Skip to main content
  • Loading metrics

Early Transcriptional Signatures of the Immune Response to a Live Attenuated Tetravalent Dengue Vaccine Candidate in Non-human Primates

  • Fiona R. Strouts,

    Affiliation Department of Microbiology and Immunology, Stanford University, Stanford, California, United States of America

  • Stephen J. Popper,

    Affiliation Department of Microbiology and Immunology, Stanford University, Stanford, California, United States of America

  • Charalambos D. Partidos,

    Current address: School of Veterinary Medicine, Department of Pathobiological Sciences, University of Wisconsin-Madison, Madison, Wisconsin, United States of America

    Affiliation Takeda Vaccines, Inc., Deerfield, Illinois, United States of America

  • Dan T. Stinchcomb,

    Affiliation Takeda Vaccines, Inc., Deerfield, Illinois, United States of America

  • Jorge E. Osorio,

    Current address: School of Veterinary Medicine, Department of Pathobiological Sciences, University of Wisconsin-Madison, Madison, Wisconsin, United States of America

    Affiliation Takeda Vaccines, Inc., Deerfield, Illinois, United States of America

  • David A. Relman

    Affiliations Department of Microbiology and Immunology, Stanford University, Stanford, California, United States of America, Department of Medicine, Stanford University, Stanford, California, United States of America, Veterans Affairs Palo Alto Health Care System, Palo Alto, California, United States of America



The development of a vaccine against dengue faces unique challenges, including the complexity of the immune responses to the four antigenically distinct serotypes. Genome-wide transcriptional profiling provides insight into the pathways and molecular features that underlie responses to immune system stimulation, and may facilitate predictions of immune protection.

Methodology/Principal Findings

In this study, we measured early transcriptional responses in the peripheral blood of cynomolgus macaques following vaccination with a live, attenuated tetravalent dengue vaccine candidate, TDV, which is based on a DENV-2 backbone. Different doses and routes of vaccine administration were used, and viral load and neutralizing antibody titers were measured at different time-points following vaccination. All 30 vaccinated animals developed a neutralizing antibody response to each of the four dengue serotypes, and only 3 of these animals had detectable serum viral RNA after challenge with wild-type dengue virus (DENV), suggesting protection of vaccinated animals to DENV infection. The vaccine induced statistically significant changes in 595 gene transcripts on days 1, 3, 5 and 7 as compared with baseline and placebo-treated animals. Genes involved in the type I interferon (IFN) response, including IFI44, DDX58, MX1 and OASL, exhibited the highest fold-change in transcript abundance, and this response was strongest following double dose and subcutaneous (versus intradermal) vaccine administration. In addition, modules of genes involved in antigen presentation, dendritic cell activation, and T cell activation and signaling were enriched following vaccination. Increased abundance of gene transcripts related to T cell activation on day 5, and the type I IFN response on day 7, were significantly correlated with the development of high neutralizing antibody titers on day 30.


These results suggest that early transcriptional responses may be predictive of development of adaptive immunity to TDV vaccination in cynomolgus macaques, and will inform studies of human responses to dengue vaccines.

Author Summary

Dengue has become the leading cause of mosquito-borne virus infections worldwide. Despite considerable effort, development of a successful vaccine against dengue virus (DENV) has been challenging due to the co-circulation of the four DENV serotypes in endemic areas—to which humans develop distinct immune responses, and the increased risk of severe disease in those with pre-existing immunity to one serotype when they are infected with a different serotype. In this study, we investigated the responses in macaques to vaccination with the tetravalent, live-attenuated vaccine, TDV, by different doses and routes of vaccine administration. We identify changes in macaque gene expression that occurred in the days immediately following vaccination with TDV, a time-period that is difficult to study during natural infection. The gene expression response was characterized by features of the innate immune response to virus, notably the type I interferon response, and began the day after TDV vaccination. This response correlated with the development of neutralizing antibodies, which means that it might serve as an early indicator of a subsequent protective immune response to dengue vaccines.


Over the last 50 years, the incidence of dengue has increased 30-fold, and now more than half of the world’s population is at risk of dengue virus (DENV) infection [1]. Transmitted by Aedes mosquitos, DENV has become the leading cause of mosquito-borne viral infections worldwide, with an estimated 390 million infections occurring each year [1]. The outcome of infection ranges from an asymptomatic state to classic dengue fever (DF) and severe and potentially life-threatening dengue hemorrhagic fever (DHF) and dengue shock syndrome (DSS).

Each of the four antigenically distinct serotypes of dengue virus (DENV1 –DENV4) is capable of causing severe disease. While infection with one serotype provides long-lasting protection against re-infection with that serotype, cross-protective immunity is temporary and lasts only several months [2]. Furthermore, secondary infection with a heterologous serotype greatly increases the risk of developing severe disease [3,4]. While there is currently no licensed vaccine against dengue, there are several dengue vaccine candidates in development [5]. Takeda Vaccines’ Tetravalent Dengue Vaccine Candidate (TDV) (formerly DENVax, Inviragen) consists of a live attenuated DENV-2 strain (TDV-2) and three chimeric viruses containing the prM and E protein genes of DENV-1, -3 and -4 expressed in the context of the TDV-2 genome backbone (TDV-1, TDV-3, and TDV-4, respectively) [6,7]. TDV has been shown to be immunogenic and efficacious in animal models [810], generally well tolerated in humans [11], and is currently in phase 2 clinical trials.

Studies of dengue infection have revealed unique transcriptional signatures during the acute phase of infection that are associated with subsequent disease severity [1216]. A recent study examined the role of the innate immune response in modulating the humoral immune response [16]. Understanding the mechanisms underlying the development of protection against dengue, and responses to dengue vaccination, may be useful in the further development of an effective vaccine against dengue. Host genome-wide transcriptional profiling provides a means to identify changes in gene expression occurring immediately following vaccination that may play a role in the development of protective immunity. This approach has revealed useful, clinically-relevant signatures following immunization with a variety of different vaccines [1722].

In this study, we used transcriptional profiling to characterize early changes in gene expression in peripheral blood cells of cynomolgus macaques following vaccination with TDV, involving different doses (single dose or double dose on day 0), and routes (subcutaneous or intradermal) of vaccine administration. We compared changes in transcript abundance following TDV vaccination with those following infection with wild-type (wt) DENV in cynomolgus macaques. Gene transcript abundances were correlated with measurements of viral load and neutralizing antibody titer to identify markers predictive of vaccine immunogenicity.


