Seroprevalence study on Ebola Virus in patients in Uganda

Örebro University

School of Health and Medical Sciences

Biomedicine and Method in Medical Diagnosis

Experimental Medicine

Medical Degree Project in Medicine 45 Credits 15/6/2015

Seroprevalence study on Ebola Virus

in patients in Uganda

Author: Nicola Änäkkälä

Supervisor: Prof. Dr. Tomas Bergström

Abstract

The Ebola virus (EBOV) causes outbreaks of hemorrhagic fever (HF) in Central and Western Africa. Antibodies against EBOV envelope glycoprotein (GP) correlate with survival. Our aim was to develop two ELISA assays using EBOV/Sudan GP (aa 1–647) and EBOV/Zaire GP (aa 1–647) as antigens for seroprevalence studies. We analyzed sera from 67 Ugandan patients, of whom 30 had a diagnosed EBOV infection. Swedish blood donors (n=40) functioned as controls. Higher antibody prevalence to EBOV/Sudan was seen in the Ugandan patients with a history of HF than in the non-EBOV patients. The Ugandan patients had a higher seroprevalence of IgG against EBOV/Sudan than against EBOV/Zaire. A serological cross-reactivity between the two EBOV antigens was discovered, explained by extensive homologies of their amino acid sequences, shown by BLAST analysis. Several Ugandan samples from patients without a history of EBOV were seropositive to EBOV, suggesting the existence of subclinical infections.

Abbreviation list:

1. EBOV: Ebola virus 2. NP: Nucleoprotein 3. VP: Viral protein 4. GP: Glycoprotein

5. ZEBOV: Zaire Ebola virus 6. SEBOV: Sudan Ebola virus 7. REBOV: Reston Ebola virus

8. CIEBOV: Côte d’Ivoire Ebola virus 9. BBEBOV: Bundibugyo Ebola virus 10. EBV: Epstein-Barr virus

11. HHV-4: Human herpes-virus 4 12. IgG: Immunoglobulin G 13. HF: hemorrhagic fever

14. BSL-4: Biosafety Level 4 laboratory 15. PCR: polymerase chain reaction 16. CHO: Chinese Hamster Ovary

17. IMAC: Immobilized metal ion affinity chromatography 18. ELISA: Enzyme-linked immunosorbent assay

19. WB: Western Blot 20. OD: Optical Density 21. Fig: figure

22. NCBI: National Center for Biotechnology Information 23. A.a.: Amino acid

Contents:

1. Introduction………..

1

1. 1. Ebola viruses (EBOV) .………...1

1. 2. Genogroups of EBOV………..1

1. 3. Outbreak of SEBOV in the District of Gulu in Uganda ...…….2

1. 4. Immunoreactivity ...……….,………...3

1. 5. Epstein-Barr virus (EBV) ...………....3

1. 6. Aims ...………....4

2. Materials and Methods ………4

2. 1. Human serum samples ...………4

2. 2. EBOV GP Antigens ...………5

2. 3. EBV gp-350 Antigens ....………....5

2. 4. Indirect ELISA Assay ………....5

2. 5. BLAST analysis and O-glycan prediction ...……….6

2. 6 Statistics………...7

3. Results...7

3. 1. Seroassays ..………7

3. 2. Compilation of seroprevalence ..………...10

3. 3. Cross reactivity ..…..……….11

3. 4. SEBOV G sequencing and O-glycan prediction ……….12

4. Discussion ………13

5. Conclusion ...………15

6. Acknowledgments ...………16

1

1. Introduction

1.1. Ebola viruses (EBOV)

The EBOV and the Marburg virus are causative agents of hemorrhagic fevers (HF) and members of the filoviridae family that all have a nucleocapsid, a polymerase complex, a matrix protein and an envelope surrounding the virus [1, 2]. The virus genome consists of a single stranded negative RNA that encodes 7 proteins in form of nucleoprotein (NP), VP35, VP30, sGP, VP24, VP40 and RNA-dependent RNA polymerase (L). Proteins controlling replication and transcription of the viral RNA are NP, VP35, VP30 and RNA-dependent RNA polymerase (L). The viral entry is regulated by glycoprotein (GP) and the assembly and the release of the virions is controlled by VP40 and VP24 [3, 4]. The structure of the virion and its genome and encoded proteins are shown in Fig. 1.

Fig. 1. A Morphology of EBOV by electron microscopy. B. Localisation of the 7 encoded

EBOV proteins on the virion. C. Organisation of the EBOV genome.

EBOVs are of many genotypes that are clustered in at least five genogroups, which are Zaire Ebola virus (ZEBOV), Sudan Ebola virus (SEBOV), Reston Ebola virus (REBOV), Côte d’Ivoire Ebola virus (CIEBOV), and Bundibugyo Ebola virus (BBEBOV). These viruses, although being genetically related and causing HF, all have different case-fatality rates

2

during their epidemics [5 - 10]. EBOVs are thought to be spread from animals such as monkeys or bats to humans but these viruses are also spread from human to human through contacts with body fluids such as infected blood, sweat and genital secretions [10].

