PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Vaccine. Author manuscript; available in PMC 2012 July 18.
Published in final edited form as:
PMCID: PMC3138857
NIHMSID: NIHMS303269

International Seroepidemiology of Adenovirus Serotypes 5, 26, 35, and 48 in Pediatric and Adult Populations

Abstract

Recombinant adenovirus serotype 5 (rAd5) vaccine vectors for HIV-1 and other pathogens have been shown to be limited by high titers of Ad5 neutralizing antibodies (NAbs) in the developing world. Alternative serotype rAd vectors have therefore been constructed. Here we report Ad5, Ad26, Ad35, and Ad48 NAb titers in 4,381 individuals from North America, South America, sub-Saharan Africa, and Southeast Asia. As expected, Ad5 NAb titers were both frequent and high magnitude in sub-Saharan Africa and Southeast Asia. In contrast, Ad35 NAb titers proved infrequent and low in all regions studied, and Ad48 NAbs were rare in all regions except East Africa. Ad26 NAbs were moderately common in adults in sub-Saharan Africa and Southeast Asia, but Ad26 NAb titers proved markedly lower than Ad5 NAb titers in all regions, and these relatively low Ad26 NAb titers did not detectably suppress the immunogenicity of 4×1010 vp of a rAd26-Gag/Pol/Env/Nef vaccine in rhesus monkeys. These data inform the clinical development of alternative serotype rAd vaccine vectors in the developing world.

1. Introduction

A limitation that has become apparent with rAd5 vaccine vectors is the high titers of Ad5 NAbs in human populations, particularly in the developing world. Baseline Ad5 NAbs have been shown to suppress the immunogenicity of rAd5 vector-based vaccines for HIV-1 in both preclinical studies [12] and clinical trials [35], although higher doses of rAd5 vectors can partially overcome this effect. To address this and other problems with rAd5 vectors, alternative human serotype rAd vectors [68], hexon-chimeric rAd vectors [2], and rAd vectors derived from other species [910] have been constructed.

In particular, rAd26 and rAd35 vectors are currently being evaluated in phase 1 HIV-1 vaccine clinical trials in both the United States and sub-Saharan Africa. Ad26 (subgroup D), Ad35 (subgroup B), and Ad48 (subgroup D) are derived from different Ad subgroups than Ad5 (subgroup C). Moreover, these alternative Ad serotypes differ from Ad5 in terms of their receptor usage [6, 8], tropism [11], dendritic cell stimulatory capacity [12], innate immune profile (D.H.B., unpublished data), adaptive immune phenotype [13], and capacity to protect against SIV challenge in rhesus monkeys [14].

Building on Ad seroepidemiology studies previously reported from our laboratory and others [6, 8, 1519], we report here a large study of Ad5, Ad26, Ad35, and Ad48 NAb titers in 4,381 pediatric and adult subjects from North America, South America, sub-Saharan Africa, and Southeast Asia. We also model the impact of typical Ad26 Nab titers found in the developing world on rAd26 vaccine immunogenicity in a pilot study in rhesus monkeys.

2. Materials and Methods

2.1 Study Populations

This study involved 4,381 subjects from pediatric and adult populations in North America, South America, sub-Saharan Africa, and Southeast Asia. Subjects included both low HIV-1 risk and high HIV-1 risk adult populations from multiple geographic regions, as well as healthy infants and schoolchildren from South Africa. Random or case-controlled baseline samples were utilized in these studies to minimize selection bias. Table 1 details the specific cohorts in this study. All samples were collected with local Institutional Review Board (IRB) approvals, and adenovirus neutralization assays utilizing these samples were approved by the Beth Israel Deaconess Medical Center IRB.