Vaccines and viruses

TDV was generated from cDNA clone-derived DENV-2 VV45R virus, based on the DENV-2 PDK-53 genome [10], and from three chimeric strains based on DENV-2 PDK-53 expressing the prM and E genes of wt DENV-1, DENV-3 or DENV-4, as described previously [10,8]. For the neutralization antibody assays, we used the wt viruses from which the prM and E genes for each TDV virus were derived [9,11,23]. DENV-2 New Guinea C and DENV-4 Dominica/81 were used for viral challenge, generously provided by Dr. Stephen Whitehead (US National Institutes of Health, Bethesda, MD).

Ethics statement

Animal work was conducted at the Charmany Instructional Facility of the School of Veterinary Medicine, University of Wisconsin-Madison, and at the Wisconsin National Primate Research Center. All animal procedures were approved by the University of Wisconsin-Madison’s Graduate School Animal Care and Use Committee, and the protocol, #G00634, in which these procedures are described, was approved on 10/22/2013. The regulations/guidelines to which animal care and the animal use protocol adhered are “The Guide for the Care and Use Of Laboratory Animals, 8th Edition”; the United States Department of Agriculture (USDA) Animal Welfare Act and Animal Welfare Regulations; US Public Health Service Policy on Humane Care and Use of Laboratory Animals; US Government Principles for the Utilization and Care of Vertebrate Animals Used in Testing, Research, and Training; and the USDA Policy Manual.

Monkeys were singly housed in standard stainless steel primate cages (Suburban Surgical, Chicago IL). All animals had visual and auditory contact with each other in the same room. They were fed twice daily with commercial chow (Harlan Teklad #2050, 20% protein Primate Diet, Madison, WI) and given a variety of fruit in the afternoons. In addition, we provided foraging activities and physical environmental enrichment at least weekly for both activities. Housing rooms were maintained at 65–75°F, 30–70% humidity and on a 12:12 light–dark cycle (ON: 0600, OFF: 1800). Standard veterinary analgesia was available to animals if necessary with either NSAIDs or opioid derivative drugs such as buprenorphine.

Animals and study design

Thirty-five adult male, DENV-seronegative cynomolgus macaques from Vietnam were placed in quarantine for 30 days prior to the start of the study. Five groups of animals (n = 6 animals per group) received TDV either subcutaneously (SC) in 0.5 mL inocula or intradermally (ID) in 0.1 mL inocula using a needle-free injector (PharmaJet device) or needle and syringe (N&S), as outlined in Table 1. Detailed experimental methods appear in S1 Text.

Serum viral RNA

Viral RNA in serum samples was measured using a quantitative reverse transcription-polymerase chain reaction (qRT-PCR) as described previously [9,23]. The limit of detection for the qRT-PCR of 3.6 log10 copies/mL was determined for each viral RNA standard by creating a standard curve consisting of nine replicates per dilution.

Microneutralization assay

Heat-inactivated serum samples (56°C for 30 min) were tested for neutralizing activity using a viral immunofocus reduction microneutralization assay with an ELISpot reader (AID, San Diego, CA), as previously described [23]. Fifty percent of the average number of foci in the negative control serum defined the cut-off point (PRNT50). The serum dilution closest to the cut-off was recorded as the reciprocal neutralizing titer.

Genome-wide transcript abundance analysis of peripheral blood

Total RNA was extracted using the PAXGene Blood RNA kits (Qiagen, Valencia, CA), and 500 ng of total RNA was amplified using the TargetAmp Aminoallyl aRNA amplification kit (Epicentre, Madison, CA). 8 ug of amplified RNA and 5 ug Universal Human Reference aRNA (Stratagene, La Jolla, CA) were labeled using Cy5 and Cy3 dyes, respectively, and hybridized to Human Exonic Evidence Based Oligonucleotide (HEEBO) microarrays. Details of the protocol have been described previously [15]. HEEBO microarrays, containing 44,544 probes, were printed by the Stanford Functional Genomics Core Facility. A detailed description of this probe set can be found at ( Microarray data were submitted to the Princeton University MicroArray database for subsequent analyses. Previous studies have shown that HEEBO arrays can accurately measure the abundance of gene transcripts in nonhuman primates [2426]. Data were normalized by local background subtraction and a global mean normalization using regression correlation. Data were filtered to exclude probes that did not demonstrate a regression correlation of ≥0.6 between Cy5 and Cy3 signal over the pixels comprising the array element, and intensity/background ratio >2 in at least one channel. The microarray data are available at Gene Expression Omnibus (GEO accession number GSE72430).

Statistical analysis

Changes in transcript abundance were calculated by subtracting the log2 abundance value at baseline (average of days -11 and -2) from each subsequent time-point (days 1, 3, 5, 7, etc.) for each animal. Baseline-transformed data were used for comparison between the different groups. Significance Analysis of Microarrays (SAM) was used to identify significantly differentially expressed genes between predetermined groups; genes and their transcripts were considered significant at a false-discovery rate (FDR) <5% and fold-change ≥1.3 [27]. The fold-change reported is the difference in the average relative abundance between each of the two groups in the comparison. Gene enrichment analysis was performed using the Database for Annotation, Visualization and Integrated Discovery (DAVID) [28]. DAVID keywords included Gene Ontology (GO) terms, biological pathways from the Kyoto Encyclopedia of Genes and Genomes (KEGG), and Swiss-Prot keywords. Keywords were considered significant when p<0.05 after Benjamini-Hochberg correction for multiple testing. Pathway analysis was performed using Gene Set Enrichment Analysis (Broad Institute) [29] and Blood Transcription Modules (BTMs) [17] on genes pre-ranked by SAM score or Spearman correlation coefficient. BTMs were considered significant at a FDR<5%. Differences in median transcript abundance levels of specific genes between immunization groups were tested using the Mann-Whitney U test. Spearman correlation was used to correlate relative transcript abundance of 15,705 variable genes (filtered for probes with an intensity/background >2 and present in at least 80% of the samples) with peak viral RNA abundance and duration of detectable viral RNA after vaccination, and DENV-2 specific neutralizing antibody titer and median neutralizing antibody titer (median PRNT50 of the 4 serotypes) on day 30.


