Age-stratified analysis of serotype-specific baseline immunity against group B streptococcus

ABSTRACT

Group B streptococcus (GBS) vaccines are currently under development. Data on the natural immunity in diverse age groups will aid establishing the GBS immunization policy. In this study, thirty serum samples were collected from three age groups (neonates/infants, pregnant women, and the elderly) between August 2016 and July 2017. Serotype-specific opsonophagocytic activity (OPA) was assessed using a GBS multiplex opsonophagocytic killing assay (MOPA) against serotypes Ia, III, and V. The mean OPA titers for serotype Ia of the three age groups were not significantly different (p = .156), but tended to be lower in neonates/infants (mean ± standard deviation, 137 ± 278). For serotype III and V, the mean OPA titer of neonates/infants (338 ± 623 and 161 ± 445, respectively) was significantly lower than that of pregnant women (1377 ± 1167 and 9414 ± 6394) and the elderly (1350 ± 1741 and 3669 ± 5597) (p = .002). In conclusion, the lower levels of OPA titers against all tested serotypes in neonates/infants, despite high maternal titers, indicates that intrapartum GBS vaccinations may be required for efficient placental transfer of serotype-specific GBS antibodies with high avidity.

KEYWORDS:

• Streptococcus agalactiae
• group B streptococcus
• GBS
• immunoassay
• immunity
• opsonophagocytic killing assay

Introduction

Group B streptococcus (Streptococcus agalactiae; GBS) is frequently isolated from the gastrointestinal and genitourinary tracts of healthy adults.¹ Because asymptomatic recto-vaginal colonization by GBS is common (10–37%) during women’s reproductive years,² vaginal colonization by GBS might be associated with neonatal invasive diseases such as sepsis, pneumonia, and meningitis, which affect 1,930 live births per year in the United States alone.³ While current recommendations, including prenatal GBS screening and intrapartum antibiotic prophylaxis, have dramatically reduced early-onset disease (EOD), they have had no effect on late-onset disease (LOD) or late maternal colonization. Alarmingly, the adult GBS infection rate is also on the rise,⁴ especially in the elderly. For example, in the United States the incidence and mortality of invasive GBS infections in the elderly has increased over three fold over the past five years.³

Administering GBS vaccinations to pregnant women may be an alternate strategy for the prevention of invasive GBS disease in infants, and pregnant women themselves, thereby reducing the risk of premature births and stillbirths.⁵ It could also be the best strategy for preventing the high disease burden in the elderly.³ Capsular polysaccharide (CPS), the most important virulence factor of GBS, is categorized into 10 distinct polysaccharides (Ia, Ib, II to IX) based on their structural and serological differences.⁴ More than 86% of EOD and 93% of LOD cases are caused by serotypes Ia, III, and V.⁴ In adults, serotype V (27.5%) was shown to be the predominant serotype, followed by Ia (24.3%) and III (16.5%).⁴ Currently, tri- or penta-valent GBS polysaccharide-protein conjugate vaccines (GBS-PCV) containing Ia, III, and V polysaccharides are under development by major pharmaceutical companies.⁶

Enrolling a large number of pregnant women in clinical trials for the licensure of GBS vaccines is challenging due to the lower baseline incidence of primary endpoint of GBS invasive disease.^2,6 Thus, reliable immunogenicity assays are essential for the process of GBS vaccine licensure, as observed with Neisseria meningococcal group C conjugate vaccine approval.⁷ Previously, we developed a multiplexed, opsonophagocytic assay (MOPA) targeting serotypes Ia, III and V.⁸ Although the predominant serotype varies by country and population age, serotypes Ia, III, and V are the most prevalent globally. In this study, we measured the levels of baseline functional antibodies against GBS-CPS Ia, III, and V in three targeted groups (neonates/infants, pregnant women, and the elderly) using MOPA. The aim of this study was to obtain indirect information on the protective immunity levels required from GBS vaccination by comparing basal serotype-specific GBS immunity between a non-high-risk group (pregnant women) and high-risk groups (neonates/infants and the elderly).

