1 1 Comparison of 2 Commercially Available Dengue NS1 Capture ...

CVI Accepts, published online ahead of print on 6 August 2008 Clin. Vaccine Immunol. doi:10.1128/CVI.00140-08 Copyright © 2008, American Society for Microbiology and/or the Listed Authors/Institutions. All Rights Reserved.

1 2

Comparison of 2 Commercially Available Dengue NS1 Capture Enzyme-Linked

3

Immunosorbant Assays for Diagnosis of Acute DENV Infection with a Single

4

Clinical Sample

D E

5 6

Kovi Bessoff¹, Mark Delorey², Wellington Sun¹, and Elizabeth Hunsperger¹*

7

T P

8

¹ Centers for Disease Control and Prevention (CDC), Division of Vector-Borne Infectious

9

Diseases (DVBID), Dengue Branch, San Juan, PR

10

² Centers for Disease Control and Prevention (CDC), Division of Vector-Borne Infectious

11

Diseases (DVBID), Fort Collins, CO

12 13

E C

C A

14

*Corresponding Author:

15

Dr. Elizabeth Hunsperger

16

United States Public Health Service

17

CDC- Dengue Branch

18

1324 Calle Cañada

19

San Juan, PR 00920

20

Phone: (787)-706-2474

21

Fax: (787)-706-2496

1

22 23

Abstract: Dengue virus (DENV) non-structural protein 1 (NS1) has shown promise as a

24

novel diagnostic marker of acute DENV infection. Current techniques used to diagnose

25

acute DENV infection including virus isolation and RT-PCR are costly and difficult to

26

perform, while traditional serological assays have low sensitivity during the acute stage

27

of infection. Two commercially available NS1 antigen capture ELISAs, the Platelia

28

Dengue NS1Ag test (Bio-Rad Laboratories, Marnes La Coquette, France) and the Pan-E

29

Dengue Early ELISA test (Panbio Diagnostics, Brisbane, Australia), were evaluated

30

against a well-characterized panel of 208 real time RT-PCR- and virus isolation-positive

31

sera, as well as 45 real time RT-PCR and serologically negative sera from patients with

32

other acute febrile illness. The overall sensitivities were 64.9% (CI95 58.2-71.1) for the

33

Panbio test and 83.2% (CI95 77.5-87.7) for the Bio-Rad test, with inter-serotype variation,

34

especially in DENV-4. Predictive models were constructed to identify factors that had a

35

significant influence on a test’s outcome with respect to this panel of samples, in order to

36

identify the conditions in which the test will be most effective as a diagnostic tool. IgG

37

titer was found to be the only covariate that significantly influenced results in the Bio-

38

Rad test, while serotype and day post-onset of infection were found to significantly

39

influence results in the Panbio test. We conclude that the NS1 capture ELISA is a useful

40

tool that can improve testing algorithms to diagnose DENV infection in single samples

41

from acute and early convalescent cases.

D E

T P

E C

C A

42 43

Introduction:

44

2

45

Dengue viruses (DENV) are members of the genus Flavivirus, family

46

Flaviviridae, and exist as four antigenically distinct viruses. While most infections result

47

in asymptomatic response or mild febrile illness, (dengue fever, DF), all four serotypes

48

are capable of producing the more severe, and sometimes fatal, dengue hemorrhagic

49

fever/dengue shock syndrome (DHF/DSS)(10, 12, 28). Over the last 50 years, factors

50

like rapid urbanization and failure to control vector mosquitoes have led to the emergence

51

of endemic dengue in over 100 countries. With 2.5 billion people at risk for infection,

52

dengue has become the most important arthropod-borne virus affecting humans (11, 13,

53

34).

54

D E

T P

E C

Current diagnostic methods are often unable to recognize emerging epidemics in a

55

timely manner or at a reasonable cost, drastically reducing the efficacy of control

56

measures (13, 34). Virus isolation is a lengthy process which requires specialized

57

laboratory equipment (29, 32), and rapid antibody response to DENV in secondary

58

infections may prevent successful isolation from serum drawn as few as three days after

59

onset of fever (17, 28, 29, 32). The development and modification of nested RT-PCR

60

(16, 21) as well as real time RT-PCR (5, 8, 30, 31) techniques have significantly reduced

61

processing time; however, these procedures remain expensive, technically difficult, and

62

laboratory contamination can yield false positive results. Serological diagnosis offers

63

many advantages including more flexible schedules for testing, lower cost, and more

64

widely available reagents (32). However, cross-reactivity between other flaviviruses and

65

“original antigenic sin” complicate specific diagnosis of secondary flavivirus infections

66

(15, 17, 28, 32). Additionally, antibody half-lives of up to two months confound

67

diagnosis in cases where the date of onset of illness is unknown (17, 28, 29).

C A

3

68

The inability of serological-based methods to reliably diagnose acute infections

69

has been recognized as a serious impediment to quick detection of epidemics and

70

effective clinical case management (34). A number of options are being explored to

71

overcome this obstacle; one of the most promising is the detection of DENV non-

72

structural protein 1 (NS1) in infected serum. NS1 is essential for viral replication (20, 24,

73

25), and the amount of secreted NS1 (sNS1) in the serum of individuals infected with

74

DENV-2 has been shown to directly correlate with viremia (23). The protein is

75

detectable by enzyme-linked immunosorbent assay (ELISA) as early as the first day of

76

fever, and detection does not appear to be hindered by the presence of anti-dengue IgM

77

antibodies (2, 35). Given these attributes and protein kinetics, NS1-based ELISAs may

78

be an important diagnostic tool for those samples in which IgM is not detectable and PCR

79

is not available. Moreover, DENV-2 sNS1 levels have been shown to be higher in

80

patients with DHF (23), making diagnosis by NS1 a valuable predictive tool for severe

81

disease.

