Vaccine escape of piliated Streptococcus pneumoniae strains

Vaccine 34 (2016) 2787–2792

Contents lists available at ScienceDirect

Vaccine journal homepage: www.elsevier.com/locate/vaccine

Vaccine escape of piliated Streptococcus pneumoniae strains夽 Gili Regev-Yochay a,b,∗ , Hanaa Jaber a , Ayob Hamdan c , Muhannad Daana d,e , Hanan Nammouz d , Amin Thalji f , Fuad Jaar h , Ziad Abdeen i , Carmit Rubin b , Amit Huppert g , Meir Raz d , Galia Rahav a , on behalf of the PICR study group a

Infectious Dis. Unit, Sheba Med Ctr, Ramat-Gan, Israel Section of Infectious Dis. Epidemiology, Gertner Institute for Epidemiology Research, Ramat-Gan, Israel Private Clinic, Nabulus, Palestine d Maccabi Healthcare Services, Jerusalem-Hashfela District, Israel e Hadassah University Medical Ctr, Jerusaelm, Israel f Private Clinic, Ramallah, Palestine g The Biostatistics Unit, Gertner Institute for Epidemiology Research, Ramat-Gan, Israel h Private Clinic, Bethlehem, Palestine i Al-Quds Nutrition and Health Research Institute, Al-Quds University, Abu-Dis, Palestine b c

a r t i c l e

i n f o

Article history: Received 17 December 2015 Received in revised form 16 April 2016 Accepted 20 April 2016 Available online 29 April 2016 Keywords: PCV7 effects Population-based study Pilus type-1 Pneumococcal vaccines

a b s t r a c t Introduction: Type1-pilus proteins were suggested as targets of future protein-based vaccines. Here we studied the effect of pneumococcal-conjugate vaccine (PCV7) implementation on the prevalence of piliated strains in a unique study setting which controls for typical confounders; the Palestinian-Israeli Collaborative Research (PICR). Methods: Annual cross-sectional surveys of pneumococcal carriage were performed during 2009–2011 among two closely related population that live under different health policies (a) Palestinian-Authority (PA) (n = 1773), where PCV7 was not yet introduced (b) East-Jerusalem (EJ) (n = 983) where PCV7 was rapidly implemented. Clinical data were collected, pneumococci identified and characterized and the presence of Type1-pilus genes was determined by rrgC PCR. Results: Following PCV7 implementation in EJ, overall carriage prevalence did not change (∼30%), but VT7 strains decreased from 61.5% to 33.8%. While prevalence of non-piliated-VT7 isolates decreased from 37% to 10%, p < 0.001, the prevalence of piliated-VT7 strains persisted ∼25%. Additionally, piliated non-VT13 strains emerged (1–15%, p < 0.001). These changes were not observed in PA. These dynamics were independent of the bacteria’s resistance pattern. Conclusions: A differential effect of PCV7 was observed with a relative resistance of piliated strains to the vaccine. This suggests that Type1-pilus confers an intrinsic advantage for colonization and may be an attractive vaccine target. © 2016 Elsevier Ltd. All rights reserved.

1. Introduction Streptococcus pneumoniae is a major cause of infectious disease morbidity and mortality, accounting for nearly 900,000 annual childhood deaths [17]. Nasopharyngeal carriage is the source of human to human transmission and serves as the first step to infection. Nearly all children are colonized by pneumococcus for some

夽 This study was partially presented at the International Symposium on Pneumococci and Pneumococcal Diseases (ISPPD9) Hyderabad, India, March 2014. ∗ Corresponding author at: Infectious Dis. Unit, Sheba Medical Center, Ramat-Gan 52621, Israel. Tel.: +972 3 5303500; fax: +972 3 5305301. E-mail addresses: [email protected], [email protected] (G. Regev-Yochay). http://dx.doi.org/10.1016/j.vaccine.2016.04.064 0264-410X/© 2016 Elsevier Ltd. All rights reserved.

