Successful Topical Respiratory Tract Immunization of Primates against Ebola Virus

Abstract

Ebola virus causes outbreaks of severe viral hemorrhagic fever with high mortality in humans. The virus is highly contagious and can be transmitted by contact and by the aerosol route. These features make Ebola virus a potential weapon for bioterrorism and biological warfare. Therefore, a vaccine that induces both systemic and local immune responses in the respiratory tract would be highly beneficial. We evaluated a common pediatric respiratory pathogen, human parainfluenza virus type 3 (HPIV3), as a vaccine vector against Ebola virus. HPIV3 recombinants expressing the Ebola virus (Zaire species) surface glycoprotein (GP) alone or in combination with the nucleocapsid protein NP or with the cytokine adjuvant granulocyte-macrophage colony-stimulating factor were administered by the respiratory route to rhesus monkeys—in which HPIV3 infection is mild and asymptomatic—and were evaluated for immunogenicity and protective efficacy against a highly lethal intraperitoneal challenge with Ebola virus. A single immunization with any construct expressing GP was moderately immunogenic against Ebola virus and protected 88% of the animals against severe hemorrhagic fever and death caused by Ebola virus. Two doses were highly immunogenic, and all of the animals survived challenge and were free of signs of disease and of detectable Ebola virus challenge virus. These data illustrate the feasibility of immunization via the respiratory tract against the hemorrhagic fever caused by Ebola virus. To our knowledge, this is the first study in which topical immunization through respiratory tract achieved prevention of a viral hemorrhagic fever infection in a primate model.

Ebola virus (EBOV), a member of the Filoviridae family of the Mononegavirales order, is an enveloped virus with a negative-sense RNA genome of 19 kb. EBOV causes the most severe viral hemorrhagic fever disease known, with a mortality rate in humans of up to 88% (37). The first recorded outbreaks of EBOV took place in 1976, and a number of sporadic outbreaks have occurred in the ensuing years. All the major human outbreaks of EBOV have occurred in or near the rain forests of central Africa (reviewed in reference 28). The natural reservoir for EBOV has not been identified, although a recent study implicated fruit bats; monkeys and apes are readily infected in the wild and probably serve as intermediate hosts (23).

The development of vaccines against EBOV is a priority because of the severity and uncontrolled nature of its outbreaks and the possibility of its use in bioterrorism or biological warfare. Direct contact with tissues of diseased individuals appears to play the major role during outbreaks in humans (15). However, a particular concern is the ability of EBOV to infect nonhuman primates by the laboratory-generated aerosol route (19). Multiple early attempts to develop an effective EBOV vaccine were unsuccessful, indicating that protection against EBOV infection is not easily achieved (reviewed in references 13 and 16). More recently, however, vectored vaccines based on adenovirus or vesicular stomatitis (VSV) were effective when administered parenterally in a nonhuman primate model (20, 33), although each of these vectors remains to be demonstrated to be safe and effective in humans. In any event, it would be advantageous to develop alternative vectors and strategies of vaccination against EBOV and other highly pathogenic viruses. In particular, studies with human respiratory viruses, such as influenza and respiratory syncytial viruses, indicate that topical immunization induces both mucosal and systemic immune responses and is the most effective means of restricting infection and viral shedding from the respiratory tract.

Because the respiratory route is often a portal of pathogen entry and egress, we have been evaluating members of the Paramyxoviridae family as live intranasal vaccine vectors. Like filoviruses, paramyxoviruses are members of the Mononegavirales order and are enveloped RNA viruses with a negative-sense RNA genome of 15 to 19 kb. They include a number of common pathogens of humans and animals, including mumps, measles, respiratory syncytial, and parainfluenza viruses and characteristically infect by the respiratory route. In the present study, we employed human parainfluenza virus type 3 (HPIV3), a common pediatric respiratory pathogen, as a vaccine vector. HPIV3 infects the superficial layer of the respiratory epithelium and does not spread significantly beyond the respiratory tract, providing increased vaccine safety (9, 40; reviewed in reference 6). Foreign coding sequences are inserted into the gene order as transcription cassettes under the control of vector-specific transcription signals and are expressed as separate mRNAs.

Mucosal immunization through the respiratory tract against any viral hemorrhagic fever had not been explored until recently, when we demonstrated that intranasal (i.n.) immunization of guinea pigs with a single dose of HPIV3 expressing the EBOV (Zaire species) surface glycoprotein (GP) alone or in combination with the nucleoprotein (NP) induced a high level of protection against a high dose of guinea pig-adapted EBOV administered by intraperitoneal (i.p.) injection (7). However, rodent models for EBOV can be poorly predictive surrogates for primates, and some previous EBOV vaccines that were effective in rodents failed completely in nonhuman primates (17, 26). In the present study, we evaluated mucosal vaccination through the respiratory tract against viral hemorrhagic fever caused by a highly lethal dose of EBOV in rhesus monkeys.

MATERIALS AND METHODS

Construction of recombinant HPIV3.