TDV-induced early changes in peripheral blood transcript abundance

To characterize the overall response to TDV, we compared changes in transcript abundance by day in all 30 vaccinated animals against their pre-vaccinated baseline values. Five hundred and ninety-five genes were significantly differentially expressed (DE) in vaccinated animals compared with their own baseline on at least one day during the first week following vaccination with TDV (FDR<5%, fold-change≥1.3). Forty-six additional genes were differentially expressed in an unpaired SAM comparison between animals vaccinated with TDV and those that received the placebo (n = 5); these genes were included in this list (S1 and S2 Tables). A more stringent fold-change cut-off of 1.5 resulted in 217 DE genes (Fig 1A), and so we decided to use the 1.3-fold change cut-off in order to include a greater number of genes in the functional analysis. Functional annotation using DAVID identified GO terms associated with these gene sets that increased and decreased in abundance by day (S1 Table). Genes associated with the GO term “antiviral defense” (including DDX58, EIF2AK2, ISG15, IRF7, MX1, IFI44, STAT1) increased in abundance on day 1, but the association with this GO term was strongest on day 5 along with additional terms related to the response to vaccination (Fig 1B). Genes associated with cell cycle and transcription (including CDC73, CCAR1, CNOT7, CREBZF, ZEB1, and FOS) decreased in abundance on days 1 and 3. No GO terms were significantly enriched on day 7.

Fig 1. Changes in gene transcript abundance following TDV vaccination.

Results from SAM analysis of transcripts from vaccinated animals compared with baseline, and vaccinated animals compared with placebo animals. (A) Numbers of differentially-expressed (DE) genes at each time-point. FDR<0.05; fold-change≥1.3 solid areas, fold-change≥1.5 dashed areas. Red bars indicate increased abundance; blue bars indicate decreased abundance. (B) Top GO terms associated with DE genes by day, p<0.05 after Benjamani correction for multiple testing. Strength of the association is shown on the abscissa. Red bars represent positively enriched GO terms; blue bars represent negatively enriched GO terms.

Transcripts for genes involved in viral recognition (DDX58 (RIG-I) and EIF2AK2 (PKR)), the type I interferon (IFN) response (IRF7, OAS2 and OASL), the antiviral response (MX1), regulation of cytokine signaling (STAT1), and apoptosis (XAF1) exhibited the highest fold-changes in abundance compared with baseline (maximum fold-change 2.4) (S1 Table). The IFN stimulated gene (ISG) ISG15 was up-regulated following TDV vaccination and is typically induced in response to type I IFN [30]. Secreted ISG15 acts on T and natural killer (NK) lymphocytes, in which it induces IFN-γ production, and has been shown to inhibit viral replication and suppress particle release of DENV-2 [30]. Transcripts for TNFSF13B, a gene associated with B cell activation [31] also increased in abundance following TDV vaccination.

To complement the gene-level analysis, we applied Gene Set Enrichment Analysis (GSEA) using the 334 Blood Transcription Modules (BTMs) constructed from publicly available microarray data specific to human blood [17], supplemented with 4 additional modules comprising cytokine-induced gene sets identified in our previous work [32] (S3 Table). Blood transcription modules also include sets of cell-type specific genes, which indicate the cell types that may contribute to the response based on the most differentially expressed genes [17]. Following vaccination, 86 BTMs were enriched for genes that increased in abundance (positively enriched), and 34 BTMs were enriched for genes that decreased in abundance (negatively enriched) (FDR<5%) (Fig 2, S4 Table). Positively enriched modules included those with functions associated with the innate immune response to viruses, antigen presentation, and activation of T cells, dendritic cells (DCs) and platelets. There was significant enrichment of genes specific to T cells, NK cells, neutrophils, monocytes and DCs (Fig 2, S4 Table). Negatively enriched modules were associated with mitosis, other aspects of the cell cycle and cell division, and inositol signaling. Inositol signaling has been shown to control the amplitude of type I IFN secretion and pDC activation, and usually inhibits cell activation [33].

Fig 2. BTMs enriched over time following TDV administration across all animals.

Modules are grouped (black vertical bars) by function and ordered by significance; only modules significantly enriched on at least one day are shown (FDR<5%). Positively enriched BTMs are shown in red, negatively enriched BTMs in blue, and no significant enrichment is shown in grey. NES, Normalized Enrichment Score [17].

Effect of vaccination route and dose on the early transcriptional response

The transcriptional response in animals that received TDV by ID route (Groups 1 and 2) differed from those that received TDV via SC route (Groups 4 and 5) primarily by the temporal pattern of transcript abundance. Following ID immunization, most transcripts increased in abundance on day 1 (n = 142) and then decreased by day through day 7 (n = 0). On day 1 there was an increase in abundance of genes involved in lymphocyte activation (THY1/CD90, CORO1A, LST1, STXBP2, TNFSF13 and TNSF13B), and viral recognition and the antiviral response (DDX58, EIF2AK2, ISG15, IRF7, IRF9, IFI44, STAT1, OASL, OAS2, IFITM3 and MX1), which persisted on days 3 and 5. Following SC immunization, transcript abundance increased over time from day 1 (n = 18), 3 (n = 157), to 5 (n = 1460), and then decreased by day 7 (n = 17). On day 5 following SC immunization, there was an increase in abundance of genes related to T and B cell activation and differentiation (including BCL2, CD28, CD47, IL15, TNFSF13B and KLRK1), regulation of cytokine production, and the JAK/STAT signaling cascade (JAK2, IL6ST, STAT1, STAT2, STAT4, and SOC2S). On day 7, only genes involved in viral recognition and the antiviral response were enriched (DDX58, EIF2AK2, ISG15, IFI44, STAT1, OAS2, OASL, XAF1 and MX1). No significant differences in gene enrichment were observed by direct comparisons of relative transcript abundance levels between animals that received ID vaccination versus SC vaccination.