Materials and methods

Ethics statement

The study was approved by the Korea University Guro hospital (KUGH IRB No. 2016GR0265), and informed written consent was obtained from all participants aged ≥18 years. The institutional review board of the KUGH waived written informed consent for the use of remnant neonatal serum samples after laboratory tests in the department of laboratory medicine. Subject records/information for neonates was anonymized and de-identified prior to analysis.

Study design and sample collection

Serum samples were collected from pregnant women, infants under 3 months of age, and the elderly ≥65 years old, who are the main targets of the GBS vaccine. Between January 2016 and March 2017, subjects who voluntarily agreed to participate in the study were recruited when they visited the outpatient clinic and 30 persons per group were sampled in order of application. Healthy subjects without comorbidities were included. In the case of neonates/infants, 30 cases were collected in the order of sample collection at the laboratory. In the twins, only the first newborn was included in the study. All study samples were obtained from subjects who had not received any antibiotics in the three days prior to blood sampling. All pregnant women underwent vaginal-rectal screening (swab and conventional culture) for GBS colonization at 35–37 weeks (at the time of admission in case of preterm delivery). They were sub-grouped by GBS colonization status, and the capsule types of the colonized GBS isolates were identified by the use of both phenotypic and genotypic assays.⁹

Multiplexed opsonophagocytic assay (MOPA)

GBS MOPA was conducted as described previously.⁸ Briefly, 10 µL of the serially diluted serum samples and GBS mixture (streptomycin-resistant serotype Ia, spectinomycin-resistant serotypes III and kanamycin-resistant serotypes V GBS strains) were added to each well of 96-well plates. After 30 min of incubation at room temperature, 42.8 μl of a differentiated HL-60 cell suspension (4.28 × 10⁵ cells per well) and 7.2 μl of baby rabbit complement (BRC; Pel-Freez Biological; Rogers, AR) were added to each well, after which the mixture was incubated for 45 min at 37°C with shaking. Afterward, 10 µL of the final reaction mixture from each well was spotted onto tryptic soy agar containing one of three antibiotics (streptomycin, spectinomycin, or kanamycin). Then, 2,3,4-triphenyltetrazolium chloride (TTC; Sigma-Aldrich) was overlaid onto the spotted plates, after which the plates were incubated overnight at 37°C. The surviving bacterial colonies on the plates were counted and the opsonophagocytic activity (OPA) of the serum samples was determined. The OPA titer was defined as the last dilution of the serum sample reducing colony forming units (CFUs) to 50% of the negative control.

Statistical analysis

All statistical analyzes were performed using SPSS 18.0 (SPSS Korea, Seoul, Republic of Korea). The one-way analysis of variance (ANOVA) is used to determine the statistical differences between the three age groups. A p-value < 0.05 is considered significant.

Results

The mean age of the neonates/infants, pregnant women, and the elderly was 1.3 months (range, 1–3 months), 31.9 years (range, 23–41 years), and 68.8 years (range 65–76 years), respectively. There were 15 premature (gestational age <37 weeks, 53.3%) and 14 full-term (46.7%) newborns in the neonate/infant group. The baseline OPA titers of each age group are shown in Tables 1 and 2.

Table 1. Comparison of baseline opsonophagocytic activity (OPA) for group B streptococcus (GBS) between risk groups.

Serotypes	Groups*	OPA titers (mean)	95% CI	p-value
Ia	neonates/infants	137	33 – 241	0.596
	pregnant women	285	17 – 552
	old adults	231	3 – 459
III	neonates/infants^a	338	105 – 570	0.002
	pregnant women^b	1377	941 – 1813
	old adults^b	1350	700 – 2000
V	neonates/infants^a	161	−5 – 327	<0.001
	pregnant women^b	9414	6825 – 12003
	old adults^c	3669	1579 – 5759

* The same letters (alphabetical superscripts) indicate non-significant difference between groups based on Tukey’s multiple comparison test.

Table 2. Comparison of baseline opsonophagocytic activity (OPA) for group B streptococcus (GBS) between risk groups: proportion in each range stratified by the magnitude of OPA titer.