1.3. Outbreak of SEBOV in the District of Gulu in Uganda

The outbreak of SEBOV in the District of Gulu in Uganda occurred in year 2000 to 2001, with an incidence of 425 cases and 224 deaths (fatality rate 53%). This strain of EBOV was closely related to a virus causing outbreak in Sudan in year 1976 to 1979, which also had a high fatality rate of about 53% [11, 12].

Fig. 2. Map of Uganda and the areas (red spots) that were affected by the outbreak of

SEBOV disease in the year 2000 to 2001. Subjects who were regarded as suspected to be positive with SEBOV disease were diagnosed according to report (notification scheme) from the laboratory. Clinical symptoms included bleeding, fever and three or more specified symptoms (headache, vomiting, anorexia, diarrhea, weakness or severe fatigue, abdominal pains, body aches or joint pains, difficulty in breathing, hiccups as well as unexplained deaths [13].

3

The Ugandan serum samples were collected in September of year 2000 in the District of Gulu during the outbreak of SEBOV disease. Gulu District is located in the Northern Uganda and has an area of 3.452 km2 with a population of about 396.600 (2012) with an increase of 2.9% since year 2002. It has a population density of 114.9/km2 (298/sq milo). The travel distance is 340 km by road from Gulu District to the capital city of Uganda, Kampala [14].

1.4. Immunoreactivity

One previous study has shown that serum samples drawn from survivors of EBOV infection after 10 years follow-up contained virus-specific antibodies that might neutralize the virus and this neutralization correlated to immunoreactivity against viral proteins such as NP, VP30, and GP1-649. The same study demonstrated that patients who did not survive had little or no antibody reactivity [15]. Another study also confirmed that the greatest seroreactivity was directed to viral proteins NP and GP1-649 and later to VP40, which showed a high correlation of immunorecognition of antibodies towards these four viral proteins of SEBOV to survival of human victims [16]. It was further shown in another study that there were persistent immune responses in the survivors after the EBOV disease in form of significant increased levels of cytokine expression that was associated with regulators of the pre-inflammatory and T-cell-derived immune response [17]. Anti- EBOV immunoglobulin-G (IgG) antibodies was shown in some serosurveillance studies in humans to be present and directed to viral proteins such as NP, VP40, VP35, and sGP of EBOV [14, 18 - 20]. The present seroprevalence study aimed at investigating the presence of IgG antibodies against SEBOV and ZEBOV in serum samples from Ugandan patients with and without a diagnosis of EBOV infection, and used Swedish blood donors as negative controls.

1.5. Epstein-Barr virus (EBV)

EBV gp350 antigen expressed in mammalian cells at the Proteomics Core Facility at the University of Gothenburg (Sweden) was used as a quality control to investigate the presence of IgG antibodies in the Ugandan serum samples collected during the outbreak of SEBOV disease in year 2000 in Gulu District. EBV, or human herpes-virus 4 (HHV-4), is a ubiquitous virus that causes mononucleosis and, in addition, Burkitt´s lymphoma in African

4

children [18]. EBV gp350 is the receptor-binding protein of the virus and it was used as control antigen in the study since most individuals are expected to be seropositive.

1.5.

 To develop serological assays to SEBOV, ZEBOV, and EBV based on their major envelope GP

 To assay seroprevalence of these three viruses in patient sera from Uganda, with and without a history of Ebola infection, and from Swedish blood donors

 To compare amino acid sequences of SEBOV GP to ZEBOV GP using BLAST and O-glycan prediction database

2. Materials and Methods

2.1. Human serum samples

The 75 serum samples from Uganda were obtained from the biobanked collection in collaboration with the Uganda Virus Research Institute at Entebbe which is handling diagnostics of EBOV outbreaks. The 67 serum samples analyzed were derived from 30 patients who had a diagnosis of EBOV (EBOV disease patients) and 37 who had other diseases and who were EBOV-negative (Non-EBOV disease patients). There were 50 females (25 EBOV-negative and 25 EBOV-positive) and 17 males (12 EBOV-negative and 5 EBOV-positive). The remaining 8 serum samples from the 75 serum samples were not analyzed due to lack of identification or diagnosis but only having the date of collection and list number. The information on ages of these subjects was not provided but a list on the date at which each serum sample was collected was provided. All the serum samples were collected about three weeks after the outbreak of SEBOV disease in the year 2000 within the District of Gulu in the Northern part of Uganda [13]. Ugandan serum samples were collected and prepared under ethical policies used by the virology department of Uganda Virus Research Institute at Entebbe [17]. To ensure non-infectiousness, all serum samples from Uganda were heat inactivated for 30 min at 56oC before analysis. As negative controls, 40 coded Swedish human serum samples from blood donors at Sahlgrenska University

5

Hospital in Gothenburg, Sweden were used. All samples were coded and anonymous under investigation, so no ethical permission was required.