In all groups, the highest fold-change in gene abundance following vaccination was observed for genes involved in the antiviral and type I IFN response. To investigate how the observed responses were modulated by dose and route of vaccine administration we created a master list of 379 genes involved in the antiviral/type I IFN response by combining the genes present in relevant BTMs. Overall, 38 of these genes were significantly differentially expressed following TDV vaccination (Fig 3A). A greater number of these genes were significantly differentially expressed in the groups of animals that received SC immunization (single or double dose) compared with those that received ID immunization (S1 Fig), suggesting that SC immunization results in a broader antiviral/type I IFN response than ID immunization. To evaluate whether dose or route of vaccine administration altered the strength of the antiviral/type I IFN response, we compared expression of each gene by vaccination group over time (Mann-Whitney U test, p<0.05). Abundances for 11 gene transcripts differed significantly between SC and ID vaccination (single and double dose groups combined); the majority had higher levels following SC vaccination (Fig 3A). For dose, we focused on the SC double (SCd) and single (SCs) dose groups (groups 4 and 5) since a greater number of changes in transcript abundance were observed in these groups compared with the ID groups (groups 1–3), and because SC administration is a route used for human trials of TDV. Following double dose SC vaccination, 20 genes had significantly higher transcript abundances during at least one time-point compared with single dose SC vaccination (Fig 3A and 3B). Most of these differences occurred on day 7 (15 genes), suggesting that increasing the dose of TDV elicits a stronger and more prolonged response for these genes.

Fig 3. Antiviral/type I IFN response following TDV vaccination.

A) Median abundances of 38 transcripts over time (days post-vaccination) in all vaccinated animals, and by group. Red indicates an increase in transcript abundance; blue indicates decreased transcript abundance (FDR<0.05, fold-change≥1.3). Genes listed when significant in comparison between intradermal (ID) and subcutaneous (SC) groups, and SC double dose (SCd) and SC single dose (SCs) groups (p<0.05, ranksum). Placebo group, and placebo animals challenged with wt DENV, shown for comparison. B) Median relative transcript abundance over time by vaccination group. Bars indicate standard error.

Transcriptional response to wild-type DENV is most similar to the response to double dose SC TDV

Wild-type (wt) challenge of placebo animals with DENV-2 (3 animals) or DENV-4 (1 animal) by SC inoculation led to significant changes in the abundance of 135 genes over time (Fig 4A). The smaller number of significant genes may reflect the smaller sample size of 4 animals. Unsupervised hierarchical clustering of these genes revealed 3 clusters (Fig 4B, S5 Table). Clusters 1 and 2 were both significantly enriched for genes we previously identified as induced by type I IFN (p<1E-10 for each cluster) [32], though only genes from cluster 1 were significantly associated with GO terms, including “response to virus” (p = 1.2E-8), “immune response” (p = 0.0001), and “GTP-binding” (p = 0.04). Cluster 3 was not associated with any GO terms. While abundance of transcripts in clusters 1 and 2 increased following both infection with wt DENV and vaccination, the magnitude of response was much greater following wt virus infection, as seen in median fold-change over time of transcripts for genes in each cluster (Fig 4B). Fold-change peaked at 19.9 (ISG15) following wt DENV infection compared with 5.7 (IFI44) following double dose SC TDV vaccination. Wt virus caused a decrease in transcript abundance for genes in Cluster 3. Proteins expressed by genes in this cluster include IL1-R2, a negative regulator of IL-1 signaling [34], and CRISPLD2, a serum protein produced by monocytes, NK cells, and T cells in response to stimulation by LPS and other PAMPs, including poly (I:C) [35,36].

Fig 4. Comparison of changes in transcript abundance following wt DENV infection and TDV immunization.

A) Hierarchical clustering of 135 genes whose transcripts differ in abundance over time following wt infection of sham-immunized animals (FDR<0.05, fold-change≥1.3). Median fold-change in transcript abundance is shown over time for each group; statistical significance (FDR<5% and fold-change ≥1.3) is indicated in right-hand grey column. Red indicates increased abundance; blue indicates decreased abundance. Black vertical bars delineate gene clusters 1, 2 and 3. B) Median fold-change of transcript abundance for each gene cluster by group. wt DENV, wild-type dengue virus; SCd, subcutaneous double dose; SCs, subcutaneous single dose.

TDV-2 viral load correlates with DENV-2 neutralizing antibody titer and features of the early response

Over half of the animals (17/30) had detectable viral RNA (vRNA) (range: 3.8–5.6 log10 copies/mL) between days 5 and 14 following primary immunization (mean duration, 5 days) (Table 2). Only TDV-2 virus RNA was detected in all cases, consistent with previous studies of TDV immunization [9,11]. Most animals that received TDV by SC (single and double dose) and ID (double dose only) routes had vRNA, compared with only 1 animal that received TDV by single dose ID route (Table 2). After challenge with wt DENV, only 3 of the vaccinated animals developed detectable vRNA, compared with all placebo (sham-immunized) animals (range: 3.7–5.8 log10 copies/mL; mean duration, 4 days), suggesting that TDV protected against wt DENV challenge (Table 2).

Table 2. Viral load (range) after TDV vaccination and DENV-2/DENV-4 challenge.

Correlation of transcriptional responses (relative transcript abundance by day) to the duration and peak of viral load revealed significant correlations with genes involved in B and T cell activation and differentiation on day 5 post-immunization. Genes included IRF4 (Spearman’s ρ = 0.70, p = 1.76E-05) and ITK (Spearman’s ρ = 0.58, p = 0.0009), which encode intracellular tyrosine kinases involved in the regulation of T cell development and differentiation, and CCR6 (Spearman’s ρ = 0.52, p = 0.006), which encodes a beta chemokine receptor important for B-lineage maturation and antigen-driven B-cell differentiation. Both the duration and peak of TDV-2 virus load were positively correlated with the development of higher DENV-2 neutralizing antibody titers on day 30, when they were highest (r2 = 0.42 and 0.40, respectively) (S2 Fig).

Features of the innate immune response to TDV correlate with neutralizing antibody titer

All animals vaccinated with TDV developed a neutralizing antibody response to each of the four dengue serotypes (S3 Fig). Median neutralizing antibody titer (median PRNT50) was calculated to represent an overall response to all serotypes of the tetravalent vaccine. Median PRNT50 differed between the groups, and was highest for animals administered SC TDV (Fig 5A). Median titers were higher in animals that received double dose TDV rather than a single dose before boost (days 30 and 53), but were similar post-boost (days 75 and 88), and did not change significantly after challenge with wt DENV (Fig 5A and S3 Fig). Six hundred and thirteen genes were significantly correlated with median PRNT50 on day 30 during at least 1 time-point following vaccination (Spearman’s rank correlation coefficient, p<0.01). The strongest correlations were seen with genes related to the type I IFN response on day 7, including DHX58, OASL, GBP1, GBP2, IFI27, XAF1 and STAT1 (Spearman’s ρ: 0.58–0.64, p<0.0007). On day 5, transcript abundance for KLRC3 (killer cell lectin-like receptor subfamily c, member 3), which is expressed primarily in NK cells and involved in T cell responses, was positively correlated with median PRNT50 on day 30 (Spearman’s ρ = 0.52, p = 0.004).