time by the age of two [22]. While pneumococcal conjugate vaccines protect against colonization and mucosal disease and are thus highly effective, they are limited in covering only several capsular types. Therefore, future multi-valent protein vaccines are being investigated. The structural proteins of the pneumococcal pilus type-1 have been studied as such potential vaccine candidates, in particular, the pilus-1 subunits RrgA and RrgB [3,11,15]. S. pneumoniae pilus is a multimeric filamentous surface structure composed of subunit proteins attached to the cell wall. The pathogenicity islet Pilus-Islet-1 (PI-1) is encoded by the rlrA accessory region. This region consists of genes encoding three different pilus subunit proteins, RrgA, RrgB and RrgC, linked by three pilus-specific sortases. The PI-1, and particularly the RrgA subunit was shown to have an important role in adherence to

2788

G. Regev-Yochay et al. / Vaccine 34 (2016) 2787–2792

human respiratory epithelial cells [5,16]. Moreover, PI-1 was shown to influence colonization, virulence and inflammatory response in murine models [13,18]. Additionally, immunization with pilus subunits was shown to be protective against invasive pneumococcal disease (IPD) and bacteremia in a mouse model [11,15]. PI-1 is present in approximately 30% of all pneumococcal isolates and strikingly, this prevalence is similar across different geographic areas and various pneumococcal clones [2,14,20,21,23]. The presence of PI-1 was shown to be strongly associated with particular clones mainly of serotypes included in the sevenvalent pneumococcal conjugate vaccine (PCV7) [2]. It was therefore not surprising that within a few years after the introduction of PCV7 in the United States, the frequency of strains carrying the pilus genes significantly declined [6]. However, several years later, re-emergence of PI-1 was observed [20]. This was attributed to emergence of replacement strains not covered by the vaccine that frequently carried PI-1. This suggested that the presence of PI-1 may confer an evolutionary advantage to the organism. The Palestinian-Israeli Collaborative Research (PICR) crossconflict setting provided an opportunity to study overall and indirect effects of PCV7 vaccination whilst controlling for seasonal effects and secular trends [9]. In this study, two closely related Palestinian populations governed by two distinct health policies, where one was vaccinated and the other was not, were screened and studied. Here, we report the effects of PCV7 on the dynamics of S. pneumoniae strains harboring PI-1. 2. Materials and methods 2.1. Study design and study population During three consecutive years, repeated cross-sectional surveillance studies were carried out in two Palestinian populations that live under different health authorities with different PCV7 vaccination policies and minimal mixing; (1) Children (<5 years) living in East Jerusalem (EJ), under Israeli Health policy who visited any of four primary-care pediatric clinics of Maccabi Healthcare Services (MHS) (a major Israeli health maintenance organization). (2) Children living in the West Bank, under the Palestinian Authority (PA) policy, who visited any of three large private pediatric clinics in Bethlehem, Nabulus and Ramallah. 2.2. Study period Screening took place during summer months (May–July) of three consecutive years (2009–2011), in the two regions. 2.3. Vaccination policies and coverage in the two regions PCV7 was approved for use in 2007 in Israel, and introduced into the Israeli National Immunization Plan (NIP) in July 2009. It was introduced under a 2, 4, 12 month schedule with a 2-dose catch-up for infants aged <2 years, free of charge. In November 2010, PCV13 was approved and gradually replaced PCV7 in the NIP. During the study period, PCVs have not yet been introduced in PA. Vaccine (PCV7) coverage in both study populations in the first year was negligible (∼3%) but within one year of PCV7 introduction to EJ, over 70% of children <2 years of age, received at least one dose of PCV7 and nearly 60% received at least two doses. Due to the policy of gradual introduction of PCV13 in Israel, implementation was slow and by the end of our study, less than 20% of children <2 years in EJ, received two doses of PCV13 [8]. In PA, the number of children <2 years, who received at least one dose of PCV7 was ∼3% and increased gradually up to 13%, none received PCV13 [9].