2.2.

EBOV GP Antigens

Two purchased viral proteins used as antigens in the study, SEBOV GP (Ebola/Sudan - Nakisamata) aa 1-647 (Genbank No. AFP2823) and ZEBOV GP (Ebola/Zaire) aa 1-647 (Genbank No. AFP2823) was produced by Immune Technology Corp. These antigens were earlier used as antigens for WB and ELISA by others. Rabbit antibodies (Immune Technology Corp) against the latter antigen were also purchased. Both antigens were purified from recombinant protein expressed in 293 cell cultures by the same company.

2.3.

EBV gp-350

Antigens

As control antigens to independently check for IgG antibody reactivity in all samples we used a recombinant expression system of a truncated form (a collection of DNA constructs encoded by vectors to different parts of gp-350/220 for expression of a full fragment of aa 1-860) of EBV envelope glycoprotein gp350. This is a ubiquitous virus to which almost all adult subjects carry IgG antibodies why it was suitable as control antigen. EBV-350 antigen aa 1-860 was expressed in Chinese Hamster Ovary (CHO) cells and prepared and purified at the Proteomics Core Facility at the University of Gothenburg. The preparation and the purification were done using protein with proper glycosylation of this mucin-like protein, which was obtained after expression and production in mammalian cells in form of CHO-S cells. The purification of the large batch of gp350 aa1-860 recombinant protein was done using immobilized metal ion affinity chromatography (IMAC). The analysis of purified gp350 aa1-860 was performed on ELISA and Western blot (Hasi G., Biochemistry Department, University of Gothenburg, Sweden) [21, 22].

2.4. Indirect ELISA Assay

Serial dilution of antigens to positive and negative control sera in an indirect ELISA assay was used to determine a suitable antigen concentration and the chosen suitable concentrations were as follows 1µg/ml for EBV antigens, 2µg/ml for SEBOV antigens and 2µg/ml for ZEBOV antigens. The 96 well plates (F96 Maxisorp NUNC-immuno plate

6

(Thermo scientific)) were used in all the seroassays. These plates were coated with the above defined concentrations of SEBOV, ZEBOV and EBV GP antigens added into carbonate buffer 0.05M pH 9.6 (Bacteriology laboratory, Sahlgrenska University hospital) and incubated at 4oC for 24h. The plates were washed x 3 (Molecular Devices, SkanWasher (Molecular Dervices Corp, USA)) with PBS solution containing 0.05% Tween 20 after the incubation and then blocked with the dilution buffer of 2% milk in PBS at room temperature for 60 min in order to prevent unspecific binding. The blocking buffer was discarded. Then 1µl of primary antibody (human sera or rabbit serum control) in 200µl dilution buffer [1% milk in PBS]. Each sample was analyzed in duplicates. The plates were incubated in 37oC for 90 min and later washed x 3. Secondary antibodies in form of alkaline phosphatase-conjugated goat-anti-human-IgG (Jackson ImmunoResearch) for human sera or alkaline phosphatase-conjugated goat-anti-rabbit-IgG (Jackson ImmunoResearch) for rabbit sera, diluted 1/1000 in dilution buffer [1% milk in PBS], were added in the wells and incubated at 37oC for 90 min. After the incubation period the plates were washed 2 x 3 to ensure through washing of all the walls. Substrate was added to the wells in form of 1mg/ml alkaline phosphatase (Sigma) in DEA – buffer (Bacteriology laboratory, Sahlgrenska University hospital). The plates were incubated at room temperature between 15 - 45min depending on the color reaction in the plates. Finally, developed plates were read in a spectrophotometer (Thermo Scientific Multiskan FC (Thermo Scientific)) at 405nm and 620nm. Measurements, for each reaction was calculated as follows: Optical density (OD) values obtained at 405nm minus background OD values at 620nm. The cut-off between positive and negative reactions was calculated as the mean of negative control on each plate + 0.2 OD units. The same assay procedure was repeated to all tests.

2.5. BLAST analysis and O-glycan prediction

A.a. sequences of the SEBOV GP and ZEBOV GPs were retrieved from Pubmed and analyzed by BLAST program (http://blast.ncbi.nlm.nih.gov/Blast.cgi) to search for homologies, and by a program for O-glycan prediction available at

7

2.6. Statistics

Statistics used was of Prism Software, corrected for multiple comparisons – results of one-way ANOVA and Tukey multiple comparison tests comparing every group with the other group (1995-2015 GraphPad Software, Inc. (Prism Software)).