HPIV3/EboGP and HPIV3/EboGP+NP were constructed previously by inserting a transcription cassette encoding the EBOV (Mayinga strain) GP gene between the HPIV3 P and M genes alone or in combination with a cassette encoding the NP inserted between the HPIV3 HN and L genes (7). For the present study, an additional virus was constructed, HPIV3/EboGP+GM-CSF. To make this virus, the plasmid encoding HPIV3/EboGP was modified to contain a cassette encoding human granulocyte-macrophage colony-stimulating factor (GM-CSF) inserted between the HN and F genes. A cDNA of the GM-CSF coding sequence (kindly provided by Elizabeth Jaffee, John Hopkins University School of Medicine, Baltimore, MD) was amplified with the forward primer TACACTTACGCGTGATGTGGCTGCAGAGCCTGCT (the beginning of the open reading frame is underlined, and an MluI site is indicated by bold type) and the reverse primer TACACTTTTCGAAGTTATCACTCCTGGACTGGCTCCC (the end of the open reading frame with the stop codon is underlined, and a BstBI site is indicated by bold type). The PCR product was then digested with MluI and BstBI and cloned into the MluI-BstBI window of the plasmid pHPIV3(XhoI-SphI)StuI (bearing nucleotides 7,437 to 11,317 of the HPIV3 antigenomic cDNA in which an oligonucleotide duplex encoding an additional set of HN gene end, intergenic, and gene start signals had been introduced into the noncoding region of the HN gene downstream of the open reading frame [31]). The modified fragment of HPIV3 cDNA with the insert of GM-CSF was then used to replace the corresponding fragment of the plasmid pHPIV3/EboGP (7). The resulting plasmid was used to recover the HPIV3/EboGP+GM-CSF virus by transfection, together with plasmids expressing the HPIV3 N, P, and L proteins, into BHK-21 cells constitutively expressing the T7 polymerase (kindly provided by Ursula J. Buchholz, National Institute of Allergy and Infectious Diseases) (3, 11). The virus was amplified in LLC-MK2 rhesus monkey kidney cells, the expression of EBOV GP was confirmed by Western blotting as previously described (7), and the expression of GM-CSF was confirmed by an enzyme-linked immunosorbent assay (ELISA) (Quantikine human GM-CSF immunoassay; R&D Systems, Minneapolis, MN) of tissue culture medium supernatants from infected cells. Infection of 10⁶ cells with 2 50% tissue culture infective doses (TCID₅₀) per cell of HPIV3/EboGP+GM-CSF resulted in the production of 468 ng of GM-CSF by 24 h postinfection compared to 2 ng by cells infected with HPIV3 lacking the GM-CSF insert and 0.1 ng by mock-infected cells. Evaluation of the growth kinetics of HPIV3/EboGP+GM-CSF in LLC-MK2 and Vero cells (multiplicity of infection of 2 TCID₅₀ per cell) showed that its replication was moderately reduced (by ∼0.5 to 1.5 log₁₀/ml) on days 1, 2, and 3 postinfection compared to HPIV3, which is an expected consequence of the addition of the transcription cassettes (6), and was comparable to that of HPIV3/EboGP and HPIV3/EboGP+NP (7; also data not shown).

Immunization and challenge of rhesus monkeys.

The immunization and challenge of rhesus monkeys were performed according to the CDC Animal Care and Use Committee-approved protocols. Sixteen 3-year-old rhesus monkeys of either sex weighing 3.7 to 6.2 kg were confirmed to be seronegative for HPIV3. During manipulations, the animals were anesthetized with Telazol injected by the intramuscular route at 5 mg/kg of body weight. The recombinant HPIV3 viruses were diluted in Leibowitz-15 medium (Invitrogen, Carlsbad, CA) and administered by a combined i.n. and intratracheal (i.t.) inoculation using a volume of 0.5 ml into each nostril and a volume of 1.0 ml into the trachea of the anesthetized animals that had been placed in dorsal recumbency. Two experiments were performed: in experiment 1, the animals received a single dose of 4 × 10⁶ TCID₅₀ of an individual HPIV3, whereas in experiment 2, the animals received doses of 2 × 10⁷ TCID₅₀ on days 0 and 28 (Table 1).

TABLE 1.

Immunization of rhesus monkeys via the respiratory tract with the HPIV3/EBOV constructs protects against challenge with EBOV

Expt no. and immunizing recombinant HPIV3	Animal no.	Serum EBOV-specific IgG response (ELISA)a				Serum EBOV-specific IgA response (ELISA)a				Serum HPIV3-specific antibody response (hemagglutination inhibition)a			Clinical response to challenge with EBOV
		Day 0	After 1st immuniz.	After 2nd immuniz. (expt 2 only)	After EBOV challenge	Day 0	After 1st immuniz.	After 2nd immuniz. (expt 2 only)	After EBOV challenge	Day 0	After 1st immuniz.	After 2nd immuniz. (expt 2 only)
Expt 1b
HPIV3 (control)	1	≤4.6	≤4.6		Deadc	≤4.6	≤4.6		Dead	<2.0	9.0		Became ill on day 4; euthanized on day 6
	2	≤4.6	≤4.6		Dead	≤4.6	≤4.6		Dead	<2.0	8.0		Became ill on day 4; euthanized on day 6
HPIV3/EboGP	3	≤4.6	10.6		18.6	≤4.6	≤4.6		12.6	<2.0	8.0		No signs of disease
	4	≤4.6	8.6		16.0	≤4.6	≤4.6		12.6	<2.0	6.0		No signs of disease except that the animal did not eat well on day 7; normal on day 8
HPIV3/EboGP+NP	5	≤4.6	8.6		16.6	≤4.6	≤4.6		12.6	<2.0	3.0		No signs of disease except that the animal did not eat well on day 7; normal on day 8
	6	≤4.6	8.6		Dead	≤4.6	≤4.6		Dead	<2.0	4.0		Became ill on day 7; died on day 8
HPIV3/EboGP+GM-CSF	7	≤4.6	8.6		14.6	≤4.6	≤4.6		14.6	<2.0	7.0		Mild symptoms of the disease on day 7; normal on day 8
	8	≤4.6	8.6		16.6	≤4.6	≤4.6		16.6	<2.0	8.0		No signs of disease
Expt 2d
HPIV3 (control)	9	≤4.6	≤4.6	≤4.6	Dead	≤4.6	≤4.6	≤4.6	Dead	<2.0	10.0	11.0	Became ill on day 4; euthanized on day 6
	10	≤4.6	≤4.6	≤4.6	Dead	≤4.6	≤4.6	≤4.6	Dead	<2.0	10.0	11.0	Became ill on day 4; euthanized on day 5
HPIV3/EboGP (1 dose)	11	≤4.6	10.6	10.6	Dead	≤4.6	6.6	6.6	Dead	<2.0	11.0	11.0	Died on day 6 (no symptoms on the previous days)
	12	≤4.6	10.6	10.6	18.6	≤4.6	6.6	6.6	14.6	<2.0	10.0	10.0	Became ill on day 6; only mild symptoms on days 10-16
	13	≤4.6	10.6	12.6	18.6	≤4.6	6.6	6.6	12.6	<2.0	10.0	9.0	No signs of disease
HPIV3/EboGP (2 doses)	14	≤4.6	10.6	14.6	18.6	≤4.6	6.6	10.6	14.6	<2.0	9.0	11.0	No signs of disease
	15	≤4.6	10.6	10.6	18.6	≤4.6	6.6	10.6	14.6	<2.0	10.0	10.0	No signs of disease
	16	≤4.6	10.6	12.6	18.6	≤4.6	6.6	10.6	14.6	<2.0	9.0	10.0	No signs of disease