2.4. Screening Nasopharyngeal S. pneumoniae carriage was detected using a rayon-tipped aluminum shaft swab placed in Amies transport media (Copan innovation, Brescia, Italy). Medical and vaccination histories and socio-demographic data were collected from medical files and the parent. 2.5. Laboratory Swabs were transported to the central PICR laboratory at Sheba Medical Center within 24 h. S. pneumoniae was isolated, identified and serotyped as previously described [9]. E-test strips for penicillin and ceftriaxone susceptibilities were used following Clinical and Laboratory Standards Institute (CLSI) guidelines. Strains with MIC >0.06 mg/L were defined as penicillin-nonsusceptible S. pneumoniae (PNSSP) and strains with MIC ≥2 mg/L as penicillin-resistant S. pneumoniae (PRSP). VT7 serotypes included: 4, 6A, 6B, 9V, 14, 18C, 19F and 23F. Serotype 6A was also defined as VT7, since it was shown to be highly cross-reactive with 6B and affected by PCV7 [10,19]. PI-1 positive strains were defined using PCR for rrgC, as described previously [20]. DNA was extracted using single bacterial colonies and rrgC and ply specific primers were used in a single reaction, where ply served as a positive control to assure pneumococci DNA was tested. 2.6. Statistical analyses Prevalence rates and proportions were calculated and compared using Chi square, or Fisher’s exact test for small numbers, and Cochran–Armitage trend tests. The potential effects of PCV7 implementation were assessed using an interaction term of year and region that allowed comparison of both pre- and post-vaccine introduction periods as well as vaccinated (EJ) and unvaccinated (PA) populations. This term was assessed in a multivariate logistic regression model of PI-1 carrying strains. This model adjusted for known and potential confounders, including demographic factors that differed in the two populations (p < 0.2) in univariate regression models (i.e. age, gender, carriage of VT7 strains and PRSP). Odds ratios and 95% confidence intervals (CIs) were calculated. All analyses were conducted using SAS 9.4 software. 2.7. Institutional Review Board (IRB) and patient consent IRB approvals were given by local committees of the Sheba Medical Center, Maccabi Healthcare Service (MHS) and An-Najah University. Written informed consent was given by a parent for each participating child before recruitment. 3. Results 3.1. Study population and S. pneumoniae carriage The study population was described in detail [9]. In brief, an average of 326 children from EJ and 590 children from PA were screened every summer during three consecutive years (2009–11). S. pneumoniae carriage was detected in 278 (28.4%) children in EJ and 583 (32.9%) in PA during the three-year study period, a difference that was not statistically significant. While prevalence of carriage in EJ did not change during the three years, carriage prevalence decreased non-significantly in PA from 36.0% to 28.8%.

G. Regev-Yochay et al. / Vaccine 34 (2016) 2787–2792

2789

Table 1 Proportion of piliated strains in EJ and PA. Study population

EJ (n = 980)

PA (n = 1771)

2009 (n = 345)

2010 (n = 311)

2011 (n = 324)

2009 (n = 620)

2010 (n = 595)

2011 (n = 556)

SP (n) Available serotype

100 91

91 91

87 77

223 223

200 198

160 145

VT7a strains % piliated among VT7 VT13-7b strains Non-VT13c strains % piliated among Non-VT13 Pilus pilus (%)

61.5% 42.6% 5.5% 33.0% 4.0% 26/89

31.9% 60.7% 3.3% 64.8% 8.5% 22/90

33.8% 68.0% 3.9% 62.3% 23.4% 31/85

50.2% 47.6% 4.1% 45.7% 15.7% 62/198

50.0% 55.6% 7.6% 42.4% 14.3% 69/200

56.5% 49.4% 9.7% 33.8% 29.2% 57/157

29.2% 95.8% 4.2% 97 16.5% 0% 83.5% 34.3% 29.9% 46.4%

24.4% 77.3% 22.7% 90 24.4% 0% 75.6% 32.4% 14.4% 69.2%

36.5% 58.6% 37.9% 87 19.5% 6.3% 80.5% 43.5% 12.6% 72.7%

31.3% 75.8% 22.6% 194 17.5% 10% 82.5% 34.5% 28.9% 66.0%

34.5% 79.7% 17.4% 199 16.1% 3.1% 83.9% 40.7% 21.1% 64.3%

36.3% 70.2% 24.6% 160 13.1% 10% 86.9% 40.2% 33.8% 63.0%

% VT7 among piliated strains % non-VT13 among piliated strains # with available susceptibility test PSSPd (% of all strains) % pilus in PSSP PNSSPe (% of all strains) % pilus in PNSSP PRSPf (% of all strains) % pilus in PRSP a b c d e f ␺ ␺␺