3. Results

The OD values in ELISA from Ugandan serum samples of EBOV disease (n = 30), from other diseases (n=37) and from Swedish blood donors (n = 40) were measured using the average of duplicates of each sample’s OD values in all the seroassays. The results from the three patient groups to the three antigens are presented in three graphs: seroprevalence to SEBOV in Fig. 3, to ZEBOV in Fig. 4 and to EBV in Fig. 5. All seroprevalences are compiled in Table 1. The results were statistically analyzed using GraphPad Software of one-way ANOVA and Tukey multiple comparison. The data on cross-reactivity of IgG between SEBOV vs ZEBOV) are presented in Fig. 6, and one-way ANOVA and Tukey multiple comparison was also used for the statistics. The results from the BLAST analysi s and O-glycan prediction of amino acid sequences from SEBOV GP and ZEBOV GP are presented in Fig. 6.

3.1. Seroassays

SEBOV IgG: The SEBOV antigen worked well in the ELISA, and the cut-off, based on the mean of negative controls + 0.2 OD units was regularly 0.3. Maximum signal was 1.0 and the signal vs. background ratio was judged as more than acceptable for this ELISA test. The results (see Fig 3.) were the following: 40% (12/30) were positive for SEBOV IgG of the EBOV disease patients, 19% (7/37) of the Non-EBOV disease patients, while 2.5% (1/40) was seropositive of the Swedish blood donors. SEBOV is prevalent since long in Uganda and the results are compatible with a former asymptomatic infection of some of the non-EBOV disease patients. We lack data on the non-EBOV patients, especially how long after onset samples were drawn, why we cannot explain why the majority of this group remained seronegative. Likewise, the Swedish blood donors were run anonymous, why we lack data on the seropositive subject.

8

Fig. 3. OD values of seroreactivity to the SEBOV GP (aa 1- 647) antigen of serum samples

collected from confirmed EBOV disease patients of Ugandan; serum samples from Non-EBOV disease patients of Uganda; and of Swedish blood donors. The data are presented as individual sample results and median values. The horizontal black line denotes the cut-off. The results of EBOV disease patients and Non-EBOV disease patients from Ugandan sera demonstrated higher seroprevalences of IgG antibodies against SEBOV antigen in both groups as compared to Swedish blood donors. The results were the following: 40% (12/30) were positive for SEBOV IgG of the EBOV disease patients, 19% (7/37) of the Non-EBOV disease patients, while 2.5% (1/40) was seropositive of the Swedish blood donors.

ZEBOV IgG: The ZEBOV antigen worked less well in the ELISA, and the cut-off, based on the mean of negative controls + 0.2 OD units was approximately 0.5. Maximum signal was 1.0 and the signal vs. background ratio was judged as acceptable for this ELISA test. The results (see Fig 4.) were the following: 10% (3/30) positive for ZEBOV IgG in the EBOV disease patients, 11% (4/37) in Non-EBOV disease patients, while 2.5% was seropositive of

9

the Swedish blood donors (1/40). ZEBOV is not prevalent in Uganda and the results are compatible with a serological cross-reaction with SEBOV. This question is addressed in the BLAST analysis (see below).

Fig. 4. OD values of seroreactivity to the ZEBOV GP (aa 1- 647) antigen of serum samples

collected from confirmed EBOV disease patients of Ugandan; serum samples from Non-EBOV disease patients of Uganda; and of Swedish blood donors. The data are presented as individual sample results and median values. The horisontal black line denotes the cut-off. The seroprevalence results were the following: 10% (3/30) positive for ZEBOV IgG in the EBOV disease patients, 11% (4/37) in Non-EBOV disease patients, while 2.5% was seropositive of the Swedish blood donors (1/40).

EBV gp-350 IgG: The EBV antigen worked excellent in ELISA with an extremely low background and the cut-off, based on the mean of negative controls + 0.2 OD units was 0.25. Maximum signal was 4.0 and the signal vs. background ratio was judged as optimal.

10

This was especially beneficial since this was the only antigen that we produced ourselves for an ELISA test. The following seroprevalences of EBV-IgG were found: 100% (30/30) in EBOV disease patients, 95% (35/37) in Non-EBOV disease patients and 90% (36/40) in Swedish blood donors.

Fig. 5. OD values from sera tested on EBV gp-350 (aa 1-860) antigen with the same three

groups of EBOV disease patients, Non-EBOV disease patients and Swedish blood donors. The data are presented as individual sample results and mean values. The horisontal black line denotes the cut-off. The following high seroprevalences of EBV-IgG were found: 100% (30/30) in EBOV disease patients, 95% (35/37) in Non-EBOV disease patients and 90% (36/40) in Swedish blood donors.

11

Here, we have compiled the data from all three assays into one table (Table 1). The data shows that the EBOV patients had the highest seroprevalence to SEBOV, which was logic since they most likely were infected with this virus. On the other hand, the high seroprevalence to EBV in both Ugandan groups of patients indicated that the samples were of high enough quality (i.e. not destroyed during transport or heat activation) to allow for IgG analysis. This means that the negative results for SEBOV IgG among the EBOV patients probably were correct.