Table 1
Study populations for adenovirus seroepidemiology studies

2.2 Adenovirus Neutralization Assay

Ad-specific NAb titers were assessed by high-throughput luciferase-based virus neutralization assays as described [20]. A549 human lung carcinoma cells were plated at a density of 1×104 cells per well in 96-well plates and infected with E1/E3-deleted, replication-incompetent rAd-Luc reporter constructs at a multiplicity of infection (MOI) of 500 with 2-fold serial dilutions of serum in 200 µ1 reaction volumes. Following a 24-hour incubation, luciferase activity in the cells was measured using the Steady-Glo Luciferase Reagent System (Promega, Madison, WI) with a Victor 1420 Multilabel Counter (Perkin Elmer, Wellesley, MA). Neutralization titers were defined as the maximum serum dilution that neutralized 90% of luciferase activity. rAd5-Luc, rAd26-Luc, rAd35-Luc, and rAd48-Luc vectors exhibited comparable virus particle to plaque-forming unit (vp/pfu) ratios (10–30) and similar infectivity. NAb assays utilizing these four vectors demonstrated similar performance characteristics and dynamic ranges.

2.3 Animals and Immunizations

Adult Indian-origin rhesus monkeys (N=12) were housed in the bio-level 3 containment facility at New England Primate Research Center (NEPRC). Animals were inoculated twice separated by four weeks by the intranasal route with either 1011 viral particles (vp) replication-competent, E1-positive rAd26-Empty vectors (N=6) or saline (N=6). After four weeks, animals were vaccinated by the intramuscular route with 4×1010 vp replication-incompetent, E1/E3-deleted rAd26-Gag/Pol/Nef/Env vectors in the quadriceps muscles. Cellular immune responses were assessed by interferon-γ ELISPOT assays and multiparameter intracellular cytokine staining (ICS) assays essentially as described [1314]. All animal studies were approved by the Harvard Medical School Institutional Animal Care and Use Committee (IACUC).

3. Results

3.1 Adenovirus NAb Titers in Multiple International Human Populations

We determined Ad5, Ad26, Ad35, and Ad48 NAb titers in 4,381 subjects in multiple international populations from North America, South America, sub-Saharan Africa, and Southeast Asia as detailed in Table 1. Ad-specific NAb titers were determined by luciferase-based virus neutralization assays as previously described [20]. Ad5, Ad26, and Ad35 NAb assays have been validated for use in clinical trials, whereas Ad48 NAb assays represent research assays. The vector characteristics, technical assay parameters, and assay performance and dynamic ranges were similar for the four vectors utilized in this study.

Table 2 depicts the raw seroepidemiology data for each population, and Figure 1 shows summary data. Ad5 seroprevalence was 87.9–89.5%, 90.5%, 86.4%, and 82.2% in adults in South Africa, Kenya, Uganda, and Thailand, respectively. In contrast, Ad26, Ad35, and Ad48 seroprevalence was significantly lower than Ad5 seroprevalence in all regions studied (p<0.0001 for each population, chi-square test). Ad26 seroprevalence was 43.1–53.2%, 66.2%, 67.8%, and 54.6% in adults in South Africa, Kenya, Uganda, and Thailand, respectively. Ad35 seroprevalence was 10.6–17.8%, 14.8%, 5.4%, and 17.1% in these populations, and Ad48 seroprevalence was 13.3–24.6%, 51.0%, 50.0%, and 12.8% in these cohorts, respectively.

Figure 1
Ad5, Ad26, Ad35, and Ad48 NAb Titers
Table 2
Neutralizing antibodies against Ad5, Ad26, Ad35 and Ad48

The majority of adults in these regions also exhibited high Ad5 NAb titers >200 (61.1–78.7%), and a substantial fraction had very high Ad5 NAb titers >1000 (25.1–46.8%). In contrast, markedly fewer individuals in these regions demonstrated high Ad26 NAb titers >200 (5.4–17.8%), Ad35 NAb titers >200 (0.4–5.0%), or Ad48 Nab titers >200 (0.9–10.8%), and only rare individuals had very high Ad26 NAb titers >1000 (0.0–3.8%), Ad35 NAb titers >1000 (0.0–1.7%), or Ad48 NAb titers >1000 (0.0–2.1%).

These data demonstrate that Ad35 seroprevalence and NAb titers were low in all regions studied. Ad48 seroprevalence was low in all regions except for East Africa. Ad26 seroprevalence was moderate in adults in the developing world, but Ad26 Nab titers proved consistently and substantially lower than Ad5 NAb titers in all regions studied. Moreover, Ad26 and Ad35 seroprevalence and NAb titers were negligible in infants (3–9 months old) and were low in children (6–12 years old) in South Africa, which represent potential ultimate target populations for an HIV-1 vaccine. NAb titers to each of these Ad serotypes appeared to be independent (data not shown).