VT7 – vaccine type strains, belonging to the serotypes covered by of xx with available serotypes. VT13-7 – strains belonging to serotypes covered by PCV13 but not PCV7. Non-VT13 – non-vaccine type strains, belonging to serotypes not covered by PCV13. PSSP – penicillin susceptible S. pneumoniae (≤0.06 mg/L). PNSSP – penicillin non-susceptible S. pneumoniae (>0.06 mg/L). PRSP – penicillin resistant S. pneumoniae (≥2.0 mg/L). p Value for 3rd year value vs. 1st year value <0.05. p Value for 3rd year value vs. 1st year value <0.01.

3.2. Prevalence of piliated strains and their association with serotypes In the pre-vaccine period (2009), the proportion of piliated strains among all isolates was ∼30% in both populations. Most piliated isolates in that year belonged to VT7 serotypes; 96% and 76% in EJ and PA, respectively (Table 1). Within one year following PCV7 introduction in EJ, the prevalence of VT7 serotypes among all S. pneumoniae isolates, decreased significantly, from 61.5% to 31.9% (p ≤ 0.0001). During the second year follow-up, no further decrease was observed. VT7 strains were replaced by non-VT7 strains (38.5–68.1%), so that overall S. pneumoniae carriage did not change in EJ. In contrast, in PA, where PCV7 was not introduced, a non-significant increase in VT7 strains was observed (from 50.2% to 56.5%) (Fig. 1; Table 1).

Fig. 1. Proportion of vaccine type strains among the carried pneumococci in EJ and PA each year. Black – VT7 strains (serotypes 4, 6B, 9V, 14, 18C, 19F, 23F, and also including 6A). Grey – VT13-7 strains (serotypes 1, 3, 5, 7F, 19A), White – non-VT13 strains. Full black – non-piliated VT7 strains, striped black piliated VT7 strains.

Despite the significant decrease in VT7 strains in EJ, and the fact that initially, most piliated isolates belonged to VT7 serotypes, only a limited, non-significant decrease (from 29.2% to 24.4%) in the prevalence of piliated strains was observed in the first year following PCV implementation (2010). This was not observed in PA. The decrease in EJ was transient, with a re-emergence of piliated isolates from 24.4% to 36.5% in 2011, reaching similar prevalence to that observed in PA that year (36.3%) (Table 1). The effect of PCV7 on the prevalence of piliated and non-piliated VT7 and non-VT7 isolates was separately assessed in each region (Fig. 2). Interestingly, non-piliated VT7 isolates decreased significantly in EJ, from 37.4% to 10.7% (p < 0.001), while prevalence of piliated VT7 isolates did

Fig. 2. Carriage prevalence of piliated and non-piliated S. pneumoniae strains by vaccine type in (a) EJ and (b) PA. Light grey – 2009, dark grey – 2010, black – 2011.

2790

G. Regev-Yochay et al. / Vaccine 34 (2016) 2787–2792

Fig. 3. Distribution of serotypes that consisted of ≥2.5% of all isolates. White bars – 2009, grey bars – 2010, dark grey – 2011. Stripped – piliated strains. (a) VT13 serotypes in EJ. (b) Non-VT13 serotypes in EJ. (c) VT13 serotypes in PA. (d) Non-VT13 serotypes in PA.