Table 1. Compilation of seroprevalence to all three antigens in the three study groups.

Values are denoted as number of sero-positive subjects/total (%).

Antigens EBOV disease patients %

Non-EBOV disease patients %

Swedish blood donors %

SEBOV 40 19 2.5

ZEBOV 10 11 2.5

EBV 100 95 90

3.3. Cross reactivity

As mentioned above, the results from the ZEBOV ELISA was most likely explained by serological cross reactivity between the two EBOV antigens, with samples showing similar OD values in the two assays, presented in Figure 6 by mean, SD, n = 67. The subjects with a red circle around are EBOV disease patients and those subjects with a black circle around are patients (Non-EBOV disease patients) with other diagnoses. The star mark is for the Swedish blood donor subject who demonstrated antibody reactivity towards both viruses. The pink round circle marks the subjects that fall within the area of high cross reactivity between these two seroassays. The two lime color lines mark the possible diversity between the two seroassays with EBOV disease patients reacting preferably with SEBOV and other patients with ZEBOV antigens. The green color line marks regression line between the seroassays. The correlation was significant between the results of the SEOBV and ZEBOV

12

seroassays, as demonstrated by Pearson r: r of 0.2987, 95% confidence interval of 0.06297 to 0.5028, R squared of 0.08920, P value of P(two-tailed) being 0.014, a P value summary of *, Significant (α = 0.05) Yes and Number of XY Pair of 67.

Fig. 6. Results from SEOBV vs ZEOBV GP seroassay demonstrated possible cross reactivity

between these two antigens, with samples showing similar OD values in the two assays, presented by mean, SD, n = 67. The subjects with a red circle around are EBOV disease patients and those subjects with a black circle around are patients (Non-EBOV disease patients) with other diagnoses. The star mark is for the Swedish blood donor subject who demonstrated antibody reactivity towards both viruses. The pink round circle marks the subjects that fall within the area of high cross reactivity between these two seroassays. The two lime color lines mark the possible diversity between the two seroassays with EBOV disease patients reacting preferably with SEBOV and other patients with ZEBOV antigens. The green color line marks regression line between the seroassays. The correlation was significant between the results of the SEOBV and ZEBOV seroassays, as demonstrated by

13

Pearson r: r of 0.2987, 95% confidence interval of 0.06297 to 0.5028, R squared of 0.08920, P value of P(two-tailed) being 0.014, a P value summary of *, Significant (α = 0.05) Yes and Number of XY Pair of 67.