82

D E

T P

E C

C A

Two commercial tests, the Platelia Dengue NS1 kit (Bio-Rad Laboratories,

83

Marnes La Coquette, France) and the Pan-E Dengue Early ELISA kit (Panbio

84

Diagnostics, Brisbane, Australia), are currently available for the detection of DENV NS1

85

in human serum. Independent evaluations of the Platelia (9, 18, 19) and Pan-E (3)

86

systems have been published, however, the performance of the two tests has never been

87

evaluated against a common serum panel. We evaluated the sensitivity of both NS1

88

detection systems with a well-characterized panel of 208 samples determined DENV-

89

positive by virus isolation and real time RT-PCR. Additionally, specificity was

90

evaluated using a panel of 45 serum samples from patients with acute febrile illnesses

4

91

that tested negative for DENV by real time RT-PCR and serological methods

92

Sensitivities varied by serotype for both tests, with the highest rates of detection

93

occurring in DENV-1 positive samples and the lowest low occurring in DENV-4 positive

94

samples. The specificity was excellent for both kits.

D E

95 96 97 98

T P

Materials and Methods:

99 100 101

E C

Clinical Samples:

Human serum samples were obtained from the reference collection at the Centers

102

for Disease Control and Prevention (CDC)- Dengue Branch in San Juan, Puerto Rico.

103

All accompanying epidemiological data was obtained from the database maintained by

104

the Dengue Branch. The panel consists of real time RT-PCR- and virus isolation-positive

105

DENV1-4 samples gathered as the result of routine surveillance activities carried out at

106

the facility between 1998 and 2005. All samples were collected in Puerto Rico, and were

107

taken from patients during the first five days post-onset (DPO) of symptoms,

108

corresponding to the acute period of infection as defined by the Dengue Branch testing

109

algorithm (CDC, unpublished data). The panel consisted of 56 DENV-1, 45 DENV-2, 52

110

DENV-3, and 55 DENV-4 samples. Additionally, the panel included 45 samples

111

collected from patients with acute febrile illness that tested negative for DENV by PCR

112

and serology. Human surveillance data from 1998-2005 suggests that DENV was the

113

only flavivirus circulating on the island of Puerto Rico (CDC, unpublished data).

C A

5

114 115

Sample Characterization:

116

Serum samples were tested for the presence of dengue virus by the diagnostic

117

multiplex real time RT-PCR assay (6), and viral isolation in C6/36 cells as previously

118

described (14). Anti-dengue IgG titers were determined for all virus-positive samples

119

using the IgG ELISA described elsewhere (4). Because all samples were from acute

120

infections (≤5 DPO), any IgG titer of ≥1:40 was considered to be evidence of a secondary

121

flavivirus infection in accordance with WHO guidelines adapted for the IgG ELISA (26,

122

33), and DENV was the only flavivirus known to be circulating in Puerto Rico at the time

123

of collection (1998-2005).

124

D E

T P

E C

Additionally, a subset of 140 samples were tested by an IgM antibody capture

125

ELISA (MAC ELISA) (4). Table 1A shows the composition of the evaluation panel

126

according to primary or secondary infection, and table 1B describes the subset of samples

127

tested by MAC ELISA. IgG and IgM titers were used to evaluate the influence of

128

possible antibody-NS1 complexes on antigen detection.

129 130

C A

PCR negative samples were further tested by MAC ELISA, and 45 samples that

tested negative for DENV in both tests were selected to evaluate specificity.

131 132 133 134

Kit Evaluation: The Platelia Dengue NS1 and Pan-E Dengue Early ELISA kits are antigen capture

135

ELISAs that provide qualitative, non-serotype specific detection of DENV NS1 antigen.

136

All testing was performed in duplicate in strict adherence to the manufacturers’

137

instructions provided in the kit insert. All reagents and control sera were provided by the 6

138

manufacturer. Prior to testing, samples were thawed and all samples and kit reagents

139

were allowed to warm to ambient laboratory temperature. All microplate wash steps

140

were carried out with a 96-well manifold automatic plate washer, and plates were read at

141

a wavelength of 450 nm with a 630 nm reference filter. All samples and controls were

142

tested in duplicate unless otherwise noted. The testing protocols for each kit are briefly

143

described below.

D E

144 145 146

T P

Bio-Rad Kit:

Fifty microliters of patient serum, positive control, negative control, or cut-off

E C

147

control were diluted 1:2 in sample diluent and combined with 100 µL of diluted

148

horseradish peroxidase (HRP)-labeled anti-NS1 monoclonal antibody (MAb). The diluted

149

serum and conjugate were added to capture anti-NS1 MAb-coated microwells, and plates

150

were incubated for 90 minutes at 37°C. Capture MAb/NS1/HRP-labeled-MAb

151

complexes were formed when NS1 was present in the serum. Complexes were detected

152

via a colorimetric reaction in positive wells which was induced by adding 160 µL of

153

3,3’,5,5’ tetramethylbenzidine (TMB) substrate and incubating for 30 minutes at room

154

temperature in the dark. The reaction was stopped with the addition of 100 µL of stop

155

solution (1N H2SO4) and the plate was read. A sample ratio was determined for each

156

sample by dividing the average optical density (OD) of the test sample by the average

157

OD of the cut-off control (tested in quadruplicate). Sample ratios of