Serotype	OPA titer	Case No. (%)			p-value
		neonates/infants	pregnant women	old adults
Ia	<100	21 (70)	24 (80)	25 (83.3)	0.156
	100–499	6 (20)	2 (6.7)	2 (6.7)
	500–999	2 (6.7)	0 (0)	0 (0)
	>1,000	1 (3.3)	4 (13.3)	3 (10)
III	<100	18 (60)	4 (13.3)	4 (13.3)	<0.001
	100–499	4 (13.3)	4 (13.3)	6 (20)
	500–999	6 (20)	6 (20)	7 (23.3)
	>1,000	2 (6.7)	16 (53.3)	13 (43.3)
V	<100	23 (76.7)	0 (0)	2 (6.7)	<0.001
	100–499	5 (16.7)	0 (0)	2 (6.7)
	500–999	1 (3.3)	2 (6.7)	5 (16.7)
	>1,000	1 (3.3)	28 (93.3)	21 (70)

For serotype Ia, the mean OPA titers of the three age groups were not significantly different (p = .596). Notably, relatively low OPA titers were seen in the neonate/infant group compared to the pregnant women and the elderly groups (Table 1, Figure 1(a)). When the magnitude of the OPA titers were stratified into four ranges (Table 2), 70% of neonates/infants, 80% of pregnant women and 83.3% of the elderly had low titers below 100, but there were no significant differences in the distribution of OPA ranges among the three age groups.

Figure 1. Comparison of baseline opsonophagocytic activity (OPA) titers for group B streptococcus (GBS) between three age groups (upper panel, mean ± standard deviation; lower panel, point distribution): (a) serotype Ia, (b) serotype III, and (c) serotype V. NOTE: the same letters (alphabetical superscripts) indicate non-significant differences between groups based on Tukey’s multiple comparison tests.

For serotype III, the mean OPA titers were significantly lower for neonates/infants than those in the other two groups (p = .002), but there was no significant difference between the OPA titer of pregnant women and the elderly (Table 1, Figure 1(b)). Overall, 60% of neonates/infants (n = 18) showed OPA titers below 100; the proportion of OPA titers below 100 was significantly higher in neonates/infants (60%) than in pregnant women (13.3%) or the elderly (13.3%) (p < .001) (Table 2).

The OPA titers of serotype V were notably lower in neonates/infants (mean ± standard deviation, 161 ± 445) than in the elderly (3669 ± 5597) or pregnant women (9414 ± 6394), with significant differences between the three age groups (p < .001) (Table 1, Figure 1(c)). Moreover, 76.7% of neonate/infants had OPA titers below 100, whereas 93.3% of pregnant women and 70% of the elderly had high titers greater than 1,000, showing the significant differences between the groups (p < .001) (Table 2)

Of the 30 pregnant women who were included in the study, 4 (13.3%) were colonized with GBS on vaginal-rectal screening. There were no significant differences in the OPA titers for each serotype in GBS-colonized and GBS non-colonized pregnant women: serotype Ia (1,230 ± 1,467 versus 139 ± 416, p = .234), III (1,705 ± 1,569 versus 1,327 ± 1,124, p = .556), and V (6,318 ± 7,465 versus 9,890 ± 6,879, p = .346) (Figure 2). When the magnitude of the OPA titers were stratified into the four ranges, there were no significant differences in the distribution of OPA titer for each serotype in GBS-colonized and GBS non-colonized pregnant women (Table 3). Regardless of GBS colonization status, pregnant women had relatively high OPA titers, above 1,000 for serotype III (≥ 50%) and serotype V (≥ 90%). The OPA titers for colonized serotypes were generally high in colonized pregnant women, but did not always correlate with colonization. Case 3 in Table 4 carried serotypes III and Ⅷ but also showed a relatively high OPA titer, above 1,000, for serotype Ia. The OPA titers for serotype V were consistently high, titers above 1,000, regardless of colonization status (Table 4).

Table 3. Comparison of baseline opsonophagocytic activity (OPA) titers between group B streptococcus (GBS)-colonized and non-colonized pregnant women: proportion in each range stratified by the magnitude of OPA titer.