3.4. BLAST analysis and O-glycan predicition -SEBOV GP vs. ZEBOV GP

As shown in Figure 7, the first 185 a.a. are highly similar, indicated that EBOV are generally genetically conserved. This region might harbor several cross reactive epitopes. Then follows a region of less homology (up to a.a. 305) whereafter a mucin-like domain with numerous predicted O-glycosylation sites follows. Most likely, type-specific epitopes may be located here. Lastly, a.a. 533-647 are highly similar. The results indicate that the homologous regions explain, at least partly, the serological cross reactivity between the two antigens. 647 (http://blast.ncbi.nlm.nih.gov/Blast.cgi) Query 6 LLQLPRDKFRKSSFFVWVIILFQKAFSMPLGVVTNSTLEVTEIDQLVCKDHLASTDQLKS 65 +LQLPRD+F+K+SFF+WVIILFQ+ FS+PLGV+ NSTL+V+++D+LVC+D L+ST+QL+S Sbjct 6 ILQLPRDRFKKTSFFLWVIILFQRTFSIPLGVIHNSTLQVSDVDKLVCRDKLSSTNQLRS 65 Query 66 VGLNLEGSGVSTDIPSATKRWGFRSGVPPKVVSYEAGEWAENCYNLEIKKPDGSECLPPP 125 VGLNLEG+GV+TD+PSATKRWGFRSGVPPKVV+YEAGEWAENCYNLEIKKPDGSECLP Sbjct 66 VGLNLEGNGVATDVPSATKRWGFRSGVPPKVVNYEAGEWAENCYNLEIKKPDGSECLPAA 125 Query 126 PDGVRGFPRCRYVHKAQGTGPCPGDYAFHKDGAFFLYDRLASTVIYRGVNFAEGVIAFLI 185 PDG+RGFPRCRYVHK GTGPC GD+AFHK+GAFFLYDRLASTVIYRG FAEGV+AFLI Sbjct 126 PDGIRGFPRCRYVHKVSGTGPCAGDFAFHKEGAFFLYDRLASTVIYRGTTFAEGVVAFLI 185 Query 186 LAKPKETFLQSPPIREAVNYTENTSSYYATSYLEYEIENFGAQHSTTLFKINNNTFVLLD 245 L + K+ F S P+RE VN TE+ SS Y ++ + Y+ FG + LF+++N T+V L+ Sbjct 186 LPQAKKDFFSSHPLREPVNATEDPSSGYYSTTIRYQATGFGTNETEYLFEVDNLTYVQLE 245 Query 246 RPHTPQFLFQLNDTIHLHQQLSNTTGKLIWTLDANINADIGEWAFWENKKNLSEQLRGEE 305 TPQFL QLN+TI+ + SNTTGKLIW ++ I+ IGEWAFWE KKNL+ ++R EE Sbjct 246 SRFTPQFLLQLNETIYTSGKRSNTTGKLIWKVNPEIDTTIGEWAFWETKKNLTRKIRSEE 305 Query 306 LSFETLSLNETEDDDATSSRTTKGRISDRATRKYSDLVPKDSPGMVSLHVPEGETTLPSQ 365 LSF +S + +RT+ ++ T + + ++S MV +H Sbjct 306 LSFTAVSNRAKNISGQSPARTSSDPGTNTTTEDHKIMASENSSAMVQVH--- 354 Query 366 NSTEGRRVDVNTQETI-TETTATIIGTNGNNMQISTIRTGLSSSQILSSSPTMAPSPETQ 424 ++GR V+ T+ T +T+ T ST T + I ++ T Sbjct 355 --SQGREAAVSHLTTLATISTSPQPPTTKPGPDNSTHNTPVYKLDISEATQAEQHHRRTD 412 Query 425 TSTTYTPKLPVMTTEEPTTPPRNSPGSTTEAP---TLTTPENITTAV---KTVLPQE 475 +T + P MT P + T+ P T T+P+N + +E Sbjct 413 NDSTTSDTPPAMTAAGPPKAENTNTSKGTDLPDPATTTSPQNHSETAGNNNTHHQDTGEE 472 Query 476 STSN---GLITSTVTGILGSLGLRKRSRRQVNTRATGKCNPNLHYWTAQEQHNAAGIAWI 532 S S+ GLIT+T+ G+ G + +R+RR+ A KCNPNLHYWT Q++ A G+AWI Sbjct 473 SASSGKLGLITNTIAGVAGLITGGRRTRREAIVNAQPKCNPNLHYWTTQDEGAAIGLAWI 532 Query 533 PYFGPGAEGIYTEGLMHNQNALVCGLRQLANETTQALQLFLRATTELRTYTILNRKAIDF 592 PYFGP AEGIYTEGLMHNQ+ L+CGLRQLANETTQALQLFLRATTELRT++ILNRKAIDF Sbjct 533 PYFGPAAEGIYTEGLMHNQDGLICGLRQLANETTQALQLFLRATTELRTFSILNRKAIDF 592 Query 593 LLRRWGGTCRILGPDCCIEPHDWTKNITDKINQIIHDFIDNPLPNQDNDDNWWTG 647 LL+RWGGTC ILGPDCCIEPHDWTKNITDKI+QIIHDF+D LP+Q ++DNWWTG Sbjct 593 LLQRWGGTCHILGPDCCIEPHDWTKNITDKIDQIIHDFVDKTLPDQGDNDNWWTG 647

14

Fig.7. Database BLAST analysis of amino acid sequences of the two EBOV GPs where

SEBOV GP is denoted as Query and ZEBOV GP as Sbjct (subject). The amino acid sequences showed extensive homologous regions, which correlates with the detected serological cross reactivity between the two antigens. By database O-glycan prediction (www.cbs.dtu.dk/services/NetOGlyc/), it is clear that the central, divergent part is a mucin region (denoted as green), which most likely harbor type-specific EBOV epitopes. (http://blast.ncbi.nlm.nih.gov/Blast.cgi)

Discussion

Due to several EBOV outbreaks that have occurred in Uganda and the new outbreaks that hit several countries of West Africa in 2014 we decided to carry out an investigation of seroprevalence to this virus. Aims for the present study were to develop serological assays to SEBOV, ZEBOV and EBV based on their major envelope gp proteins and to assay seroprevalence of three viruses in the patients’ sera from Uganda, with and without a history of Ebola infection, and that from Swedish blood donors. The study also aimed to analyze the SEBOV sequence to ZEBOV sequence using database of BLAST sequencing and O-glycan prediction. The investigation on the presence of IgG antibodies against EBOV disease was performed on serum samples collected from the affected District of Gulu in the Northern part of Uganda during the outbreak of the year 2000. Randomly selected serum samples from Swedish blood donors were used as method control for the study. The study used three types of antigens for the investigation, which were SEBOV GP (aa 1- 647), ZEBOV GP (aa 1-647) and EBV gp-350 GP (aa 1-860). The expectation was to observe an absence of IgG antibodies to SEBOV and ZEBOV but also to demonstrate a presence of IgG antibodies to SEBOV in the serum samples from Ugandan EBOV patients.

