tests for independent samples and anova, at a 95% confidence level.

RESULTS

Reference ranges

The proportions of the various leucocyte subsets expressing activation markers were established using blood samples taken from specimens which were received for routine postnatal investigation, with complete blood pictures demonstrating no evidence of infection. The percentages of leucocyte subsets expressing activation markers are shown in Table 2 for the different control age groups. TF was expressed on significantly more cord blood derived monocytes (P< 0.009) from patients <35 weeks’ gestational age than from patients born at 35-40 weeks. CD25 and HLA-DR was expressed on significantly more peripheral blood-derived T cells (P< 0.05) from patients 7-14 days old compared to 1-7 days old. CD25 and CD69 was expressed on significantly more monocytes (P< 0.05) from patients 7-14 days old compared to 1-7 days old. HLA-DR was expressed on significantly more NK cells (P< 0.05) from blood from patients 7-14 days old compared to 1-7 days old.

Table 2.

The percentages of leucocyte subsets expressing activation markers are shown for the different control age groups

Group		TRA/RO	T RO	T 25	T 69	T 71	T DR	T 122	M25	M69	M71	M TF	NK69	NKDR
35-40 weeks	mean	4.2	8.2	9.4	9.6	5.3	4.6	4.4	4.0	11.6	6.7	3.6	10.2	8.1
n = 30	SD	2.5	5.3	3.6	7.1	2.5	3.1	2.6	4.4	5.1	6.5	3.2	7.3	5.1
1–7 days	range	0–9.3	0–18.5	0–16.6	0–23.6	0–10.2	0–10.8	0–9.5	0–12	0–23	0–19.4	0–9.9	0–24.8	0–18.3
<35 weeks	mean	4.9	7.3	10.5	8.1	6.4	4.5	4.8	4.1	7.7	6.1	6.9	12.9	8.0
n = 25	SD	3.2	3.9	3.5	5.0	3.1	3.5	3.5	2.9	5.1	4.5	4.8	5.4	5.9
1–7 days	range	0–11.3	0–15	0–17.5	0–18.1	0–12.4	0–11.5	0–11.8	0–10.1	0–16.6	0–17	0–17*	0–23.6	0–19.8
	mean	7.7	14.7	13.9	11.1	10.1	10.8	8.4	10.3	17.7	7.2	6.6	9.7	14.9
n = 25	SD	3.6	6.9	6.6	5.5	6.3	7.0	4.9	11.0	10.8	5.3	3.5	4.8	6.8
7–14 days	range	0–15	0–28	0–29.5*	0–22	0–22	0–24.5*	0–19	0–32*	0–39*	0–19	0–17	0–19.5	0–27.5*

T, CD3 or CD4 positive T cells: RA, CD45RA: RO, CD45RO: 25, CD25: 69, CD69: 71, CD71: DR, HLA-DR: 122, CD122: M, monocyte: TF, tissue factor: NK, natural killer cells. TF was expressed on significantly more monocytes (P< 0.009) from cord blood from patients < 35 weeks’ gestational age than at 35–40 weeks. CD25 and HLA-DR was expressed on significantly more T cells (P< 0.05) from blood from patients at 7–14 days old compared to 1–7 days old. CD25 and CD69 was expressed on significantly more monocytes (P< 0.05) from blood from patients at 7–14 days old compared to 1–7 days old. HLA-DR was expressed on significantly more NK cells (P< 0.05) from blood from patients at 7–14 days old compared to 1–7 days old.

Infants with confirmed infection

The percentage of leucocyte subsets expressing activation markers from infants with positive blood culture is shown in Table 3. A variety of strains of bacteria were isolated. All positive cultures were associated with up-regulation of activation markers. The most sensitive indicator of infection was the up-regulated expression of CD69 on NK cells. This marker was increased significantly in 13 of 16 patients with positive cultures. CD69 expression was increased significantly on T cells from 11 of 17 patients. Up-regulated CD25 expression on T cells was the next most sensitive marker for infection from 10 of 17 patients. The expression of dual CD45RA/CD45RO on T cells was the next most sensitive markers of infection. CD45RO was also expressed on T cells from approximately 50% of infected infants. Other markers were less sensitive, and ranged down to the expression of CD25 and TF on monocytes from approximately 7% of infected patients.