^aAntibody titers are expressed as the reciprocal log₂. Immuniz., immunization.

^bIn experiment 1, on day 0, the animals were immunized i.n. and i.t. with 4 × 10⁶ TCID₅₀ of the indicated recombinant HPIV3; on day 28, the animals were challenged i.p. with 1,000 TCID₅₀ of EBOV. The postimmunization serum samples were from day 28 postimmunization, and the postchallenge serum samples from surviving animals were from day 14 postchallenge. The antibody titers in serum samples that were taken on day 21 postimmunization were not substantially different from that on day 28, and therefore are not shown.

^cPostchallenge serum samples were not available on day 14 (experiment 1) or day 16 (experiment 2) postchallenge because the animal had previously died with or had been euthanized for the symptoms of severe EBOV infection.

^dIn experiment 2, on day 0, the animals were immunized i.n. and i.t. with 2 × 10⁷ TCID₅₀ of the indicated recombinant HPIV3; on day 28, animals 9 to 13 (the control group and the one-dose group) received a second mock dose consisting of empty HPIV3 vector, while animals 14 to16 (the two-dose group) received a second dose of HPIV3/EboGP; on day 67 all animals were challenged i.p. with 1,000 PFU of EBOV. The serum samples taken after the first and second immunization were from days 28 and 67, respectively, following the first immunization, and the postchallenge serum samples from surviving animals were from 16 days postchallenge.

In experiment 1, the animals were challenged on day 28 with 1,000 TCID₅₀ of EBOV (Zaire species, Mayinga strain), which had been isolated in 1976 (21) and passaged once in Vero E6 cells and once in a cynomolgus macaque; a spleen homogenate from this infected animal was used for the challenge. The infectivity of this virus suspension was characterized by determination of its TCID₅₀ in Vero E6 cells because it does not efficiently form plaques in those cells. In experiment 2, the animals were challenged on day 67 (39 days following the second vaccine dose) with 1,000 PFU of the same isolate of EBOV, in this case using the preparation that had undergone only the single aforementioned passage in Vero E6 cells and readily formed plaques in these cells. These two virus preparations are approximately equivalent with regard to infectivity in rhesus monkeys (sequence analysis of the GP and VP24 genes showed no differences). For both experiments, EBOV was diluted in 1 ml of Hanks' buffered salt solution (Invitrogen) and administered by i.p. injection. Experiment 1 was performed in its entirety at biosafety level 4 (BSL-4). In experiment 2, because it was found that the expression of the EBOV proteins by HPIV3 did not increase vector replication or cause signs of disease in guinea pigs (7) and rhesus monkeys (experiment 1 of the present study), immunizations with the HPIV3 vectors were performed at BSL-2, and the challenge was performed at BSL-4 (CDC, Atlanta, GA). Moribund animals with symptoms of a severe EBOV infection that were found lying at the bottom of the cage were euthanized, as required by the CDC Animal Care and Use Committee-approved protocol.

Quantitation of HPIV3 and EBOV replication.

Shedding of the HPIV3 recombinants following immunization was monitored by tracheal lavage. Fifteen milliliters of sterile phosphate-buffered saline (PBS) was instilled into the trachea via a catheter positioned via a laryngoscope, and the tracheal lavage fluid was harvested. For quantitation of the HPIV3 titers, Vero E6 cell monolayers in 96-well plates were infected in quadriplicate with serial dilutions of the tracheal lavage fluids in Dulbecco's minimal essential medium with a high level of glucose (Invitrogen) containing 2% fetal bovine serum and were incubated for 6 days at 37°C. Thereafter, medium was removed from the wells, the cells were fixed for 10 min with 10% formalin in PBS, washed three times with PBS, and dried, and the plates were subjected to 2 Mrad radiation (⁶⁰Co) on ice. Each well was treated for 1 h at 37°C with 100 μl of 5% (wt/vol) of skim milk in PBS containing 0.1% Tween 20 and 1% normal goat serum. The cells were then incubated for 1 h at 37°C with a 1:200 dilution of serum from a guinea pig that had previously been immunized with HPIV3/EboGP+NP (7). Thereafter, the plates were washed three times with PBS containing 0.1% Tween 20 (wash buffer), 100 μl of a 1:1,000 dilution of goat anti-guinea pig immunoglobulin G (IgG)-horseradish peroxidase (HRP) conjugate (KPL, Gaithersburg, MD) was added into each well, and the plates were incubated for 1 h at 37°C and washed three times with the wash buffer. Wells infected with HPIV3 recombinants were detected by the addition of 100 μl of ABTS (2,2′-azino-di-[3-ethylbenzthiazolinesulfonate]) substrate (KPL) to each well, followed by incubation for 30 min at 37°C; the titers are expressed as TCID₅₀/ml.