The developed serological assays on SEBOV, ZEBOV and EBV functioned well with good signal and reasonable background due to the Spectrophotometer measurements (405nm -620nm). The EBV antigen provided the best serological assay (i.e. high signal over low background) compared to the SEBOV and ZEBOV antigens. EBV assay needed the least antigen concentration and gave a high signal with less background compared to the other

15

two assays. The background of the ZEBOV antigen was the highest, indicating possible impurities in this protein.

The seroprevalence assays for the three viruses in patients’ serum from Uganda and from Swedish blood donors demonstrated the following results, which were graded according the median values in these seroassays: There was a high prevalence of IgG antibodies against the SEBOV in confirmed EBOV disease patients (fig. 3.) compared to the serum samples of confirmed Non-EBOV disease patients according to the analyzed results. Only one Swedish blood donor was seropositive to these two antigens seen (fig. 3. and 4.). Furthermore it was observed (fig. 4.) that more patients had IgG antibodies against ZEBOV in the Non-EBOV disease patients’ serum samples compared to the EBOV disease patients. Also here, the same Swedish blood donor was seropositive. Thus, although IgG antibodies to SEBOV and ZEBOV were commonly demonstrated in Ugandan patients’ serum, we cannot explain why one Swedish blood donors’ serum was positive to both antigens. The anonymity hindered questions to this person on subjects such as eventual history of travels to Central Africa, contacts with exotic animals such as bats and monkeys etc. Seroprevalence (fig. 3.) seemed to be slightly higher to SEBOV than to ZEBOV in Ugandan subjects, although overall the seroprevalence to both viruses was similar in patients with Ebola and those without this disease.

We also observed probable cross reactivity (fig. 6.) between SEOBV and ZEOBV seroassay, whereby some subjects demonstrated similar reactivity of OD values towards both antigens. The correlation between these two seroassays was significant (Pearson R= 0.2987, P (two-tailed) = 0.014). Database BLAST analysis (fig. 7.) of SEBOV G protein compared to ZEBOV G protein demonstrated extensive homologous regions in the N-terminal and C-N-terminal parts of the protein, which may explain the detected cross reactivity (fig. 6.) between the two antigens. On the other hand, several EBOV disease patients showed higher reactivity to SEBOV than to ZEBOV (fig. 6.), indicating a type-specific EBOV IgG response. The BLAST analysis showed a central, divergent region (area mark in green color), which was predicted to carry numerous O-linked glycans to form a mucin region. Most likely, this mucin region harbored EBOV type-specific epitopes.

16

Seroassays of both groups of patients (fig. 5.) demonstrated higher prevalence of IgG antibodies against EBV virus compared to the Swedish blood donors. This functioned as a control that the IgG antibodies in the Ugandan samples were not destroyed by storage and transport why the data on the EBOV antibodies could be trusted. It is of interest that almost all Ugandans showed seropositivity to EBV while 10% of the Swedish subjects were seronegative and therefore susceptible to this virus.

A higher seroprevalence of IgG antibodies against SEBOV (fig. 3.) in the EBOV disease patients than Non-EBOV disease patients is suspected since this virus caused the outbreak. But why were not all EBOV disease patients seropositive? Most likely, antibodies might not have developed in all patients, which might also be the cause of death in these patients due to lack of antibodies. Although we analyzed all the samples in a coded fashion and therefore did not have access to data on which patients that died, one study [15] demonstrated that patients who did not survive had little or no immunoreactivity. High demonstration of IgG antibodies in the Non-EBOV disease patients (fig. 3.) might represent subclinical infections of EBOV, i.e. asymptomatically infected subjects. The findings of the present study can be related to studies [16, 19] on SEBOV disease in human survivors of year 2000 outbreak, who had high immunoreactivity of antibodies directed to four viral proteins such as NP and GP 1-649 in the beginning and later to VP40. We had received more Ugandan female serum samples compared to Ugandan male serum samples but we were unable to obtain more information as to why there was such an imbalance in the gender. Taken together, the results of this study showed seroprevalence of IgG antibodies to two EBOV genogroups in Ugandan patients, and suggested that several cases of infection with these viruses failed to develop IgG antibodies. Finally, findings of specific IgG also in non-EBOV patients (fig. 3.) supported the existence of asymptomatic infections. Our results stimulate further efforts in form of large seroprevalence studies in different African (and European) populations to better understand the natural infection of EBOV.

4. Conclusion

All the three seroassays functioned well with good signals, but the ZEBOV assay had a somewhat high background. IgG antibodies to SEBOV and ZEBOV was demonstrated in

17

some patients and in one Swedish blood donor. Seroprevalence seemed to be higher to SEBOV than to ZEBOV in Ugandan patients, both those with a history of Ebola and those without, indicating a more common circulation of the former virus in Uganda. Extensive homologous regions regarding amino acid sequences of SEBOV and ZEBOV G proteins correlated with serological cross reactivity observed in these two seroassays. Subclinical infections of both viruses seem to exist.