Table 3 and Figure 2 depict the median and geometric mean Ad NAb titers in each geographic region. Median titers with interquartile ranges (IQRs) were calculated for each population, and geometric mean titers were determined among seropositives in each population. These data confirm that Ad26, Ad35, and Ad48 NAb titers were significantly lower than Ad5 NAb titers in all regions studied (p<0.001 for each region, ANOVA with post-hoc Dunnett correction). In particular, median (IQR) Ad5 NAb titers were 391 (94–1,069), 877 (259–2,369), 505 (77–1,501), and 521 (88–1,328), and geometric mean Ad5 NAb titers were 461, 967, 598, and 669 in adults in South Africa, Kenya, Uganda, and Thailand, respectively. In contrast, median (IQR) Ad26 NAb titers were 22 (18–65), 37 (18–115), 38 (18–132), and 30 (18–134), and geometric mean Ad26 NAb titers were 67, 97, 86, and 127 in these populations, respectively. Geometric mean Ad35 NAb titers were 75, 103, 50, and 86 in these regions, and geometric mean Ad48 NAb titers were 38, 90, 87, and 54 in these populations, respectively. These data show that median Ad26, Ad35, and Ad48 NAb titers were approximately 10-fold lower than median Ad5 NAb titers in these regions in the developing world. Median Ad26, Ad35 and Ad48 NAb titers were also substantially lower than median Ad5 NAb titers in the developed world.

Figure 2Figure 2Figure 2
Median and Geometric Mean Ad NAb Titers
Table 3
Descriptive statistics on neutralizing antibodies against Ad5, Ad26, Ad35 and Ad48

3.2 Impact of Low Ad26 NAb Titers on rAd26 Vaccine Immunogenicity in Rhesus Monkeys

Given the moderately common but low titer Ad26 NAbs in the developing world, we evaluated the potential impact of such Ad26 NAb titers on rAd26 vaccine vector immunogenicity in a pilot study in nonhuman primates. 12 adult rhesus monkeys were pre-immunized twice by the intranasal route with either 1011 vp replication-competent rAd26-Empty (N=6) or saline (N=6). Animals that received rAd26-Empty developed serum Ad26 NAb titers that were similar in magnitude to those found in humans in sub-Saharan Africa and Southeast Asia (median titer 102; geometric mean titer 85). Four weeks later, all animals were immunized once by the intramuscular route with 4×1010 vp of the rAd26-Gag/Pol/Env/Nef vaccine.

Env-specific antibody responses were assessed by ELISA, and Gag-, Pol-, Env-, and Nef-specific cellular immune responses were assessed by interferon-γELISPOT assays at week 2 following vaccination. As shown in Figure 3A, Env-specific binding antibody responses were comparable in monkeys with and without low baseline Ad26 NAb titers (p=NS, Mann-Whitney test). Similarly, as shown in Figure 3B, Gag-, Pol-, Env-, and Nef-specific cellular immune responses were comparable in monkeys with and without low baseline Ad26 NAb titers (p=NS, Mann-Whitney test). Humoral and cellular immune responses also proved comparable between groups at week 4 and week 12 following vaccination (data not shown).

Figure 3
Immunogenicity of rAd26-Gag/Pol/Env/Nef in Rhesus Monkeys

We also evaluated Ad26 NAb titers in these monkeys both before and after rAd26 vaccination. As shown in Figure 3C, Ad26 NAb titers proved high (range 1,157 – 16,384) in the Ad26 pre-immunized monkeys following rAd26-Gag/Pol/Env/Nef vaccination. These titers were significantly higher than those observed in naïve monkeys following vaccination (p=0.002, Mann-Whitney test) but did not detectably suppress rAd26-elicited humoral or cellular immune responses and did not alter the ratio of CD4/CD8 cellular immune responses (Figure 3D). These data demonstrate the dynamic range of the Ad26 NAb assay and confirm that the relatively low Ad26 NAb titers observed in human populations (Figures 12) did not simply reflect technical characteristics of the Ad26 NAb assay.