Fig 5. Correlation of viral load and relative transcript abundance to neutralizing antibody titer, by animal.

A) Median neutralizing antibody titer (median log PRNT50) over time, by vaccination group. Colors represent different vaccination groups. B) Relative transcript abundance for the 11 type I IFN genes significantly positively correlated with neutralizing antibody titer on day 30. Median expression in 5 lowest responders compared with median expression in 5 highest responders over time. C) Median relative transcript abundance of the 11 type I IFN genes over time in high responders (black line) and low responders (grey line). IDd, intradermal double dose; IDs, intradermal single dose; IDs N&S, intradermal single dose needle and syringe; SCd, subcutaneous double dose; SCs, subcutaneous single dose.

Forty-two of these genes also showed significant changes in abundance following vaccination (S6 Table), and 11 were involved in the antiviral/type I IFN response described above. The temporal transcript abundance pattern of these 11 genes differed between all ‘low responders’ (n = 5 animals whose PRNT50 on day 30 was less than 4-fold greater than baseline) and ‘high responders’ (n = 5 animals with the highest PRNT50 on day 30) (Fig 5B and 5C). The type I IFN response is induced in response to viral sensing, but the transcript median relative abundance of this type I IFN gene set was not significantly correlated with either the extent or duration of viremia (peak viral load r2 = 0.34, duration of viremia r2 = 0.32) (S4 Fig). Pathway analysis using BTMs of genes ranked by correlation score also showed that modules of genes related to T cell activation and differentiation were enriched on day 5, and the type I IFN response was enriched on day 7, supporting these findings (S7 Table).

In most cases, the dominant neutralizing antibody response was directed against serotype 2 (S3 Fig), as seen previously [9,11]. Comparison of transcriptional responses by day to DENV-2 specific neutralizing antibody titers pre-boost on day 30 revealed a positive correlation with the relative transcript abundance for genes involved in T cell proliferation and activation on day 5. These genes included PRKCQ, which encodes the enzyme, protein kinase C theta, required for T cell activation (Spearman’s ρ = 0.6, p = 0.001), and IL6ST, which encodes a cytokine signal transducer (Spearman’s ρ = 0.62, p = 0.0004). When serotype responses were examined separately, 37 transcripts were significantly correlated with the day 30 neutralizing antibody response for at least 3 of the 4 DENV serotypes. While there were no significant gene ontologies, positive regulation of apoptosis was a common signature, and included the genes XAF1, BARD1, GCH1, MAGED1, PRKCA. Four of these transcripts, which are all involved in the antiviral/type I IFN response, were also significantly enriched following TDV vaccination (DHX58, IFI27, ISG15, and XAF1).


Understanding the mechanisms that underlie the development of protective immunity against dengue infection may assist in the development of an effective vaccine against dengue. In this study, we used genome-wide transcriptional profiling to identify early transcriptional responses to vaccination that may act as predictors of an effective vaccine response. All animals vaccinated with TDV developed neutralizing antibodies to each of the 4 dengue serotypes, and in most cases the dominant antibody response was to serotype 2, as seen previously [9,11]. TDV-2 was the only strain with vRNA detected after vaccination, and the abundance of TDV-2 vRNA was positively correlated with the development of DENV-2 neutralizing antibodies. Furthermore, we observed changes in transcript abundance for genes related to T cell activation on day 5 that were positively correlated with the presence of TDV-2 vRNA on days 5–14, and the development of higher DENV-2 specific neutralizing antibody titers on day 30, suggesting that TDV-2 is more immunogenic than TDV-1, TDV-3 and TDV-4, as seen in humans [11].

Changes in abundance of transcripts related to the antiviral and type I IFN response were the most notable features of the response to TDV vaccination These changes were strongest following subcutaneous compared with intradermal vaccination, and double dose administration on day 0 compared with single dose. The type I IFN response is well described for natural DENV infection [1216], and appears to be associated with vaccination using live attenuated viruses. The live, attenuated yellow fever vaccine (YF-17D) [17,19,22] and influenza vaccine (LAIV), both resulted in a type I IFN response, but trivalent inactivated influenza vaccine (TIV) did not, suggesting that viral replication may also increase immunogenicity of LAIV vaccines [20]. It has recently been shown that IFN-α production by pDCs following dengue infection requires active viral replication in neighboring infected cells, but is triggered by internalization in pDCs of non-infectious viral components [37]. In this study, animals that experienced higher viral loads tended to have higher abundances of gene transcripts related to the type I IFN response, but the correlation was not significant. While we cannot conclude from these data that a stronger IFN response clears the virus, studies of dengue infection in humans and mice indicate that the innate response is important for controlling viral replication and pathology, and that stronger IFN responses are beneficial to the host.

Type I IFNs (IFN-α/β) are pleiotropic cytokines that play important roles in both innate and adaptive immune responses [32,3840]. During dengue infection, viral recognition of dsRNA by pattern recognition receptors and the cytoplasmic helicases retinoic-acid-inducible gene I (RIG-I) and melanoma differentiation-associated gene 5 (MDA5), leads to an intracellular signaling cascade responsible for the production of IFN-α/β[38]. IFN-α/β activates the JAK/STAT pathway, inducing expression of many interferon stimulated genes (ISGs). Transcripts for several genes in this pathway became more abundant following TDV vaccination, including DDX58 (RIG-I), EIF2AK2 (PKR), STAT1, IRF7, OASL and OAS2, MX1, IFI44, ISG15, and XAF1. The magnitude of the type I IFN response was much greater following infection with wt DENV, which might reflect greater replication and increased immunogenicity of the wt virus compared with vaccine viruses. However, the response observed in NHPs following wt DENV infection was more subtle compared with natural infection in humans [15], consistent with the lack of pathology in non-human primates following dengue infection.