Serotype	OPA titer	Case No. (%)		p-value
		GBS colonized (N = 4)	GBS non-colonized (N = 26)
Ia	<100	2 (50)	22 (84.6)	0.065
	100–499	0 (0)	2 (7.7)
	500–999	0 (0)	0 (0)
	>1,000	2 (50)	2 (7.7)
III	<100	0 (0)	4 (15.4)	0.764
	100–499	1 (25)	3 (11.5)
	500–999	1 (25)	5 (19.2)
	>1,000	2 (50)	14 (53.9)
V	<100	0 (0)	0 (0)	0.566
	100–499	0 (0)	0 (0)
	500–999	0 (0)	2 (7.7)
	>1,000	4 (100)	24 (92.3)

Table 4. Group B streptococcus (GBS) serotypes and opsonophagocytic activity (OPA) titers among four GBS-colonized pregnant women.

No.	Age	Serotypes	OPA titer
			Serotype Ia	Serotype III	Serotype Ⅴ
1	30	NT	2	11	2,773
2	34	Ⅰa	2,001	3,218	3,020
3	29	III, Ⅷ	2,916	2,876	17,496
4	27	Ⅰb	2	614	1,984

Figure 2. Comparison of baseline opsonophagocytic activity (OPA) titers for group B streptococcus (GBS) between GBS-colonized and GBS non-colonized pregnant women at their 35–37 weeks prenatal screening: distribution of OPA titers in each group (red circles, GBS-colonized women and blue circles, GBS non-colonized women).

When we compared OPA titers (mean ± standard deviation) in preterm (N = 16) and full-term (N = 14) neonates/infants, there were no statistically significant differences for serotype Ia (207 ± 365 versus 57 ± 72, p = .127), III (355 ± 599 versus 319 ± 670, p = .878), and V (244 ± 599 versus 65 ± 96, p = .278) (Table 5). Prenatal maternal GBS screening was performed in 24 mothers of 30 neonates/infants, and three of them (12.5%) were colonized with GBS. When the OPA titers (mean ± standard deviation) in neonates/infants with GBS-colonized (N = 3) and GBS non-colonized mothers (N = 21) were compared, they were higher in neonates/infants with GBS-colonized mothers than in neonates/infants with GBS non-colonized mothers, though the differences were not statistically significant,: serotype Ia (411 ± 686 versus 115 ± 226, p = .534), III (1,640 ± 1,360 versus 209 ± 292, p = .209), and V (845 ± 1,339 versus 99 ± 168, p = .437). The neonates/infants with GBS-colonized mothers showed remarkably high OPA titers for all three serotypes, except for one preterm neonate/infant (preterm birth at 30 weeks), (Supplementary Table 1).

Table 5. Comparison of baseline opsonophagocytic activity (OPA) titers between preterm and full-term neonates/infants.

	Groups	Preterm (N = 16)	Full-term (N = 14)	p-value
Gestational age at birth, mean weeks (95% CI)		35.0 (33.6–36.4)	39.6 (39.0–40.3)	<0.001
Birth weight, mean Kg (95% CI)		2.2 (1.9–2.4)	3.2 (3.0–3.4)	<0.001
Twin pregnancy		2 (12.5%)	0 (0%)	0.171
Maternal GBS colonization		2/14 (14.3%)	1/10 (10%)	0.754
OPA titers, mean (95% CI)	Serotype Ia	207 (12–402)	57 (15–99)	0.127
	Serotype III	355 (35–674)	319 (−68 – 705)	0.878
	Serotype V	244 (−75 – 564)	65 (9–120)	0.278

GBS, group B streptococcus

Discussion

This is the first immunoassay performed using validated GBS-MOPA, to evaluate naturally acquired functional activity against serotypes Ia, III, and V in diverse age groups.⁸ Significantly lower OPA titers were seen in the neonates/infants, with the highest disease burden. In adults, OPA titers were higher, especially in pregnant women compared to the elderly. Even though maternal anti-GBS OPA titers were high, transplacental transfer of preexisting maternal antibodies can be quite limited. Similar to the Tdap (tetanus, diphtheria and pertussis) vaccine, GBS vaccination would be required for neonatal protection during each pregnancy. Although prenatal maternal GBS colonization increases the risk of neonatal GBS infection, GBS exposure during the late intrapartum period might induce high serotype-specific anti-GBS immunity in neonates/infants.