Following the challenge, peripheral blood samples were taken, and the titer of EBOV in the plasma fraction was determined by limiting dilution using Vero E6 cells and expressed as TCID₅₀/ml. Wells positive for EBOV were identified in a manner similar to that used to identify wells containing HPIV3-infected cells (described above), except that the primary antibody used for the detection of infected monolayers was a 1:1,000 dilution of rabbit hyperimmune serum raised against all known species of EBOV and the secondary antibody was a goat anti-rabbit IgG (heavy and light chains)-HRP conjugate (KPL). In addition, in experiment 2, viremia was also evaluated by analysis of EBOV RNA in plasma by quantitative reverse transcription-PCR (RT-PCR) as previously described (29) using the following EBOV GP-specific primers: forward primer, 5′-GCTGCAGTGTCGCATCTAACA; reverse primer, 5′-TGGTTTGGTTGTGAGGGATTG; and TaqMan probe, 5′-6-carboxyfluorescein-CCCTTGCCACAATCTCCACGAGTCC-Dark Quencher.

Measurement of serum antibodies.

Serum antibody titers to HPIV3 were determined by hemagglutination inhibition as described previously (35). EBOV-specific IgA and IgG serum antibodies were measured by ELISA using inactivated purified EBOV virions as antigen as previously described (7), except that the secondary antibody was HRP-labeled goat antibodies specific to monkey IgG or IgA (KPL).

Measurement of cell-mediated responses.

To analyze EBOV-specific cell-mediated responses, peripheral blood mononuclear cells were isolated from peripheral blood using Vacutainer cell preparation tubes (BD Biosciences, San Jose, CA) according to the manufacturer's recommendations, suspended in fetal bovine serum containing 10% dimethyl sulfoxide, aliquoted, and placed in a nitrogen vapor freezer. Immediately before analysis, cells were thawed, washed twice with RPMI 1640 medium (Invitrogen) containing 10% bovine serum albumin, and resuspended in 0.5 ml of the same medium containing a pool of 14- to 21-amino-acid-long overlapping peptides specific to the EBOV GP protein (EBOV Zaire) at a combined final concentration of approximately 0.5 μM. In addition, antibodies specific to human CD28 (clone CD28.2) and CD49 (clone 9F10) (both from BD Biosciences) at 1 μg/ml each as stimulators of cytokine production and Golgi-Plug (BD Biosciences; final concentration of 1 μg/ml) were added. The cells were incubated for 6 h at 37°C and washed twice. Thereafter, the cells were simultaneously stained with the following antibodies (all from BD Biosciences): anti-human CD8 (clone RPA-T8) labeled with fluorescein isothiocyanate, anti-human CD4 (clone L200) labeled with perdinin-chlorophyll a protein, and anti-human CD3 (clone SP34-2) labeled with phycoerythrin. The cells were then washed twice. For intracellular staining of cytokines, the cells were fixed and permeabilized by incubation in Cytofix/Cytoperm solution (BD Biosciences) for 15 min at 4°C, washed twice with PermWash solution (BD Biosciences), and stained with anti-human gamma interferon (IFN-γ) antibodies (clone 4S.B3) labeled with allophycocyanin or similarly labeled anti-human tumor necrosis factor alpha (TNF-α) antibodies (clone MAB11) (both from BD Biosciences). After the staining, the cells were washed twice with PermWash solution and resuspended in 0.3 ml of PBS containing 2% bovine serum albumin. The cells were analyzed in a FACSCalibur flow cytometer (BD Biosciences) with at least 100,000 events run for each sample, and the statistical analysis was performed using FlowJo software (Tree Star, Ashland, OR).

Blood chemistry analysis.

Peripheral blood samples were collected into heparin-containing Vacutainer tubes (BD Biosciences), and whole-blood samples were analyzed using the Piccolo point-of-care chemistry and electrolyte system with comprehensive metabolic panel 400-0028 (Abaxis, Union City, CA).

RESULTS

Recombinant HPIV3 vectors.

Recombinant HPIV3 was engineered to express the EBOV glycoprotein GP (HPIV3/EboGP) alone or in combination with NP (HPIV3/EboGP+NP) (7). We also prepared an additional construct, HPIV3/EboGP+GM-CSF, that expresses EBOV GP in combination with the human cytokine GM-CSF as a possible means to augment the immune response without increasing the pathogenicity of the vaccine construct (5). These constructs were evaluated as topical respiratory tract vaccines against EBOV in juvenile rhesus monkeys in two experiments.

Experiment 1: response to immunization.