not change (27.7–22.7% p = 0.43). In addition, piliated non-VT7 isolates increased in EJ from 1.2% to 14.7% (p < 0.001). These changes were not observed in PA. Indeed, since non-piliated VT7 strains decreased and piliated VT7 persisted in EJ, the proportion of piliated strains among VT7 strains only, increased significantly, from 42.6% in the pre-PCV year to 68% (ptrend = 0.02), while this proportion did not change in PA (∼45%) (Table 1). The decrease in VT7 strains in EJ was attributed to a decrease in prevalence of the common VT7 serotypes; 19F, 14, 6B, and 6A. The proportion of piliated isolates among these serotypes, particularly 6B, 19F and 14 increased, while most of 6A isolates were not piliated (Fig. 3a). These changes were not observed in PA (Fig. 3c). VT7 serotypes were replaced by non-VT13 serotypes in EJ. During the first (pre-PCV7) year in EJ, non-VT13 serotypes were all non-piliated, apart from a single isolate of serogroup 22. During the 2-years follow-up, non-VT13 isolates that emerged were of serotypes 15B/C, 35B, 11A and 19B, of these, all but 15B/C were frequently piliated (Fig. 3b). All together, the proportion of piliated isolates among non-VT13 strains in EJ increased from 4.0% to 23.4% (p = 0.01). In PA, a few piliated 11A and 35B isolates were detected in the first study year, but these did not increase in the following years (Fig. 3d) and no change was observed for the overall prevalence of piliated-non-VT13 isolates (7–10%) (Fig. 2). 3.3. Correlation between antibiotic resistance and prevalence of the pilus As previously reported, antibiotic use and carriage of antibiotic resistant S. pneumoniae were high in both regions, with over 80% of isolates being non-susceptible to penicillin [9]. The major effect of PCV7 on carriage of antibiotic resistant isolates, was observed for carriage of penicillin resistant S. pneumoniae (PRSP)

(MIC ≥2.0 mg/L) and multi-drug resistant (MDR) isolates (at least resistant to macrolides and non-susceptible to penicillin) that decreased dramatically (over 2-fold) in EJ, but did not change in PA. This finding was not surprising since most of the antibiotic resistant strains were of VT7 serotypes. Yet, unexpectedly, the piliated, PRSP isolates did not decrease, moreover, the proportion of piliated PRSP isolates increased from 46% to 73% in EJ. The proportion of piliated PRSP isolates was higher in PA (66% compared to 46% in EJ in the pre-vaccine period), but it did not change in PA during the study years (Table 1). In a multivariate logistic analysis, the main independent predictors for carrying a piliated strain were carriage of VT7 serotypes and carriage of PRSP strains. Yet, despite the reduction of VT7 and PRSP strains following PCV7 implementation, and despite the fact that most piliated strains were of VT7 and PRSP strains, PCV implementation was an independent predictor for carriage of piliated strains in the 2nd year post implementation (adjusted OR: 3.05, 95% CI: 1.2–7.8) (Table 2).

4. Discussion Here, for the first time, we show a differential effect of PCV on piliated strains so that within a particular VT-serotype, non-piliated strains were rapidly eliminated, while piliated strains persisted, at least during the first two years following PCV implementation. In addition, during this time period piliated non-VT strains emerged. Thus, piliated strains seem to have been selected by PCV. Moreover, while overall prevalence of antibiotic resistant isolates decreased, due to the decrease in VT strains, the proportion of piliated resistant isolates increased significantly. Thus, the piliated resistant strains also escaped the vaccine effect.

G. Regev-Yochay et al. / Vaccine 34 (2016) 2787–2792

2791

Table 2 Predictors for carrying piliated strains. Variable Child age group

Gender VT7 strains* PRSP strains Region Year

N

% Piliated

aOR

<6 m 6–11 m 1 years 2 years 3–5.5 years

133 220 222 125 118

33.8 33.2 35.1 30.4 27.1

Ref 0.627 0.714 0.770 0.680

0.4–1.1 0.4–1.2 0.4–1.4 0.3–1.3

Ref 0.095 0.223 0.417 0.247

Male Female

502 313

33.1 31.3

Ref 0.966

0.7–1.4

Ref 0.849

No Yes

399 387

16.0 51.4

Ref 4.208

2.9–6.2

Ref <0.001

No Yes

591 195

22.5 62.6

Ref 3.801

2.6–5.7

Ref <0.001

PA EJ

555 264

33.9 29.9

Ref 0.680

0.4–1.3

Ref 0.254

2009 2010 2011

287 290 242

30.0 31.4 36.4

Ref 1.343 1.304

0.8–2.2 0.8–2.2

Ref 0.252 0.333

1.297 3.045

0.5–3.3 1.2–7.8

0.579 0.02

PCV7 effect (2010) PCV7 effect (2011)

95% CI

ap-value

All variables entered to the model are shown. N – number of children. aOR – adjusted odds ratio. 95% CI – 95% confidence interval. ap-value – adjusted p-value.