The type I IFN response that dominated the response to TDV was modulated by both dose and route. After ID immunization, most of the increases in abundance of genes related to the type I IFN response and lymphocyte activation occurred on day 1, whereas after SC immunization these genes increased in abundance on days 5 and 7. ID immunization targets Langerhans and dermal dendritic cells, and macrophages in the epidermis and dermis. The dermis is highly vascularized, and skin DCs, antigen presenting cells, and monocytes and neutrophils recruited from the peripheral blood, transport antigens and vaccine components to draining lymph nodes, limiting transfer to the peripheral blood circulation [41]. The earlier type I IFN transcriptional response, and reduced detectable viremia, following ID immunization may be due to the skin’s resident immune cells and faster clearance of the vaccine. The stronger, and later response in a subset of type I IFN genes following subcutaneous immunization, particularly double dose administration, may have resulted from the larger volume and/or dose of vaccine inoculum (S1 Materials and Methods), and prolonged persistence of the vaccine virus in the adipose tissue of the SC layer [41,42]. Numerous toll-like receptors (TLRs) are expressed by the skin’s immune cells, and by fibroblasts, adipocytes and macrophages of the subcutaneous tissue. TLRs recognize microbial pathogens and trigger signaling cascades involved in both innate and adaptive immune responses. Transcripts encoding TLR4 and TLR5 increased in abundance following TDV vaccination, yet there were no significant differences in expression between ID and SC routes.

Increased abundances of transcripts for genes involved in the type I IFN response were correlated with the development of higher neutralizing antibody titers on day 30, suggesting that type I IFN may play a role in the activation of B cells. The sequence of events linking type I interferon responses to neutralizing antibody titer in our study is not known, but the importance of type I IFN for the activation of the production of antibody by B cells in other systems has been described previously [4345]. Deal et al. demonstrated that pDC-derived type I IFN was required to activate B cells for production of virus-specific antibodies in human in vitro and mouse in vivo models of rotavirus infection [46]. While neutralizing antibody titers are the primary measure of immunity to dengue virus infection and the response to vaccination, they do not always reflect protection. In fact, T-cell responses may also play an important role in immunity to dengue [47]. The live-attenuated tetravalent chimeric yellow fever-dengue vaccine (CYD23) resulted in no protection against DENV-2 in a phase 2b efficacy trial, and lower efficacy (35%) in a phase 3 trial, despite high neutralizing antibody titers against all four serotypes [48,49]. This may be due to the absence in CYD23 of the dengue backbone that harbors important epitopes targeted by CD8T+ T-cells [47]. In our study, increased abundance of genes related to T cell activation on day 5 correlated with the concomitant presence of TDV-2 vRNA (days 5–14) and subsequent DENV-2 specific neutralizing antibody titers on day 30, suggesting a TDV-2 specific adaptive response. We were unable to correlate the transcriptional response with the development of cell-mediated immune responses, as these were only measured for animals in Group 4 (SCd PhJ) [23]. In these animals, TDV induced CD4+ and CD8+ T cells that targeted DENV-2 NS1, NS3 and NS5 proteins and that cross-reacted with DENV-4 NS3 and NS5 proteins [23]. Additional studies will be useful for evaluating links between the early transcriptional response and the development of cell-mediated immune responses.

This study focused on the response to TDV in non-human primates, which serve as important animal models for understanding vaccine responses and efficacy [50]. The responses we observed are similar to those elicited by DENV infection in humans; however, these non-human primates do not develop signs of disease following DENV infection. Characterization of gene expression changes that occur in humans vaccinated with dengue vaccines will be valuable for investigating the role of specific cell types in shaping innate and adaptive immunity.

Supporting Information

S1 Text. Supporting information on viruses, animals and study design.


S1 Table. Genes differentially expressed after TDV vaccination in 30 animals compared to baseline.


S2 Table. Baseline-transformed (average of d-11 and d-2) relative transcript abundances (log2) by animal for 595 differentially expressed genes after TDV vaccination.


S3 Table. Additional modules of IFN-induced gene sets.


S4 Table. Blood transcript modules enriched following vaccination (FDR<5%).


S5 Table. Fold-change over time of significant genes in Clusters 1–3 for animals in the placebo (PBS) group following wt DENV challenge.


S6 Table. Overlap of genes with significant changes in transcript abundance following vaccination and correlated with median neutralizing antibody titers on day 30.


S7 Table. Enriched blood transcript modules from genes ranked by Spearman correlation coefficient r (FDR<5%).


S1 Fig. The antiviral/type I IFN response following TDV vaccination.

Unsupervised clustering of median expression of 282 genes (379 genes, filtered for reliably measured transcript abundance in 2 out of 3 samples) over time in all vaccinated animals, and median expression by group. Red indicates an increase in transcript abundance, blue indicates a decrease in transcript abundance (FDR<0.05, fold-change≥1.3). Significance by group marked by grey column at the right of each heatmap, with red (increase in abundance) and blue (decrease in abundance). Placebo vaccination and placebo recipient challenge with wt DENV shown for comparison.


S2 Fig. Correlation of DENV-2 neutralizing antibody titer (PRNT50) on day 30 with (A) duration of viremia; and (B) peak viral load.


S3 Fig. Geometric mean neutralizing antibody titer (PRNT50) over time for each treatment group.


S4 Fig. Correlation between abundance of type I IFN genes and (A) duration of TDV-2 viremia and (B) peak TDV-2 viral load.



We wish to acknowledge Dr. Saverio “Buddy" Capuano, Associate Director of Animal Services at the University of Wisconsin-Madison Primate Center (WNPRC), and Dr. Joseph Brewoo for help with sample transfer, experimental design and data interpretation. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Author Contributions

Conceived and designed the experiments: FRS SJP JEO DAR. Performed the experiments: FRS SJP CDP DTS. Analyzed the data: FRS SJP DAR. Contributed reagents/materials/analysis tools: FRS SJP CDP DTS JEO DAR. Wrote the paper: FRS SJP CDP DTS JEO DAR.