Eight rhesus monkeys in groups of two were immunized by the i.n. and i.t. routes with a total of 4 × 10⁶ TCID₅₀ of the control virus (empty HPIV3 vector) or with one of the three vaccine constructs, HPIV3/EboGP, HPIV3/EboGP+NP, or HPIV3/EboGP+GM-CSF (Table 1). On days 4 and 7 postimmunization, tracheal lavage fluid samples were collected to monitor replication of the HPIV3 recombinants. Each of the animals shed the immunizing virus on both days with viral titers between 3.0 and 4.7 log₁₀ TCID₅₀/ml (data not shown). The titers of the vaccine constructs were similar to that of the HPIV3 control, indicating that the expression of EBOV antigens did not detectably alter replication of the HPIV3 vector. None of the animals showed any sign of illness from the infection.

The serum antibody responses to the HPIV3 vector and to the expressed EBOV antigen(s) were determined on days 0, 21 (not shown), and 28 (Table 1) postvaccination. Strong HPIV3-specific serum antibody responses were detected in all animals, except that the magnitude of the response was substantially lower in animals immunized with HPIV3/EboGP+NP. The reduced response to the HPIV3 vector in this group could not be explained by differences in the magnitude of vector replication, and its basis is not understood. Analysis of sera by ELISA against inactivated EBOV detected substantial titers of EBOV-specific IgG antibodies following immunization (1:400 to 1:1,600 on day 28). As an indirect way to evaluate the mucosal immune response to vaccination (see Discussion), EBOV-specific serum IgA was analyzed. However, in this experiment, IgA response could not be detected at 21 (not shown) and 28 days postimmunization (Table 1).

Experiment 1: response to challenge.

On day 28 postimmunization, the animals were challenged with 1,000 TCID₅₀ of EBOV by the i.p. route. The 50% lethal dose in primates of this preparation of EBOV has not been determined, although experience from other experiments indicates that 1 PFU or TCID₅₀ of EBOV is equal to at least 100 lethal doses (unpublished data), and very low doses of EBOV cause uniformly fatal illness in primates (1).

The two control animals (animals 1 and 2) developed serious EBOV disease within 4 days postchallenge. By day 6 postchallenge, both control animals were moribund and were euthanized (Table 1). Of the two animals immunized with HPIV3/EboGP, one animal remained healthy, while one had no symptoms of infection except loss of appetite for a single day (animals 3 and 4, respectively). Of the two animals vaccinated with HPIV3/EboGP+NP, one had no symptoms except loss of appetite for a single day, while the other became ill on day 7 and died on day 8 (animals 5 and 6, respectively). Of the two animals vaccinated with HPIV3/EboGP+GM-CSF, one remained healthy, while the other had mild signs of disease (some petechiae and loss of appetite) on a single day (animals 8 and 7, respectively).

Peripheral blood was collected on days 4, 7 (or 6 for moribund animals 1 and 2) and 14 after EBOV challenge, and the plasma fraction was assayed for infectious EBOV by a limiting dilution assay. Infectious EBOV was detected at titers of 10² to 10⁴ TCID₅₀/ml in samples from each of the control animals on days 4 and 6 (before euthanasia) (Fig. 1A). In addition, a low level of EBOV (10² TCID₅₀/ml) was detected in blood from animal 6 on day 7, the day before its death (Fig. 1A). Infectious EBOV was not recovered from the other animals. EBOV antigen was detected by immunohistochemical analysis (22) of liver, lung, spleen, kidney, heart, testis, and lymph node tissue taken from animals 1, 2, and 6 on the day of their deaths (data not shown). The tissues of the remaining animals, which were sacrificed on day 28 postchallenge, were negative for EBOV antigen (data not shown). Analysis of peripheral blood for alanine aminotransferase, aspartate aminotransferase, and total bilirubin, which are markers of liver disease and typically are elevated during EBOV infection (reviewed in reference 28), showed increases in the latter two markers in the two control animals and in all three markers in animal 6, whereas an increase was not detected in any other immunized animal (Fig. 2). Similarly, substantial increases in the level of blood urea nitrogen and creatinine were found in the same three animals, indicating a disruption of normal kidney function, while no increase was found in any other animal (Fig. 2). Increases in the titers of EBOV-specific IgA and IgG serum antibodies were detected by day 14 or 16 postchallenge in all of the animals except for the two control animals that died on day 6 and animal 6 that died on day 8.

FIG. 1.

Detection of EBOV in peripheral blood samples from rhesus monkeys following EBOV challenge. Animal numbers (animals 1 to 8 or 9 to 16) are indicated at the top of the graphs. Blood was taken on the indicated days postchallenge in experiment 1 (A) and experiment 2 (B, C) and assayed for infectious EBOV titer determined by TCID₅₀ titration (A, B) or EBOV RNA detected by quantitative RT-PCR (C). For the samples in which EBOV was not detected, the value 1.5 log₁₀ TCID₅₀/ml were assigned. For the quantitative RT-PCR (QPCR) data, glyceraldehyde-3-phosphate dehydrogenase mRNA served as an endogenous reference molecule in calculating relative expression. X indicates that the monkey died after the last indicated blood sample was collected. Det. limit, detection limit.

FIG. 2.

Peripheral blood markers of EBOV disease in rhesus monkeys following EBOV challenge in experiments 1 (left) and 2 (right). Animal numbers (animals 1 to 16) are indicated at the top of the graph. ALT, alanine aminotransferase; AST, aspartate aminotransferase; TBIL, total bilirubin; BUN, blood urea nitrogen; CRE, creatinine. The values (in international units per liter or milligrams per decaliter as indicated) are shown for each animal on the indicated days postchallenge. Markers of liver function (*) and markers of kidney function (**) are indicated. X indicates that the monkey died after the last indicated blood sample was collected.

Experiment 2: response to immunization.