We have previously shown that within one year of PCV7 implementation, a rapid decrease in VT7 strains was observed, yet, during the second year this decrease was abated [9]. A potential explanation for that observation is the differential effectiveness of PCV7 on non-piliated strains; while during the first year, most of the VT7 non-piliated strains were eliminated, these were relatively rare in the second year, while piliated-VT7 strains which were relatively resistant to the vaccine persisted. Our findings have potential implication for future protein-based pneumococcal vaccines. The importance of the PI-1 proteins as candidate targets of future protein vaccines has been debated. The fact that piliated strains are relatively resistant to the vaccine effect, strongly suggests that the pilus confers an advantage to carried strains. Some of the arguments against a pilus-protein based vaccine, were that only ∼30% of pneumococcal strains carry the PI-1 and that it is not necessarily associated with invasive disease. Yet, our findings suggest that the pilus confers a significant advantage to colonization, so that targeting it would potentially enhance the elimination of pneumococcal colonization. Moreover a vaccine that would target piliated strains would potentially prevent the emergence of piliated non-VT strains, which may be more efficient colonizers and thus a potential source for transmission and for endogenous invasive disease. It has been previously shown that within a few years after the introduction of PCV in Massachusetts, USA, piliated strains were nearly eradicated, but a few years later they re-emerged to previous levels [20]. While that observation already suggested that the pilus confers some advantage for colonization, confounding by secular trends could not be ruled out in that study. The unique PICR setting and novel study design allowed assessment of the actual dynamics of piliated strains following PCV implementation, controlling for seasonality and secular trends. This study has several limitations: The study follow-up duration was only three years, of these, the period of follow up after implementation of PCV7 was two years. Vaccine selection impacts were reported to be most prominent within 4–6 years of implementation [12] depending mainly on vaccine coverage rates but also on geography, pre-vaccine serotype distribution and population

characteristics (e.g., contact rates). A longer follow up is thus required to determine the magnitude of the vaccine selection pressure on piliated strains. Yet, even if this selection by PCV is short-lived, and further follow-up will show an eventual eradication of piliated VT strains, the fact that elimination of piliated VT strains is slower than that of non-piliated and that piliated non-VT strains emerge, implies that piliated strains have an evolutionary advantage and thus that pilus proteins should be considered as important candidate targets as one of the components for a future protein-based vaccine, which aims to target several conserved proteins. We did not directly assess the expression of the pilus proteins, but rather the genes encoding them, PI-1. Despite the report that the expression of PI-1 proteins is bi-stable [7], where in a single clone, at a certain time point, some cells will express the pilus proteins and other will not, the fact that isolates that carried these genes were selected by the vaccine probably suggests that the pilus is expressed to some degree, at some time point. Our study was also limited to studying PI-1 while another pilus islet (PI-2) has been recently identified. Yet, PI-2 has been reported to be carried by a smaller proportion of pneumococci (9–15%) and its expression is thought to have only a marginal contribution to bacterial adhesion to epithelial cells [4,14]. However, PI-2 was detected solely on serogroups 1, 7 and 19 and were reported to increase in the post-PCV era [1,24]. Further studies assessing the dynamics of PI-2 following PCV implementation would be interesting. In conclusion, this study demonstrated that piliated strains are selected by PCV7 at the population level. While PCV7 induced a rapid decrease of non-piliated VT7 strains, piliated VT7 strains were not affected by the vaccine during the first two years following PCV7 implementation. The association between the presence of pilus islet genes and the relative resistance to PCV does not necessarily implicate that pilus structural proteins confer this resistance, and other factors associated with the presence of these genes may promote such an effect. Yet, this observation suggests that further study of pilus proteins as potential candidates for future pneumococcal protein vaccines should be considered.

