Development of a sensitive novel diagnostic kit for the highly pathogenic avian influenza A (H5N1) virus

Influenza viruses belong to the Orthomyxovirus family, whose genome is composed of eight segments of negative-sense RNA encoding 12 proteins. Two major glycoproteins, hemagglutinin (HA) and neuraminidase (NA), are located on the viral envelope, and 16 HA subtypes and 9 NA subtypes of avian influenza A have been identified on the basis of their antigenicities [1, 2]. There are three types of influenza virus, namely, A, B, and C. Influenza A (both H1N1 and H3N2 subtypes) and influenza B viruses circulate among the human population each year and are the causative agents of seasonal flu. There have been several pandemics of influenza A infections, which have resulted in the deaths of many humans and animals [2, 3]. The high variability of influenza A viruses is driven by frequent mutations in genomic RNA (drift) and by genetic reassortment among avian, porcine, and human strains [4]. This has hampered the development of a universal cross-protective flu vaccine.

The 2003 and 2004 outbreaks of the highly pathogenic avian influenza (HPAI) virus of subtype H5N1 that occurred in poultry and wild birds were genetically traced back to the H5N1 HPAI virus that caused the first outbreak in Hong Kong in 1997, and also those of 2001 and 2002 [5]. This report stated that although the H5N1 HPAI virus remained endemic to that region, it had the potential to become pandemic. The H5N1 HPAI virus has occasionally crossed the species barrier to humans in Asia, resulting in human fatalities [6, 7], the first of which was recorded in Vietnam in December 2003 [8]. Numerous clinical cases of H5N1 HPAI virus infections have since been reported in Vietnam, and the disease has spread to other countries in Southeast Asia and the Middle East such as Indonesia, Cambodia, Thailand, Egypt, and Turkey [7, 9].

The H5N1 HPAI virus has evolved into many phylogenetically distinct clades and subclades, and these diverse lineages have been largely geographically separated since 2005 [9]. During 2007 in northern Vietnam, the clade 1 virus was displaced by the clade 2.3.4 strain that has a different antiviral susceptibility profile [10, 11]. The diversity of this virus is continuing to expand [3, 12, 13]. Although the earlier endemic outbreak of the avian H5N1 HPAI virus appears to be under control, the threat of a human influenza pandemic remains.

PCR-based molecular tests are one of the most sensitive ways to detect the influenza virus, and conventional and real-time RT-PCR methods have been developed to diagnose H5N1 HPAI virus infections in humans [14–16]. However, only centralized and well-equipped laboratories with trained personnel can perform these analyses. Viral antigen detection using antibodies (Abs) offers an easier and quicker diagnostic test; however, commercially available rapid detection kits for influenza A and B have poor clinical sensitivity for the identification of H5N1 HPAI infection [9]. Using previously prepared monoclonal antibodies (mAbs) against influenza A virus HA of the H5 subtype [17], we developed a rapid, sensitive, and H5N1 HPAI virus-specific diagnostic test kit for H5 HA in combination with detection of the universal NP antigen of type A influenza (H5/A kit), which is processed by a compact and fully automated luminescence analyzer, POCube (Toyobo Co Ltd., Osaka, Japan). In this study, we evaluated this novel H5/A diagnostic kit using clinical specimens infected with the H5N1 HPAI virus (genetically confirmed) in Vietnam and demonstrated the sensitive dual detection of H5 HA and type A nucleoprotein (NP) antigens for the first time. Despite of a limited number of available H5N1 clinical specimens, our results strongly suggest that this diagnostic test is a useful tool in the rapid and reliable identification of H5N1 HPAI virus infections.

Different combinations of anti-H5 HA mAbs [17] and a pair of biotinylated OM-b and ALP-conjugated 1C10 mAbs were selected for H5 HA detection.

The H5/A kit contains two sets of mAb mixtures to detect H5 HA simultaneously with NP antigens which are common to all influenza A viruses. This combination is important for the definitive diagnosis of H5 HPAI virus infection. When a clinical specimen is loaded into a reservoir tank in the cartridge, which is placed in the machine along with two reaction vessels, an automatic diagnosis for H5N1 HPAI virus infection is provided (Figure 1). To determine the sensitivity of the H5/A kit, inactivated H5N1 HPAI virus of representative clades (Table 1) were used. The quantities of viruses were determined by TCID₅₀ and quantitative real-time RT-PCR targeting the universally conserved matrix gene of influenza A. These viruses were serially 10-fold diluted and the luminescence counts were measured. Then, the virus TCID₅₀ titers at each dilution were calculated and plotted as shown in Figure 2. The dotted line indicates the lowest detection threshold, which was determined by adding four standard deviations to the average of eight measurements of the negative control. Thus, the POCube detected these viruses at a titer equivalent to 10¹–10³ TCID₅₀.

	Virus strain	Clade	TCID50/50 μl	RNA copies
			(Log10)	(×1010/ml)
1	A/Vietnam/1194/2004 (NIBRG-14)	1.1	7.0	2.88
2	A/Indonesia/5/2005 (PR8-IBCDC)	2.1.3.2	7.4	3.81
3	A/turkey/Turkey/1/2005 (NIBRG-23)	2.2.1	5.9	1.60
4	A/Anhui/1/2005 (PR8-IBCDC RG-5)	2.3.4	8.3	3.73

We also confirmed that the kit detected type A antigens, but not H5 antigens, in all other subsets of type A influenza viruses examined (H1N1, H2N3, H3N8, H4N6, H6N2, H7N1 H8N4, H9N2, H10N7, H11N6, H12N5, H13N6, H14N5, and H15N8) including strains recently endemic in Japan (mostly H1N1pdm or type B), indicating that the kit is highly specific to H5 HA (data not shown). Furthermore, we tested clinical swab specimens taken from 15 seasonal influenza and 30 non-influenza patients and found no false positives of H5. Thus, the POCube system offers a simple, safe, and rapid diagnostic system, and is expected to be highly sensitive compared with existing commercial kits that use immunochromatography (IC) technology.

When samples are H5-positive but type A-negative, such as was the case for patient M (luminescence counts of H5 and type A were 4,071 and 2,622, respectively), the test needs to be repeated or confirmed by genetic analysis. This may occur when the level of virus antigen is at the threshold of detection, such as was the case for sample No. 2 of patient A. Again, the commercially available RIDT showed poor sensitivity, with only one of nine samples testing positive for type A.

H5 HA HPAI virus detection using POCube and the H5/A kit is a highly sensitive and simple rapid diagnostic system that can be completed within 15 minutes. During the preparation of this manuscript, Sakurai et al. reported a 10–100-fold more sensitive detection of H5 HA using fluorescent beads and an improved IC method [21]; however, the clinical effectiveness of this method remains to be evaluated. We recently determined the epitopes of our anti-H5HA mAbs and found that OM-b, one of mAbs used for the H5/A kit, was broadly reactive to various clades of H5N1 influenza virus isolates in Asia, indicating that ther kit is quite useful for the diagnosis of H5N1 infection in Asian counties [22].

The POCube system using the H5/A kit is useful for the rapid detection of H5N1 HPAI virus infections in humans in hospitals and other clinical settings where technical resources are limited.

YTY worked in NIID for this study and recently moved to the Department of Medical Technology, School of Human Sciences, Tokyo University of Technology, 5-23-22 Nishi-Kamata, Ohta-ku, Tokyo 144–8535, Japan.