  1. 1. Bhatt S, Gething PW, Brady OJ, Messina JP, Farlow AW, Moyes CL, et al. The global distribution and burden of dengue. Nature. 2013;496:504–7. pmid:23563266
  2. 2. Green S, Rothman A. Immunopathological mechanisms in dengue and dengue hemorrhagic fever. Curr Opin Infect Dis. 2006;19:429–36. pmid:16940865
  3. 3. Guzman MG, Vazquez S. The complexity of antibody-dependent enhancement of dengue virus infection. Viruses. 2010. p. 2649–62. pmid:21994635
  4. 4. Halstead SB, Mahalingam S, Marovich MA, Ubol S, Mosser DM. Intrinsic antibody-dependent enhancement of microbial infection in macrophages: disease regulation by immune complexes. Lancet Infectious Diseases. 2010;10:712–22. pmid:20883967
  5. 5. Whitehead SS, Blaney JE, Durbin AP, Murphy BR. Prospects for a dengue virus vaccine. Nat Rev Microbiol. 2007;5:518–28. pmid:17558424
  6. 6. Bhamarapravati N, Yoksan S, Chayaniyayothin T, Angsubphakorn S, Bunyaratvej A. Immunization with a live attenuated dengue-2-virus candidate vaccine (16681-PDK 53): clinical, immunological and biological responses in adult volunteers. Bull World Health Org. 1987;65:189–95. pmid:3496985
  7. 7. Dharakul T, Kurane I, Bhamarapravati N, Yoksan S, Vaughn DW, Hoke CH, et al. Dengue virus-specific memory T cell responses in human volunteers receiving a live attenuated dengue virus type 2 candidate vaccine. J Infect Dis. 1994;170:27–33. pmid:7912253
  8. 8. Huang CY-H, Butrapet S, Tsuchiya KR, Bhamarapravati N, Gubler DJ, Kinney RM. Dengue 2 PDK-53 virus as a chimeric carrier for tetravalent dengue vaccine development. J Virol. 2003;77:11436–47. pmid:14557629
  9. 9. Osorio JE, Brewoo JN, Silengo SJ, Arguello J, Moldovan IR, Tary-Lehmann M, et al. Efficacy of a tetravalent chimeric dengue vaccine (DENVax) in Cynomolgus macaques. Am J Trop Med Hyg. 2011;84:978–87. pmid:21633037
  10. 10. Osorio JE, Huang CYH, Kinney RM, Stinchcomb DT. Development of DENVax: A chimeric dengue-2 PDK-53-based tetravalent vaccine for protection against dengue fever. Vaccine. 2011;29:7251–60. pmid:21777638
  11. 11. Osorio JE, Velez ID, Thomson C, Lopez L, Jimenez A, Haller AA, et al. Safety and immunogenicity of a recombinant live attenuated tetravalent dengue vaccine (DENVax) in flavivirus-naive healthy adults in Colombia: a randomised, placebo-controlled, phase 1 study. Lancet Infect Dis. 2014;14:830–8. pmid:25087476
  12. 12. Hoang LT, Lynn DJ, Henn M, Birren BW, Lennon NJ, Le PT, et al. The early whole-blood transcriptional signature of dengue virus and features associated with progression to dengue shock syndrome in Vietnamese children and young adults. J Virol. 2010;84:12982–94. pmid:20943967
  13. 13. Tolfvenstam T, Lindblom A, Schreiber MJ, Ling L, Chow A, Ooi EE, et al. Characterization of early host responses in adults with dengue disease. BMC Infect Dis. 2011;11:209. pmid:21810247
  14. 14. Simmons CP, Popper S, Dolocek C, Chau TNB, Griffiths M, Dung NTP, et al. Patterns of host genome-wide gene transcript abundance in the peripheral blood of patients with acute dengue hemorrhagic fever. J Infect Dis. 2007;195:1097–107. pmid:17357045
  15. 15. Popper SJ, Gordon A, Liu M, Balmaseda A, Harris E, Relman DA. Temporal dynamics of the transcriptional response to dengue virus infection in Nicaraguan children. PLoS Negl Trop Dis. 2012;6:e1966. pmid:23285306
  16. 16. Kwissa M, Nakaya HI, Onlamoon N, Wrammert J, Villinger F, Perng GC, et al. Dengue virus infection induces expansion of a CD14(+)CD16(+) monocyte population that stimulates plasmablast differentiation. Cell Host Microbe. 2014;16:115–27. pmid:24981333
  17. 17. Li S, Rouphael N, Duraisingham S, Romero-Steiner S, Presnell S, Davis C, et al. Molecular signatures of antibody responses derived from a systems biology study of five human vaccines. Nat Immunol. 2014;15:195–204. pmid:24336226
  18. 18. Obermoser G, Presnell S, Domico K, Xu H, Wang Y, Anguiano E, et al. Systems scale interactive exploration reveals quantitative and qualitative differences in response to influenza and pneumococcal vaccines. Immunity. 2013;38(4):831–44. pmid:23601689
  19. 19. Querec TD, Akondy RS, Lee EK, Cao W, Nakaya HI, Teuwen D, et al. Systems biology approach predicts immunogenicity of the yellow fever vaccine in humans. Nat Immunol. 2009;10:116–25. pmid:19029902
  20. 20. Nakaya HI, Wrammert J, Lee EK, Racioppi L, Marie-Kunze S, Haining WN, et al. Systems biology of vaccination for seasonal influenza in humans. Nat Immunol. 2011;12:786–95. pmid:21743478
  21. 21. Caskey M, Lefebvre F, Filali-Mouhim A, Cameron MJ, Goulet J-P, Haddad EK, et al. Synthetic double-stranded RNA induces innate immune responses similar to a live viral vaccine in humans. J Exp Med. 2011;208:2357–66. pmid:22065672
  22. 22. Gaucher D, Therrien R, Kettaf N, Angermann BR, Boucher G, Filali-Mouhim A, et al. Yellow fever vaccine induces integrated multilineage and polyfunctional immune responses. J Exp Med. 2008;205:3119–31. pmid:19047440
  23. 23. Ambuel Y, Young G, Brewoo JN, Paykel J, Weisgrau KL, Rakasz EG, et al. A rapid immunization strategy with a live-attenuated tetravalent dengue vaccine elicits protective neutralizing antibody responses in non-human primates. Front Immunol. 2014;5:263. pmid:24926294