2792

G. Regev-Yochay et al. / Vaccine 34 (2016) 2787–2792

Funding This work was supported by the Gertner Institute, Maccabi Healthcare Services Research Institute (MIHSR-250809) and the Israel National Institute for Health Policy Research (NIHP) (NIHP25-10). All funding sources had no involvement in the study design, collection, analysis or interpretation of the data, nor did they have a role in writing the report or in the decision to submit the paper for publication. Acknowledgments PICR study group: Ziad Abdeen (PA), Izzeldin Abullaish (GZ), Muhammed Affiffi (EJ), Yair Amit (IL), Yunes Bassem (EJ), Adi Cohen (IL), Muhannad Daana (EJ), Ibrahim Dandis (EJ), Abedalla ElHamdany (GZ), Ayob Hamdan (PA), Samantha Hasselton (EJ), Amit Hupert (IL), Muhammed Husseini (EJ), Fuad Jaar (PA), Laduyeh Kawather (EJ), Marian Osher (IL), Galia Rahav (IL), Meir Raz (IL), Gili Regev-Yochay (IL), Avraham Rodity (IL), Hector Roizin (IL), Waeel Siag (EJ), Ora Stern (IL), Amin Thalji (PA), Luba Yakirevitch (IL), Khairi Zecayra (EJ). In addition, we would particularly like to thank Miriam Varon and Efrat Steinberger for coordinating the study, Aviva Goral for data management, Rula Haj Yehia for serotyping. Conflict of interest: All authors – no financial or personal conflicts of interest to declare. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.vaccine.2016.04. 064. References [1] Aguiar SI, Melo-Cristino J, Ramirez M. Use of the 13-valent conjugate vaccine has the potential to eliminate pilus carrying isolates as causes of invasive pneumococcal disease. Vaccine 2012;30(37):5487–90. [2] Aguiar SI, Serrano I, Pinto FR, Melo-Cristino J, Ramirez M. The presence of the pilus locus is a clonal property among pneumococcal invasive isolates. BMC Microbiol 2008;8:41. [3] Amerighi F, Valeri M, Donnarumma D, Maccari S, Moschioni M, Taddei A, et al. Identification of a monoclonal antibody against pneumococcal pilus-1 ancillary protein impairing bacterial adhesion to human epithelial cells. J Infect Dis 2015. [4] Bagnoli F, Moschioni M, Donati C, Dimitrovska V, Ferlenghi I, Facciotti C, et al. A second pilus type in Streptococcus pneumoniae is prevalent in emerging serotypes and mediates adhesion to host cells. J Bacteriol 2008;190(15): 5480–92. [5] Barocchi MA, Ries J, Zogaj X, Hemsley C, Albiger B, Kanth A, et al. A pneumococcal pilus influences virulence and host inflammatory responses. Proc Natl Acad Sci U S A 2006;103(8):2857–62. [6] Basset A, Trzcinski K, Hermos C, O’Brien KL, Reid R, Santosham M, et al. Association of the pneumococcal pilus with certain capsular serotypes but not with increased virulence. J Clin Microbiol 2007;45(6):1684–9.