5. Acknowledgments

Lots of thanks go to Dr. Tomas Bergström for the supervision of this study and to Maria Johansson for the technical supervision, and the Staff at the 3rd floor for your support, (Microbiolog, Virology Department, Gothenburg University), to Prof. Mikael Ivarsson for course supervision (Biomedicine Department, Örebro University), lots of thanks go also to all the professors we collaborated with: Pro. Dr. Julius J. Lutwama and Pro. Dr. Edward K. Mbidde (Uganda Virus Research Institute, Entebbe, Uganda), and Pro. Klas Ola Blixt, (Chemistry Institute, University of Copenhagen, Denmark)

7. References

1. Büchen-Osmond, Comelia. ICTVdB Virus Description-01.025.0.02. Ebola virus. International Committee on Taxonomy of Viruses. Retrieve 2006;2009-06-02. 2. Ascenzi P., Bocedi A., Heptonstal J. et al. “Eblavirus Marburgvirus: insight the Filviridae family”. Mol Aspects Med. 2008;29:151-85.

3. Feldmann H., Geisbert TW. Ebola haemorrhagic fever. Lancet. 2011;377:849-62.

4. Isaacson M., Sureau P., Courteille G., Pattyn SR. Clinical Aspects of Ebola Virus Disease at the Ngaliema Hospital, Kinshasa, Zaire, 1976. Retrieved 2009-07-08.

5. Feldmann H., Geisbert TW. Ebola haemorrhagic fever. The Lancet. 2011;377 (9768): 849-862.

6. ab Special Pathogens Branch CDC . Known cases and outbreaks of Ebola Hemorrhagic Fever. Center for Disease Control and Prevention. Retrieved 2008-08-02.

7. McCormick & Fisher-Hoch 1999, p. 300

8. Waterman, Tara. Ebola Côte D’Ivoire Outbreaks. Stanford University. Retrieved 1999;2009-05-30.

18

9. Wamala J., Lukwago L., Malimbo M. et al. Ebola Hemorrhagic Fever Associated with Novel Virus strain, Uganda. 2007-2008. Emerging Infectious Disease 16 (7). Retrieved 2010-06-24.

10. Barrette RW., Xu L., Rowland J M., et al. Current perspectives on the phylogeny of Filoviridae. Infect Genet Evol. 2011;201:1514-9.

11. Kuhn JH., Dodd LE., Wahl-Jensen V. et al. Evaluation of perceived threat differences posed by filovirus variants. Biosercur Bioterror. 2011;9:361-71.

12. Outbreak of Ebola hemorrhagic fever, Uganda, 2000 – 2001. Can Commun Dis Rep. 2001;27:49-53.

13. CDC. Outbreak of Ebola Hemorrhagic Fever ---Uganda, August 2000--January 2001. Centers for Disease Control and Prevention. 2001;50(05);73-7. Available from:

http://www.cdc.gov/mmwr/preview/mmwrhtml/mm5005a1.htm#top

14. Road Distance Between Kampala And Gulu With Map. Globefeed.com. Retrieved. 2014-05-25.

15. Sobarzo A., Groseth A., Dolnik O. et al. Profile and Persistence of the Virus-Specific Neutralizing Humoral Immune Response in Human Survivors of Sudan Ebolavirue. The Journal of infectious Diseases. 2013;208:299-309.

16. Sobarzo A., Perelman E., Groseth A. et al. Profiling the Native Specific Human Humoral Immune response to Sudan Ebola Virus Strain Gulu by Chemiluminescence Enzyme-Linked Immunosorbent Assay. American Society for Microbiology. Clin Vaccine Immunol. 2012;19(11):1844-52.

17. Sobarzo A., Lutwama JJ., Guttman O. et al. Persistent Immune Responses after Ebola Virus Infection. N ENGL J MED. 2013;369;5

18. Maeda E, Akahane M, Kiryu S et al. Spectrum of Epstein–Barr virus-related diseases: a pictorial review. Jpn J Radiol. 2009;27 (1): 4–19.

19. Leroy EM., Baize S., Volchkov VE., et al. Human asymptomatic Ebola infection and strong inflammatory response. Lancet. 2000;355:2210-5.

20. Johnson BK., Wambui C., Ocheng D. et al. Seasonal variation in antibodies against Ebola virus in Kenyan fever patients. 1986;1(8490):1160. 1:1160.

21. D’Arrigo I.,Clo E., Bergstrom T. et al. Diverse IgG serum response to novel

glycopeptide epitopes detected within immunodominant stretches of Epstein-Barr virus glycoprotein 350/220: diagnostic potential of O-glycopeptide microarrays. Glycoconjugate journal. 2013;30(7):633-40.

19

22. Liu C., Dalby B., Chen W. Transient Transfection Factor for High-Level Recombinant Protein Production in Suspension Cultured Mammalian Cells. Mol Biotechnol.