A correlation between low maternal serotype-specific capsular antibody concentration and increased risk of invasive GBS disease among newborns was observed,^10,11 and the specific concentration of maternal IgG required to protect neonates against invasive EOD was evaluated, mainly through enzyme-linked immunosorbent assays (ELISAs).^10,12–15 A matched case-control study conducted by Baker et al.¹³ reported that a threshold of ≥1 µg/mL in the mother at birth may provide a measure of protection against invasive GBS disease in newborns for serotype Ia, III and V. Lin et al.^14,15 had consistent findings, but described a 10-fold higher antibody concentration. Although these studies are not directly comparable due to differences in the methodologies used and the absence of a standardized immunological assay, the proportion of invasive GBS disease was generally lower in cases with serotype-specific capsular antibody titers ≥2 µg/mL compared to the controls in the meta-analysis.¹⁰ In a cross-sectional survey, conducted to assess the protective effect of serotype-specific antibodies in the elderly, older adults were found to lack the serotype V CPS-specific serum IgG, and they were more likely to carry serotype V, the leading cause of invasive disease in the elderly.¹²

Data on opsonophagocytic killing assays with GBS is very limited.^16–21 The opsonophagocytic killing assay was conducted to assess the in vitro function of serotype Ia, III or V GBS antibodies elicited in response to the GBS glycoconjugate vaccine in pregnant women,¹⁶ neonates,¹⁷ healthy non-pregnant adults,^18,19,21 and the elderly,²⁰ where each GBS vaccine uniformly promoted opsonophagocytosis against its homologous serotype. Only a few studies were conducted using GBS opsonophagocytic killing assays in a longitudinal cohort of pregnant women. A longitudinal study of 507 pregnant women, by Kwatra et al.²² found that women who were never colonized during pregnancy had higher OPA titers (≥8) in early pregnancy than those who had newly acquired GBS colonization during the study. In a mother/infant paired cohort study, Doare et al.²³ proposed a maternally-derived functional antibody threshold, above which no infant colonization was observed. In another cohort study by Kim et al.,²⁴ serotype-specific functional GBS antibodies among Korean infants were evaluated. Only a limited proportion of infants showed detectable OPA titers against serotypes Ia, Ib, and III GBS. In particular, the seropositive rate for serotype Ia was especially low, which is consistent with this study.

The protective cutoff values of OPA titers for GBS have not yet been established. Pneumococcus is known to have a protective effect at the relatively low OPA titers of ≥1:8 in pediatric patients;²⁵ however, the protective titer in adults has not been established even in pneumococcus. Given the results of this study and previous longitudinal cohort studies,^22–24 it is likely that transfer of GBS protection to neonates may require substantially higher OPA titers in pregnant women. The naturally acquired OPA titer was higher in the elderly than in children in this study, but they may require a higher protective OPA titer, as is seen with pneumococcus.²⁵ It is possible that the higher OPA titer seen in adults compared to neonates/infants is due to repeated exposure to diverse GBS strains, leading to nonspecific OPA against teichoic acid and surface proteins.²⁴ The clinical significance of OPA against these non-capsular antigens is unknown. The protective thresholds of serotype-specific OPA still need to be established and may vary substantially by age group.²⁶

In this study, pregnant women had high OPA titers, whereas neonates/infants showed relatively low titers, highlighting the possibility of ineffective placental transfer of pre-existing maternal antibodies. The variable efficiency for transfer among the IgG subclasses and the importance of timing in gestational week on antibody transfer has been demonstrated.²⁷ The efficiency of antibody transfer increases with gestational age; neonatal Fc receptors might be more highly expressed with advancing gestation.²⁷ Considering the efficiency of transfer and possibility that antibody persistence may be transient,²² the vaccine might need to be given in the third trimester. In contrast, there is evidence supporting vaccination during early pregnancy or before pregnancy. High OPA titers were associated with a lower risk of recto-vaginal acquisition of strains belonging to the homologous serotype,²² resulting in reduced infection of infants during birth,^22,28 as well as of the pregnant women themselves.²⁸ In addition, there are reports that the functionality of serotype-specific IgG persists for at least 18 months after immunization.¹⁸ There remains uncertainty about the immunization timing that would provide the best protection to newborns.