4. Discussion

Our data show that Ad26, Ad35, and Ad48 NAb titers were substantially lower than Ad5 NAb titers in multiple large international human populations (N=4,381). Ad35 NAbs were rare in all populations studied, and Ad48 NAbs were uncommon in all cohorts except for East Africa. In contrast, Ad26 NAbs were moderately common in sub-Saharan Africa and Southeast Asia, but Ad26 NAb titers proved markedly lower than Ad5 NAb titers in all regions studied. To the best of our knowledge, this is the largest and most comprehensive international Ad seroepidemiology study to date.

Our findings confirm and extend prior studies showing that Ad35 seroprevalence and NAb titers were low in the developing world [6, 17]. Ad26 seroprevalence in adults in the present study proved somewhat higher than our initial estimate from a different population [8] and comparable with more recent estimates [16, 1819]. However, this study shows that Ad26 NAb titers were markedly and consistently lower than Ad5 Nab titers across all geographic populations, risk groups, and age ranges studied. Our data also address several discrepancies between two recent studies that have compared Ad5 and Ad26 NAb titers in the developing world [16, 19]. For example, we find substantially higher magnitude Ad5 NAb titers in sub-Saharan Africa and higher Ad26 seroprevalence in Southeast Asia than reported by one of these groups [19]. We suspect that these differences may relate to differences in the sample sizes and specific assays utilized. Additional novel features of the present study that are not addressed in these previous reports [16, 19] include direct comparisons of Ad NAb titers in infants, schoolchildren, and adults in South Africa; NAb titers to Ad35 and Ad48 in addition to Ad5 and Ad26; and Ad NAb titers in low and high HIV-1 risk cohorts in both the United States and Africa.

The low Ad26 and Ad35 seroprevalence and NAb titers in pediatric populations suggest that seroconversion to these Ads is an age-dependent process. We previously reported that Ad5 seroconversion was also age-dependent, with maternal antibodies disappearing by approximately 6 months of age and high Ad5 NAb titers often developing by age 2 in sub-Saharan Africa [15]. Our current data from infants, schoolchildren, and adults from South Africa suggest that Ad26 and Ad35 NAbs increase at a slower rate than do Ad5 NAbs, likely reflecting a lower rate of natural infection with these serotypes. These data may be relevant for HIV-1 vaccine development, since infants and pre-adolescents represent potential ultimate vaccine target populations.

We also observed that relatively low baseline Ad26 NAb titers did not detectably suppress humoral or cellular immune responses elicited by a rAd26-Gag/Pol/Env/Nef vaccine in rhesus monkeys. Baseline Ad26-specific T cell responses were also induced in these animals (data not shown), although the impact of Ad-specific cellular immune responses on rAd vector immunogenicity remains unclear. Although these data were derived utilizing a preclinical model, analogous rhesus monkey immunogenicity studies [12] accurately predicted the suppressive impact of high titers but not low titers of Ad5 NAbs on rAd5 vaccine immunogenicity in humans [35, 21], thus suggesting the potential clinical relevance of these data. However, the nonhuman primate model also has several important differences from the natural human setting, including a substantially shorter time interval between pre-exposure and vaccination and poor replication of the rAd26-Empty vector in this host species. Clinical trials are therefore required to determine the utility of these rAd vectors as vaccine candidates in humans.

The phase 2b Step study (HVTN 502/Merck 023) demonstrated that rAd5 vectors expressing clade B HIV-1 Gag/Pol/Nef antigens failed to afford protection in high-risk individuals in North and South America, the Caribbean, and Australia and may have increased HIV-1 acquisition risk in certain subgroups [3]. Alternative human serotype rAd vectors, such as rAd26 and rAd35, are therefore being explored as alternatives and are currently in phase 1 clinical trials. Our studies show that baseline NAb titers to these serotypes are substantially lower than baseline Ad5 NAb titers in multiple populations worldwide. These data suggest that further clinical studies are warranted with these and other rAd vectors in both the developing world and in the developed world.