In the second experiment, we evaluated a single vaccine construct, HPIV3/EboGP, using a fivefold-higher immunizing dose and comparing one versus two immunizing doses (Table 1). One group of two control animals (animals 9 and 10) received the empty HPIV3 vector on day 0 and a second dose on day 28. A second group of three animals (animals 11 to 13) received a single dose of HPIV3/EboGP vaccine on day 0 and a dose of empty HPIV3 vector on day 28. A third group of three animals (animals 14 to 16) received two doses of HPIV3/EboGP administered on days 0 and 28. Replication of the HPIV3 constructs was monitored by viral titration of tracheal lavage fluid samples taken on days 4 and 7 after the first dose and on day 5 after the second dose (data not shown). Following the first dose, HPIV3 shedding was detected in all animals on day 4 with no consistent differences between groups, whereas on day 7, virus was detected only in animals 10, 11, 12, 13, and 14. On day 5 after the second dose, HPIV3 shedding was not detected in any animal.

A high titer of serum antibodies specific to the HPIV3 vector was detected in all animals on day 28. An increase was not observed following the second immunization (Table 1). EBOV-specific serum IgG and IgA were also detected beginning on day 13 in all animals (the titers on day 28 are shown in Table 1), indicating rapid development of the response. The detection of IgA following a single vaccine dose in this experiment but not in experiment 1 might reflect the fivefold-higher dose. Substantial increases in EBOV-specific IgG and IgA were observed in 2/3 and 3/3 animals, respectively, that received a second dose of HPIV3/EboGP (animals 14 to 16). At the time of challenge on day 67, the titers of serum IgG were 1:1,600 to 1:6,400 (animals 11 to 13) and 1:1,600 to 1:25,600 (animals 14 to 16), and the titer of serum IgA in each animal that received a single dose of HPIV3/EboGP versus the animals that received two doses was 1:100 and 1:1,600, respectively (Table 1). The increases in IgG and IgA antibodies specific to EBOV in each of the three monkeys that received two doses of HPIV3/EboGP indicates that the second dose of HPIV3/EboGP was immunogenic despite the lack of detection of replication of the vaccine vector in any of the animals.

We also analyzed the development of an EBOV-specific cell-mediated immune response following immunization in experiment 2. Peripheral blood mononuclear cells were isolated at days 13 and 28 following the first immunization and on day 13 following the second dose. The cells were stimulated in vitro with EBOV GP-specific peptides, stained for surface marker CD8 or CD4, fixed, permeabilized, stained for cytokine IFN-γ or TNF-α, and analyzed by flow cytometry (Fig. 3). CD8⁺ cells secreting IFN-γ were detected in most of the HPIV3/EboGP-immunized animals on days 13 and 28 and did not increase in abundance after the second dose (Fig. 3A). In contrast, CD8⁺ cells secreting TNF-α could not be detected above the background level (day 0) (not shown). CD4⁺ cells secreting IFN-γ were detected in several of the HPIV3/EboGP-immunized animals (Fig. 3B), and CD4⁺ cells secreting TNF-α were detected in a number of animals, including all three animals that received two vaccine doses (Fig. 3C).

FIG. 3.

Detection of EBOV-specific cell-mediated responses to immunization in experiment 2. Peripheral blood mononuclear cells were isolated on the indicated days after primary or secondary immunization, stimulated in vitro with peptides to GP, and analyzed by flow cytometry for the indicated markers. Values for individual animals (animals 9 to 16) are shown, expressed as percentages of the CD8⁺ or CD4⁺ lymphocyte populations.

Experiment 2: response to challenge.

On day 67 after the first immunization (39 days after the second dose), all of the animals were challenged with 1,000 PFU of EBOV by the i.p. route (this challenge dose is essentially equivalent to that in experiment 1 [see Materials and Methods]). By day 5 postchallenge, both of the control animals developed signs of severe EBOV disease and were euthanized on day 5 or 6 (Table 1). In the group that had received a single dose of HPIV3/EboGP, one animal (animal 11) was found dead on day 6 postchallenge without detectable signs of disease on the preceding days. Since viremia was detected in this animal on day 4 (see below and Fig. 1), most likely, development of clinical signs of the infection could have been observed only hours before the death. Another animal (animal 12) exhibited only mild symptoms on days 6 to 10 but rapidly recovered, and the third (animal 13) showed no signs of disease. In the group that had received two doses of HPIV3/EboGP, signs of disease were not observed in any of the three animals and all animals survived (Table 1).

Peripheral blood samples were collected on days 4, 7, 10, 16, and 28 postchallenge, and the plasma fractions were assayed for EBOV. A high-level viremia was detected in both of the control animals by virus titration (Fig. 1B) and by quantitative RT-PCR assay for EBOV RNA (Fig. 1C). A single dose of HPIV3/EboGP provided partial protection against challenge virus replication as evidenced by a decrease in the magnitude of viremia in two animals (animals 11 and 13) and the absence of viremia in the remaining animal (animal 12) (Fig. 1). Two doses of HPIV3/EboGP prevented viremia in each of the three recipient animals (animals 14 to 16). The two control animals and animal 11 had widespread distribution of EBOV antigen at the time of death in liver, lung, spleen, kidney, heart, testis, and lymph node tissue (data not shown). The remaining animals were sacrificed on day 28 and were negative for EBOV antigen (data not shown).