[7] Basset A, Turner KH, Boush E, Sayeed S, Dove SL, Malley R. An epigenetic switch mediates bistable expression of the type 1 pilus genes in Streptococcus pneumoniae. J Bacteriol 2012;194(5):1088–91. [8] Ben-Shimol S, Givon-Lavi N, Greenberg D, Dagan R. Pneumococcal nasopharyngeal carriage in children <5 years of age visiting the pediatric emergency room in relation to PCV7 and PCV13 introduction in southern Israel. Hum Vaccin Immunother 2016;12(2):268–76. [9] Daana M, Rahav G, Hamdan A, Thalji A, Jaar F, Abdeen Z, et al. Measuring the effects of pneumococcal conjugate vaccine (PCV7) on Streptococcus pneumoniae carriage and antibiotic resistance: the Palestinian-Israeli Collaborative Research (PICR). Vaccine 2015;33(8):1021–6. [10] Feikin DR, Kagucia EW, Loo JD, Link-Gelles R, Puhan MA, Cherian T, et al. Serotype-specific changes in invasive pneumococcal disease after pneumococcal conjugate vaccine introduction: a pooled analysis of multiple surveillance sites. PLoS Med 2013;10(9):e1001517. [11] Gianfaldoni C, Censini S, Hilleringmann M, Moschioni M, Facciotti C, Pansegrau W, et al. Streptococcus pneumoniae pilus subunits protect mice against lethal challenge. Infect Immun 2007;75(2):1059–62. [12] Hanage WP, Finkelstein JA, Huang SS, Pelton SI, Stevenson AE, Kleinman K, et al. Evidence that pneumococcal serotype replacement in Massachusetts following conjugate vaccination is now complete. Epidemics 2010;2(2): 80–4. [13] Hilleringmann M, Ringler P, Muller SA, De Angelis G, Rappuoli R, Ferlenghi I, et al. Molecular architecture of Streptococcus pneumoniae TIGR4 pili. EMBO J 2009;28(24):3921–30. [14] Hjalmarsdottir MA, Petursdottir B, Erlendsdottir H, Haraldsson G, Kristinsson KG. Prevalence of pilus genes in pneumococci isolated from healthy preschool children in Iceland: association with vaccine serotypes and antibiotic resistance. J Antimicrob Chemother 2015;70(8):2203–8. [15] Moschioni M, De Angelis G, Harfouche C, Bizzarri E, Filippini S, Mori E, et al. Immunization with the RrgB321 fusion protein protects mice against both high and low pilus-expressing Streptococcus pneumoniae populations. Vaccine 2012;30(7):1349–56. [16] Nelson AL, Ries J, Bagnoli F, Dahlberg S, Falker S, Rounioja S, et al. RrgA is a pilus-associated adhesin in Streptococcus pneumoniae. Mol Microbiol 2007;66(2):329–40. [17] O’Brien KL, Wolfson LJ, Watt JP, Henkle E, Deloria-Knoll M, McCall N, et al. Burden of disease caused by Streptococcus pneumoniae in children younger than 5 years: global estimates. Lancet 2009;374(9693):893–902. [18] Orrskog S, Rounioja S, Spadafina T, Gallotta M, Norman M, Hentrich K, et al. Pilus adhesin RrgA interacts with complement receptor 3, thereby affecting macrophage function and systemic pneumococcal disease. MBio 2012;4(1), e00535-12. [19] Park IH, Moore MR, Treanor JJ, Pelton SI, Pilishvili T, Beall B, et al. Differential effects of pneumococcal vaccines against serotypes 6A and 6C. J Infect Dis 2008;198(12):1818–22. [20] Regev-Yochay G, Hanage WP, Trzcinski K, Rifas-Shiman SL, Lee G, Bessolo A, et al. Re-emergence of the type 1 pilus among Streptococcus pneumoniae isolates in Massachusetts, USA. Vaccine 2010;28(30):4842–6. [21] Regev-Yochay G, Lipsitch M, Basset A, Rubinstein E, Dagan R, Raz M, et al. The pneumococcal pilus predicts the absence of Staphylococcus aureus co-colonization in pneumococcal carriers. Clin Infect Dis 2009;48(6): 760–3. [22] Simell B, Auranen K, Kayhty H, Goldblatt D, Dagan R, O’Brien KL. The fundamental link between pneumococcal carriage and disease. Expert Rev Vaccines 2012;11(7):841–55. [23] Turner P, Melchiorre S, Moschioni M, Barocchi MA, Turner C, Watthanaworawit W, et al. Assessment of Streptococcus pneumoniae pilus islet-1 prevalence in carried and transmitted isolates from mother-infant pairs on the ThailandBurma border. Clin Microbiol Infect 2012;18(10):970–5. [24] Zahner D, Gudlavalleti A, Stephens DS. Increase in pilus islet 2-encoded pili among Streptococcus pneumoniae isolates, Atlanta, Georgia, USA. Emerg Infect Dis 2010;16(6):955–62.