ACKNOWLEDGEMENTS

We thank A. Riggs, D. Lynch, A. La Porte, N. Simmons, J. Iampietro, F. Stephens, D. Casimiro, M. Robertson, J. Shiver, J. McElrath, B. Farrah, J. Cox, P. Fast, H. Jaspan, G. Stevens, P. Chetty, T. Tarragona, L. Clark, A. Raxworthy-Cooper, M. Price, E. Cormier, K. Higgins, and M. Pensiero for generous advice, assistance, and reagents. We acknowledge support from the U.S. National Institutes of Health (AI066305, AI066924, AI078526), the Bill & Melinda Gates Foundation (#38614), the Elizabeth Glaser Pediatric AIDS Foundation, and the Ragon Institute of MGH, MIT, and Harvard. The IAVI samples were collected with funding from the U.S. Agency for International Development (GPO-A-00-06-00006-00). The MIRA and PIP samples were collected with funding from the Bill & Melinda Gates Foundation.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

The authors declare no competing financial interests. S.V.K., G.J.W., M.G.P., and J.G. are employees of Crucell. Crucell participated in the analysis of data but had no role in the study design, data generation, or funding of this study.

REFERENCES

1. Casimiro DR, Chen L, Fu TM, Evans RK, Caulfield MJ, Davies ME, et al. Comparative immunogenicity in rhesus monkeys of DNA plasmid, recombinant vaccinia virus, and replication-defective adenovirus vectors expressing a human immunodeficiency virus type 1 gag gene. J Virol. 2003 Jun;77(11):6305–6313. [PMC free article] [PubMed]
2. Roberts DM, Nanda A, Havenga MJ, Abbink P, Lynch DM, Ewald BA, et al. Hexon-chimaeric adenovirus serotype 5 vectors circumvent pre-existing anti-vector immunity. Nature. 2006 May 11;441(7090):239–243. [PubMed]
3. Buchbinder SP, Mehrotra DV, Duerr A, Fitzgerald DW, Mogg R, Li D, et al. Efficacy assessment of a cell-mediated immunity HIV-1 vaccine (the Step Study): a double-blind, randomised, placebo-controlled, test-of-concept trial. Lancet. 2008 Nov 29;372(9653):1881–1893. [PMC free article] [PubMed]
4. Priddy FH, Brown D, Kublin J, Monahan K, Wright DP, Lalezari J, et al. Safety and immunogenicity of a replication-incompetent adenovirus type 5 HIV-1 clade B gag/pol/nef vaccine in healthy adults. Clinical infectious diseases. 2008 Jun 1;46(11):1769–1781. [PubMed]
5. Catanzaro AT, Koup RA, Roederer M, Bailer RT, Enama ME, Moodie Z, et al. Phase 1 safety and immunogenicity evaluation of a multiclade HIV-1 candidate vaccine delivered by a replication-defective recombinant adenovirus vector. J Infect Dis. 2006 Dec 15;194(12):1638–1649. [PMC free article] [PubMed]
6. Vogels R, Zuijdgeest D, van Rijnsoever R, Hartkoorn E, Damen I, de Bethune MP, et al. Replication-deficient human adenovirus type 35 vectors for gene transfer and vaccination: efficient human cell infection and bypass of preexisting adenovirus immunity. J Virol. 2003 Aug;77(15):8263–8271. [PMC free article] [PubMed]
7. Barouch DH, Pau MG, Custers JH, Koudstaal W, Kostense S, Havenga MJ, et al. Immunogenicity of recombinant adenovirus serotype 35 vaccine in the presence of pre-existing anti-Ad5 immunity. J Immunol. 2004 May 15;172(10):6290–6297. [PubMed]
8. Abbink P, Lemckert AA, Ewald BA, Lynch DM, Denholtz M, Smits S, et al. Comparative seroprevalence and immunogenicity of six rare serotype recombinant adenovirus vaccine vectors from subgroups B and D. J Virol. 2007 May;81(9):4654–4663. [PMC free article] [PubMed]
9. Farina SF, Gao GP, Xiang ZQ, Rux JJ, Burnett RM, Alvira MR, et al. Replication-defective vector based on a chimpanzee adenovirus. J Virol. 2001 Dec;75(23):11603–11613. [PMC free article] [PubMed]