Analysis of peripheral blood for markers of liver and kidney disease showed strong increases in all parameters in both of the control animals except for blood urea nitrogen and creatinine in animal 9 (Fig. 2). In all of the animals that had received one dose of the HPIV3/EboGP, all five parameters were normal with the exception of a transient (days 7 and 10) increase in blood urea in animal 12. The lack of increase in these parameters in animal 11, which died on day 6 without previous symptoms of EBOV infection, might be indicative of a very rapid development of EBOV infection. In contrast, in the group that had received two doses of HPIV3/EboGP, all five parameters were normal in all animals (Fig. 2).

DISCUSSION

This study demonstrates that the immunization of primates via the mucosa of the respiratory tract with a respiratory virus-vectored vaccine confers substantial protection against EBOV, a viral agent of severe systemic hemorrhagic fever. The HPIV3 vector, based on a wild-type HPIV3 backbone, is naturally attenuated for rhesus monkeys and causes an asymptomatic infection, thus mimicking an attenuated vaccine virus. Therefore, the level of protective efficacy conferred by this asymptomatic infection in rhesus monkeys should be achievable in humans with a comparably attenuated paramyxovirus vector. Importantly, the HPIV3 vector expressing the EBOV GP protein was, like the HPIV3 vector, attenuated for the monkeys, which remained clinically well following infection with the HPIV3/EboGP recombinant virus. This was important since we had previously shown that the EBOV GP is incorporated into the HPIV3 vector particle (7). Interestingly, the HPIV3 vector containing EBOV GP acquired resistance to neutralization by HPIV3-specific antibodies as well as sensitivity to neutralization by EBOV-specific antibodies, indicating that the incorporated EBOV GP was active in initiating infection (7). The incorporation of functional GP had the potential to alter the pathogenesis of the vector and confer EBOV-like characteristics, since as has been shown with explanted human vessels, the protein is responsible for vascular cytotoxicity and injury (38). However, an in vivo study using the cynomolgus macaque model fully permissible for EBOV did not confirm that finding (18). Moreover, the two injectable adenovirus-vectored and VSV-vectored vaccine candidates expressing GP (in the second case, GP was incorporated into the vaccine particles) demonstrated no toxicity in monkeys (20, 33); recent clinical trials of a DNA vaccine expressing GP revealed no toxicity (24). Similarly, in the present study in primates and in a previous study in guinea pigs (7), there was no apparent change in the efficiency of replication of the HPIV3 vector due to the presence of GP, and the infection was completely asymptomatic even when administered at a dose that was extremely high compared to a lethal dose of EBOV. Testing the liver, spleen, kidney, and blood samples of guinea pigs infected with HPIV3/EboGP demonstrated the lack of detectable virus in these organs/tissues (A. Bukreyev et al., unpublished data). These findings provide a precedent in which incorporation of a functional foreign attachment/penetration protein from a highly virulent pathogen into a vector particle had no apparent effect on the replication or pathogenesis of the vector. Even if more detailed future studies demonstrate possible subtle changes in infection by the vector expressing GP, the vaccine construct can be easily modified by replacement of the GP with its modified form lacking cytotoxicity in vitro (32, 39).

Despite the attenuated nature of the replication of the HPIV3/EboGP recombinant virus in the monkeys, one or two doses of vaccine delivered to the respiratory tract induced moderate to high levels of antibody to EBOV. A single dose of HPIV3 expressing EBOV GP induced a titer of serum IgG antibody of approximately 1:500 to 1:1,600, whereas two doses increased the titer to approximately 1:6,000. We did not specifically measure EBOV-specific mucosal IgA levels in the monkeys, but the level of serum EBOV-specific IgA was determined, since serum IgA responses during acute respiratory tract infection can be an indirect measure of a mucosal immune response to vaccination (2, 8). A single dose of HPIV3/EboGP recombinant vaccine virus induced only a weak serum IgA EBOV antibody response, but a 16-fold increase occurred following the second dose. This showed that, even though replication of the second dose of HPIV3/EboGP infected was highly restricted as evidenced by the lack of detectable vaccine virus shedding, the virus infected sufficiently well to induce substantial boosting of EBOV-specific IgG or IgA antibodies. The substantial boosting of EBOV-specific IgG and IgA occurred even though each monkey had developed antibodies after the first dose of vaccine to both the HPIV3 and EBOV GP attachment proteins present in the virions. Perhaps the presence of two independent modes of entry, mediated by HPIV3 HN and F or by EBOV GP, increased the ability of the vector virus to initiate an infection in a partially immune host. This is presently under investigation.

EBOV-specific cell-mediated immune responses were also documented in monkeys infected with the HPIV3/EboGP recombinant vaccine virus. Responses were measured by analyzing peripheral blood for CD4⁺ and CD8⁺ cells that secreted IFN-γ or TNF-α following in vitro stimulation with peptides specific to EBOV GP. EBOV GP-specific T cells in both the CD4⁺ and CD8⁺ cell populations were present in animals immunized with one or two doses of vaccine. Sampling from the peripheral blood was not considered the optimal method to quantitate the CD4⁺ and CD8⁺ T-cell responses, since peripheral blood cells typically contain a much lower frequency of activated T cells than do cells isolated from lungs following infection with respiratory viruses (10), but the design of the experiment precluded sacrificing monkeys to obtain lung T cells. Thus, it is likely that the pulmonary T-cell response was of greater magnitude than that detected in the peripheral blood. The T-cell responses tended to remain elevated on day 28 following immunization and thus may have been sufficiently long-lived to play a role in protective immunity against challenge, which occurred 28 and 39 days following vaccination in experiments 1 and 2, respectively.