  24. 24. Smith LM, Hensley LE, Geisbert TW, Johnson J, Stossel A, Honko A, et al. Interferon- therapy prolongs survival in rhesus macaque models of ebola and marburg hemorrhagic fever. J Infect Dis. 2013;208:310–8. pmid:23255566
  25. 25. Yen JY, Garamszegi S, Geisbert JB, Rubins KH, Geisbert TW, Honko A, et al. Therapeutics of Ebola hemorrhagic fever: Whole-genome transcriptional analysis of successful disease mitigation. J Infect Dis. 2011;204 Suppl 3:S1043–52. pmid:21987740
  26. 26. Garamszegi S, Yen JY, Honko AN, Geisbert JB, Rubins KH, Geisbert TW, et al. Transcriptional correlates of disease outcome in anticoagulant-treated non-human primates infected with ebolavirus. PLoS Negl Trop Dis. 2014;8:e3061. pmid:25079789
  27. 27. Tusher VG, Tibshirani R, Chu G. Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci USA. 2001;98:5116–21. pmid:11309499
  28. 28. Huang DW, Sherman BT, Zheng X, Yang J, Imamichi T, Stephens R, et al. Extracting biological meaning from large gene lists with DAVID. Current Protocols in Bioinformatics. 2009;Chapter 13:Unit 13.11. pmid:19728287
  29. 29. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci USA. 2005;102(43):15545–50. pmid:16199517
  30. 30. Hishiki T, Han Q, Arimoto K, Shimotohno K, Igarashi T, Vasudevan SG, et al. Interferon-mediated ISG15 conjugation restricts dengue virus 2 replication. Biochem Biophys Res Commun. 2014;448:95–100. pmid:24769207
  31. 31. Rickert RC, Jellusova J, Miletic A V. Signaling by the tumor necrosis factor receptor superfamily in B-cell biology and disease. Immunol Rev. 2011;244:115–33. pmid:22017435
  32. 32. Waddell SJ, Popper SJ, Rubins KH, Griffiths MJ, Brown PO, Levin M, et al. Dissecting interferon-induced transcriptional programs in human peripheral blood cells. PLoS One. 2010;5:e9753. pmid:20339534
  33. 33. Swiecki M, Colonna M. Unraveling the functions of plasmacytoid dendritic cells during viral infections, autoimmunity, and tolerance. Immunological Reviews. 2010;234:142–62. pmid:20193017
  34. 34. Peters VA, Joesting JJ, Freund GG. IL-1 receptor 2 (IL-1R2) and its role in immune regulation. Brain Behav Immun. 2013;32:1–8. pmid:23195532
  35. 35. Wang T, Wang Z, Wang L, Yan L, Wan J, Zhang S, et al. CRISPLD2 is expressed at low levels during septic shock and is associated with procalcitonin. PLoS One. 2013;8:e65743. pmid:23799041
  36. 36. Wang Z-Q, Xing W-M, Fan H-H, Wang K-S, Zhang H-K, Wang Q-W, et al. The novel lipopolysaccharide-binding protein CRISPLD2 is a critical serum protein to regulate endotoxin function. J Immunol. 2009;183:6646–56. pmid:19864597
  37. 37. Décembre E, Assil S, Hillaire MLB, Dejnirattisai W, Mongkolsapaya J, Screaton GR, et al. Sensing of immature particles produced by dengue virus infected cells induces an antiviral response by plasmacytoid dendritic cells. PLoS Pathog. 2014;10:e1004434. pmid:25340500
  38. 38. Green AM, Beatty PR, Hadjilaou A, Harris E. Innate immunity to dengue virus infection and subversion of antiviral responses. J Mol Biol. 2014;426:1148–60. pmid:24316047
  39. 39. Colonna M, Trinchieri G, Liu Y-J. Plasmacytoid dendritic cells in immunity. Nat Immunol. 2004;5:1219–26. pmid:15549123
  40. 40. Katze MG, He Y, Gale M. Viruses and interferon: a fight for supremacy. Nat Rev Immunol. 2002;2:675–87. pmid:12209136
  41. 41. Lambert PH, Laurent PE. Intradermal vaccine delivery: will new delivery systems transform vaccine administration? Vaccine. 2008;26:3197–208. pmid:18486285
  42. 42. Combadiere B, Liard C. Transcutaneous and intradermal vaccination. Hum Vaccin. 2011;7:811–27. pmid:21817854
  43. 43. Le Bon A, Schiavoni G, D'Agostino G, Gresser I, Belardelli F, Tough DF. Type I interferons potently enhance humoral immunity and can promote isotype switching by stimulating dendritic cells in vivo. Immunity. 2001;14:461–70. pmid:11336691
  44. 44. Jego G, Palucka AK, Blanck JP, Chalouni C, Pascual V, Banchereau J. Plasmacytoid dendritic cells induce plasma cell differentiation through type I interferon and interleukin 6. Immunity. 2003;19:225–34. pmid:12932356
  45. 45. Coro ES, Chang WLW, Baumgarth N. Type I IFN receptor signals directly stimulate local B cells early following influenza virus infection. J Immunol. 2006;176:4343–51. pmid:16547272
  46. 46. Deal EM, Lahl K, Narváez CF, Butcher EC, Greenberg HB. Plasmacytoid dendritic cells promote rotavirus-induced human and murine B cell responses. J Clin Invest. 2013;123(6):2464–74. pmid:23635775
  47. 47. Weiskopf D, Sette A. T-cell immunity to infection with dengue virus in humans. Front Immunol. 2014;5:93. pmid:24639680
  48. 48. Sabchareon A, Wallace D, Sirivichayakul C, Limkittikul K, Chanthavanich P, Suvannadabba S, et al. Protective efficacy of the recombinant, live-attenuated, CYD tetravalent dengue vaccine in Thai schoolchildren: a randomised, controlled phase 2b trial. Lancet. 2012;380:1559–67. pmid:22975340
  49. 49. Capeding MR, Tran NH, Hadinegoro SRS, Ismail HIHM, Chotpitayasunondh T, Chua MN, et al. Clinical efficacy and safety of a novel tetravalent dengue vaccine in healthy children in Asia: a phase 3, randomised, observer-masked, placebo-controlled trial. Lancet. 2014;384:1358–65. pmid:25018116
  50. 50. Sariol CA, White LJ. Utility, limitations, and future of non-human primates for dengue research and vaccine development. Front Immunol. 2014;5:452. pmid:25309540