Table 3.

The percentages of leucocyte subsets expressing activation markers from patients with positive culture

Patient	Culture	TRA/RO	T RO	T 25	T 69	T 71	T DR	T 122	M25	M69	M71	M TF	NK69	NKDR
1	gbs	12	14	8	14	24*	8	5	3	2	14	5	23	30
2	s.aureus	†		32	42				4	4			27	11
3	s.aureus	8	20	16	35	7	4	11				8	70	15
4	c + s	13	15	23	23	17	18	3		11	3		80	7
5	ser	6	9	18	12	2	2	5	4	3	23	15	44	20
6	c-s	16	17	33	43	29	8	7	1	1	3	15		10
7	gbs	15	6	13	11	13	3	6	4	1	5		75	8
8	ecoli	13	23	15	21			8	12	7	9	6	23	21
9	ser	18	29	7	31	23	28	15	1	1	15	9	12	6
10	c + s/ecoli	7	12	28	32	13	12	5	1	25	5	5	43	16
11	mycoplas	2	4	19	5	10	6	4	6	36	10	3	26	4
12	ser	19	37	15	16	6	20	17	1	1	12	16	18	13
13	klebs	4	7	18	25	4	8	5	5	5	21	3	12	11
14	c + s	8	21	32	22	8	3	5	4	3	3	18	26	11
15	c + s/ser	10	30	29	15	10	10	15	2	2	8	5	33	18
16	c-s	15	18	19	53	17		9	23	20		14	37	39
17	c + s/gbs	11	32	2	26	2	14	18	15	22	4	17	47	10
Score		9/16	7/16	10/17	11/17	2/15	3/14	4/16	1/15	3/16	2/14	1/15	13/16	6/16

Gbs, group B streptococcus: s.aureus, Staphylococcus aureus: c +s, coagulase positive staphylococcus: ecoli, Escherichia coli: ser, Serratia marscecens: klebs, Klebsiella: mycoplas, Mycoplasma pneumoniae.

^*Bold type indicates significantly increased percentage of leucocyte subset.

^†Blank data-test not performed due to insufficient sample. Expression of CD69 on NK cells is the most sensitive marker for infection (13 of 16 patients show significantly increased levels). Expression of CD25 and tissue factor (TF) on monocytes, are the least sensitive markers of infection.

Results of serial testing

Serial testing on blood from 14 infants with proven infection was performed. There were some general trends in the duration of raised activation markers on leucocyte subsets from infants with proven infection during their hospitalization in the intensive care unit. In patients where levels of CD69 was significantly raised on T and NK cells, the expression of this marker was up-regulated transiently on the first day or two following patient reports of positive cultures. CD25 expression on T cells returned to ‘normal’ levels within several days while the up-regulated expression of CD45RO remained significantly increased for approximately a week following reports of negative blood cultures after which time ‘normal’ levels were detected (data not shown).

Infants with possible infection

Forty patients were included in the ‘possible infected’ group based on clinical suspicion of infection. These infants were subsequently shown to fulfil none of the selection criteria for infected infants. Of these patients, 12 showed elevated levels of some leucocyte activation markers. Ten of these patients showed elevated levels of CD69 expression on NK cells, eight patients showed elevated levels of CD25 and CD69 on T cells, seven patients showed increased expression of CD45RA/CD45RO and six patients showed elevated levels of CD45RO on CD4 positive T cells.