GBS colonization in Korean pregnant women is increasing. Previously, the GBS carriage rate in Korean pregnant women (2–6%) was estimated to be lower than that in the United States (15–20%).^29,30 However, several recent studies have reported an increase in carriage rates since 2005 (6.3–17.3%).^31–33 Serotypes III, V, and Ia were the most common in pregnant women (35.2–56.2%, 13.9–24.1%, and 12.1–26.2%, respectively) and infants (44.6–51.0%, 18.4–28.6%, and 14.3–15.3%, respectively).^30,32–35 In this study, conducted in South Korea, the GBS-OPA titer of serotype V tended to be higher in adults, including pregnant women, regardless of vaginal colonization. In contrast, the OPA titer of serotype Ia was generally low in all three age groups. This phenomenon could be partly explained by differences in the GBS strains. GBS capsular constituents, apart from surface adherence proteins, could influence cervicovaginal adherence or vaginal persistence.³⁶ In vitro analysis by Lubas et al.³⁷ found significantly lower adherence of serotype Ia GBS strains to the vaginal epithelium. A mouse model study carried out by Patras et al.³⁸ characterized the distinct phenotypes of serotype Ia, III, and V strains. The serotype V strain displayed greater intracellular survival and less cytokine stimulation in cervical cells, and increased persistence in the mouse vaginal tract while serotype Ia was the least adherent overall. These results highlight that serotype V acts differently from serotype Ia and III in its interactions with the host epithelium, which is consistent with its long-term colonization phenotype in vivo. The serotype-specific long-term colonization of GBS might be explained by the differences in their repertoire of cellular adhesins/invasins, their ability to outcompete normal flora, and/or their dialog with the host immune system. Another potential basis may lie in the structural complexity of their capsular polysaccharides. Serotype V possesses a set of 20 unique genes not found in other GBS strains, and is adept at forming biofilms in vitro.^38,39 The high OPA titers of serotype V might be explained by previous repeated colonization. In addition, a substantial proportion of immune response isotypes raised to serotype V are IgM and IgA, as well as IgG, so placental transfer could be lower than expected even in with high maternal OPA titers.^12,16,20 Careful interpretation is needed to establish the maternal protective threshold for neonates.

There are some limitations in this study. First, the relatively small sample size may have biased the results. Nevertheless, this is the first immunoassay using validated GBS-MOPA, and the results should be useful as reference data to establish thresholds for protection in the further clinical trials. Second, pregnant women could not be matched with their neonates/infants because the IRB did not permit the use of cord blood samples. Thus, this study could not provide any information on the transplacental transfer of maternal anti-GBS immunity. However, the OPA titers were consistently high in the pregnant women, even compared to healthy old adults, so it can be inferred that most healthy pregnant women would have similarly high OPA titers. Third, ELISA data was not available in this study because a standardized assay was not established. Fourth, because this study was conducted at a tertiary hospital, more than 50% (16 among 30) of the neonates/infants were preterm newborns with low birth weights. Although there was no statistically significant difference between preterm and full-term neonates/infants, it is possible that the OPA titers in this study are lower than those in healthy full-term neonates/infants. Finally, although GBS-MOPA was validated in a single center,⁸ this method needs to be further validated and standardized in collaboration with other institutions.^40,41

In conclusion, this study indicates that GBS OPA titers differ by serotype and between age groups. Despite the high serotype-specific OPA titers in pregnant women, these titers were consistently low in neonates/infants. GBS vaccines should be developed to ensure sufficient transplacental delivery of serotype-specific GBS antibodies irrespective of preexisting maternal immunity, and this needs to be confirmed by clinical trials. Although invasive GBS infection has increased in elderly people, most healthy elderly showed high levels of immunity against GBS in this study. GBS vaccination may not recommended for healthy older adults. Basal immunity in elderly people with comorbidities (diabetes, liver cirrhosis, malignancy, etc.) might differ from that in healthy older adults, so further assessment is required. The results in this study might be used as reference data in further vaccine clinical trials. Additional studies are required to assess the correlation between OPA and ELISA titers.

Disclosure of potential conflicts of interest

The authors disclosed no conflicts if interest.

Supplemental Material

Supplemental data for this article can be accessed online at http://dx.doi.org/10.1080/21645515.2019.1688036.

Additional information

Funding

This work was supported by the National Research Foundation of Korea under Grant [NRF-2017M2A2A6A02020925] to HSS and the Ministry of Food and Drug Safety under Grants [16172MFDS264] and [18172MFDS253] to JYS.