10. Fitzgerald JC, Gao GP, Reyes-Sandoval A, Pavlakis GN, Xiang ZQ, Wlazlo AP, et al. A simian replication-defective adenoviral recombinant vaccine to HIV-1 gag. J Immunol. 2003 Feb 1;170(3):1416–1422. [PubMed]
11. Waddington SN, McVey JH, Bhella D, Parker AL, Barker K, Atoda H, et al. Adenovirus serotype 5 hexon mediates liver gene transfer. Cell. 2008 Feb 8;132(3):397–409. [PubMed]
12. Lore K, Adams WC, Havenga MJ, Precopio ML, Holterman L, Goudsmit J, et al. Myeloid and plasmacytoid dendritic cells are susceptible to recombinant adenovirus vectors and stimulate polyfunctional memory T cell responses. J Immunol. 2007 Aug 1;179(3):1721–1729. [PMC free article] [PubMed]
13. Liu J, Ewald BA, Lynch DM, Denholtz M, Abbink P, Lemckert AA, et al. Magnitude and phenotype of cellular immune responses elicited by recombinant adenovirus vectors and heterologous prime-boost regimens in rhesus monkeys. J Virol. 2008 May;82(10):4844–4852. [PMC free article] [PubMed]
14. Liu J, O'Brien KL, Lynch DM, Simmons NL, La Porte A, Riggs AM, et al. Immune control of an SIV challenge by a T-cell-based vaccine in rhesus monkeys. Nature. 2009 Jan 1;457(7225):87–91. [PMC free article] [PubMed]
15. Thorner AR, Vogels R, Kaspers J, Weverling GJ, Holterman L, Lemckert AA, et al. Age dependence of adenovirus-specific neutralizing antibody titers in individuals from sub-Saharan Africa. J Clin Microbiol. 2006 Oct;44(10):3781–3783. [PMC free article] [PubMed]
16. Mast TC, Kierstead L, Gupta SB, Nikas AA, Kallas EG, Novitsky V, et al. International epidemiology of human pre-existing adenovirus (Ad) type-5, type-6, type-26 and type-36 neutralizing antibodies: correlates of high Ad5 titers and implications for potential HIV vaccine trials. Vaccine. 2010 Jan 22;28(4):950–957. [PubMed]
17. Kostense S, Koudstaal W, Sprangers M, Weverling GJ, Penders G, Helmus N, et al. Adenovirus types 5 and 35 seroprevalence in AIDS risk groups supports type 35 as a vaccine vector. Aids. 2004 May 21;18(8):1213–1216. [PubMed]
18. Curlin ME, Huang ML, Lu X, Celum CL, Sanchez J, Selke S, et al. Frequent detection of human adenovirus from the lower gastrointestinal tract in men who have sex with men. PLoS ONE. 2010;5(6):e11321. [PMC free article] [PubMed]
19. Chen H, Xiang ZQ, Li Y, Kurupati RK, Jia B, Bian A, et al. Adenovirus-based vaccines: comparison of vectors from three species of adenoviridae. J Virol. 2010 Oct;84(20):10522–10532. [PMC free article] [PubMed]
20. Sprangers MC, Lakhai W, Koudstaal W, Verhoeven M, Koel BF, Vogels R, et al. Quantifying adenovirus-neutralizing antibodies by luciferase transgene detection: addressing preexisting immunity to vaccine and gene therapy vectors. J Clin Microbiol. 2003 Nov;41(11):5046–5052. [PMC free article] [PubMed]
21. McElrath MJ, De Rosa SC, Moodie Z, Dubey S, Kierstead L, Janes H, et al. HIV-1 vaccine-induced immunity in the test-of-concept Step Study: a case-cohort analysis. Lancet. 2008 Nov 29;372(9653):1894–1905. [PMC free article] [PubMed]
22. Pitisuttithum P, Gilbert P, Gurwith M, Heyward W, Martin M, van Griensven F, et al. Randomized, double-blind, placebo-controlled efficacy trial of a bivalent recombinant glycoprotein 120 HIV-1 vaccine among injection drug users in Bangkok, Thailand. The Journal of infectious diseases. 2006 Dec 15;194(12):1661–1671. [PubMed]
23. Flynn NM, Forthal DN, Harro CD, Judson FN, Mayer KH, Para MF. Placebo-controlled phase 3 trial of a recombinant glycoprotein 120 vaccine to prevent HIV-1 infection. The Journal of infectious diseases. 2005 Mar 1;191(5):654–665. [PubMed]