We also sought to determine whether it was possible to augment the immunogenicity of the EBOV GP expressed by recombinant HPIV3 vector by coexpression of GM-CSF, since this cytokine can act as a immunological adjuvant (36). In the present study, expression of GM-CSF did not provide a significant increase in the induction of serum antibodies, cellular immunity, or protective efficacy. This supports a previous study with PIV3 vectors in rhesus monkeys in which an increase in immunogenicity associated with GM-CSF was observed only with a highly attenuated version of HPIV3 (5), but not with wild-type HPIV3 (unpublished data). These findings suggest that, for a live virus vector in nonhuman primates, the adjuvant effect of GM-CSF might be limited to versions of the vector that are highly attenuated.

One or two doses of the HPIV3/EboGP recombinant vaccine virus protected monkeys against challenge with a high dose (1,000 infectious units) of homologous EBOV Zaire administered intraperitoneally. A single dose of vaccine protected seven of nine monkeys against EBOV, which was uniformly fatal in the nonimmunized control monkeys. Protection was accompanied by decreased replication of challenge virus and decreased signs of disease, including a reduction of chemical markers of hepatic and renal disease. Of the nine animals that received a single immunizing dose, the two that died developed EBOV infections indistinguishable from those of control animals, despite the fact that they developed specific immune responses against EBOV. The reasons why these responses failed to prevent fatal EBOV disease in these two animals and how they differed from those induced in protected animals are unclear. There was a reasonably good correlation between the magnitude of the virus-specific serum IgG titer and protection against death caused by EBOV. A titer of 8.6 to 10.6 log₂ was protective against death in seven of nine animals (78%), and a higher titer (12.6 to 14.6 log₂) was protective in three of three animals. In addition, a virus-specific serum IgA titer of 10.6 log₂, which was induced after the second dose, was associated with complete protection (against death and signs of disease), although the effect could not be cleanly dissociated from that of serum IgG given the small number of animals. In contrast, there was no obvious association between the level of virus-specific peripheral T lymphocytes and protection. An increase in protective efficacy was not apparent using another HPIV3 expressing both EBOV GP and NP, indicating that this dual-antigen virus did not have an apparent advantage over constructs expressing only GP. Importantly, two doses of the HPIV3/EboGP recombinant vaccine virus afforded complete protection to each of the three immunized animals against disease and detectable virus replication. However, each of these animals was presumably infected at a low level with the EBOV challenge virus, since each developed increases in serum IgA and IgG antibodies to the EBOV GP antigen, and it is unlikely that the input EBOV antigen would have been sufficient to be immunogenic on its own. Thus, it is possible to provide complete protection against EBOV disease in nonhuman primates by delivery of vaccine to the respiratory tract, a route of immunization that will greatly facilitate outbreak control during epidemics in areas where EBOV is endemic.

Three vector systems have now been shown to be protective against EBOV in primates, namely, vectors based on adenovirus (33), VSV (20), and HPIV3 (7). None of these systems is ideal in its present form. Adenovirus vectors expressing the EBOV GP presently are based on nonreplicating versions of adenovirus 5, which is a common human pathogen (reviewed in reference 34). The extensive seroprevalence in the adult population resulting from infection with this common virus likely will interfere with the extent of infection with the adenovirus type 5 vector and thus reduce the overall expression and immunogenicity of the foreign antigen encoded by the virus (reviewed in reference 34). This might be overcome by using vectors based on other adenoviruses that are not common human pathogens or by relying on a very large vaccine dose of the adenovirus type 5 vaccine virus to overcome vector immunity. VSV is not a common pathogen, is infectious at reasonable doses, and can readily be engineered, attenuated, and grown in tissue culture cells in vitro as a candidate live vaccine vector. However, there is little or no experience with its administration to humans, and central nervous system involvement with wild-type VSV has been described in rodents and in at least one human case (14, 27). Thus, safety in humans remains to be established, although VSV vectors that are attenuated in nonhuman primates in vivo have been identified (25). Whereas adenovirus and VSV vectors are usually administered by injection (20, 33), HPIV3 is naturally tropic for the respiratory tract and is readily administered to humans by the i.n. route (reviewed in reference 9). As already noted, the virus replicates at the surface of the respiratory tract and has little propensity to spread beyond that site, which is probably an advantage for vaccine safety (9, 40). This results in the induction of an effective mucosal immune response in the respiratory tract in the context of a self-contained infection that, in case of vaccination against EBOV or any pathogen capable of infection by the respiratory route, is important for protection against transmission by aerosol, as might be the route of exposure in an act of bioterrorism. The protective efficacy of HPIV3/EboGP against aerosolized EBOV will be tested in a separate study. In addition, numerous vaccine candidates based on HPIV3 or related respiratory paramyxoviruses have been evaluated or continue to be evaluated in clinical trials, and there are considerable data on their safety, including the lack of significant shedding, and means of attenuation (12, 30). One limitation of the HPIV3 vector strategy is that it too is a common human pathogen, and the widespread seroprevalence to this virus in adults would likely reduce its effectiveness. It is possible that the presence of EBOV GP in the HPIV3 particle, and the resistance it confers to neutralization by HPIV3-specific antibodies, will mediate limited infection of HPIV3-immune humans. Our second dose of vaccine is able to induce a booster immune response in monkeys that developed vector immunity following the first dose. Indeed, our preliminary data using the guinea pig model indicate that guinea pigs immune to HPIV3 can be immunized with HPIV3 expressing the EBOV GP. As an alternative strategy, we are developing vectors based on animal paramyxoviruses, such as Newcastle disease virus, that are not common pathogens of humans, should not have extensive seroprevalence in the human population, and produce little or no shedding following administration through the respiratory tract (4).