DISCUSSION

This is the first report of the up-regulation of CD69 on NK cells in neonatal infection. We have shown previously that CD45RO expression on CD4 positive T cells was a reasonably sensitive marker for neonatal infection and that the up-regulation of CD45RA/CD45RO occurred early in the infective process [2]. We hypothesized that the incorporation of other early activation markers may contribute additional sensitivity to this test, especially in the early stages of infection, as we have shown previously that up-regulation of CD69 occurs within 2 h and CD25 occurs within 4 h following cord blood stimulation in vitro[4]. In the present study, up-regulated expression of CD25 and CD69 was detected at a higher rate than CD45RO on T cells from neonates with proven infection. However, the single most sensitive marker of neonatal infection was the up-regulated expression of CD69 on NK cells. The levels of expression of CD69 on NK cells in newborns were consistent with those found in a previous study (17% compared to 18%) [4]. CD69 is a molecule involved very early in signal transduction, contributing to signalling of cytokine and cytokine receptor synthesis [5]. Our findings indicate that a variety of infectious agents cause up-regulation of this signalling molecule on NK cells in the neonate, more consistently than any other surface activation marker. Up-regulated expression of other markers of activation on T cells i.e. CD71, the transferrin receptor, reported to be present on most proliferating cells [7]; HLA-DR, the major histocompatability class II complex, associated with antigen-presentation; and CD122, the IL-2βR, associated with IL-2 signal transduction, were less sensitive indicators for neonatal infection. Although the expression of HLA-DR on NK cells was up-regulated on 6/16 patients with proven infection, the inclusion of this marker did not improve predictability of infection compared to the expression of NK CD69. Activation markers on monocytes proved to be less reliable indicators of infection compared to T and NK cell markers. Infants infected with the same microorganisms did not show any obvious similarity in the pattern of up-regulation of surface markers (other than the more sensitive markers). However, numbers were too small to draw statistical significance, and a larger study would needed to be undertaken to further examine this observation.

Serial testing of neonates with proven infection showed trends consistent with those from our studies of stimulated in vitro whole blood cultures [4]. We showed that CD69 was up-regulated first followed by CD25 after in vitro stimulation. We also showed up-regulation of CD45RA/CD45RO occurred within 48 h followed by CD45RO on CD4 positive T cells at 72 h following in vitro stimulation [4]. Larger numbers of patients also need to be studied to confirm these ex vivo findings.

To eliminate the problem of contaminating nucleated erythrocytes in standard flow cytometric gating techniques used to identify lymphocytes based on forward versus side-scatter profiles, PC5 versus side-scatter gating was performed. This is a technique similar to a previous study using CD45 versus SSC gating to eliminate nucleated erythrocytes in patients with thalassaemia [6]. While differences in activation markers between leucocyte subsets were not significantly different between groups less than 35 weeks’ gestational age and term newborns (except monocyte expression of tissue factor), there was a significant increase in T cell expression of CD25 and HLA-DR, monocyte CD25 and CD69 and expression of HLA-DR on NK cells. The increase in these activation markers may reflect a maturation of these functional surface molecules, as we have shown similar up-regulation of these molecules between adults and cord blood [4]. Recently, the up-regulation of expression of CD64 on neutrophils was shown to be a sensitive marker of neonatal sepsis [1]; however, the processing technique involved to stain neutrophils with CD64 is time-consuming and can be performed on only heparin anticoagulated blood. We have shown previously significant levels of apoptosis of neutrophils from blood collected into heparin after 24 h [8]. Apototic neutrophils have been shown previously to down-regulate surface receptors during apoptosis [9], suggesting that patient samples collected for neutrophil quantification of CD64 must be processed immediately and cannot be left for analysis until the following day. In contrast, our current study has shown that leucocyte phenotyping can be performed within 24 h and can often be performed on EDTA anticoagulated blood ‘left over’ following blood collected for a complete blood picture, a laboratory test which is usually performed on an infant with suspected infection. We have shown previously that lymphocyte subsets do not undergo detectable levels of apoptosis when blood is collected into heparin at 72 h or EDTA at 24 h [8].

We have also shown recently that l-selectin, an adhesion molecule present on leucocytes, is down-regulated from the surface of lymphocytes, neutrophils and monocytes in patients with B. pertussis infection [10]. Measurement of l-selectin levels may also prove to be a sensitive marker for other infective organisms and we are currently undertaking a study of suspected infected newborns to test this hypothesis.

In conclusion, this is the first report of the up-regulation of CD69 on NK cells as a sensitive marker of neonatal infection. A combination of this marker with CD45RA, CD45RO, CD25 and CD69 expression on T cells is the most sensitive and specific for early detection of neonatal infection.