Influenza virus is a lipid-enveloped orthomyxovirus, possessing a segmented, single stranded RNA genome with negative orientation. Based on their genetic and antigenic differences they are divided into three types: A, B and C, among which influenza A virus causes the most significant morbidity and mortality in the human population and domestic animals on a global scale. There have been several global pandemics for the last 100 years. The 1918 'Spanish influenza' killed as many as 50 million people worldwide ; the 1957'Asian influenza' H2N2, 1968'Hong Kong influenza' H3N2 [15, 21] and 2009 swine-origin H1N1  all gave rise to the world-wide panic. Besides, influenza A virus causes yearly epidemics of respiratory illness of varying severity worldwide in people of all ages, and results in thousands of deaths and enormous economic losses .
Vaccination is the primary method to prevent or lower the burden of influenza disease. As the polymerase complex of influenza virus does not possess a proof reading activity, it is easy to bring in mutations to the progeny virus (called "antigenic drift"), especially to the HA and NA parts. Those mutations lead to the conformational alterations of epitopes recognized by neutralizing antibodies, thus making existing vaccines less-or non-effective. On the other hand, due to the nature of the segmented genome, influenza virus can independently recombine segments upon the infection to a cell (called "antigenic shift"). For influenza A virus, 16 HA-subtypes (H1-H16) and 9 NA-subtypes (N1-N9) have been discovered, and recombinations among the subtypes may lead to new antigenic properties [32, 51]. Consequently, the vaccines against influenza virus need to be adjusted annually. Each year, the World Health Organization (WHO) identifies new strains of influenza viruses through the international surveillance system, summaries the epidemiological data twice a year in two meetings (one in February, the other in September), and decides which strains will be included in the trivalent influenza vaccine production .
Most of present available influenza vaccines are generated in embryonated hen's eggs. Seed strain is inoculated in the allantoic cavity of embryo. Three days after inoculation, allantoic fluid is collected for the purification of progeny virions, which is then inactivated by formalin or β-propriolactone and detergent treatment. Then we can either use the harvested virus as wholevirus vaccine, or further get the purified hemagglutinin and neuraminidase proteins as subvirion or subunit vaccines [10, 18]. Although this traditional vaccine production system has served well for decades, there are several insurmountable defects for the egg-based vaccine: 1) In order to produce enough influenza vaccines, a large amount of SPF level embryonated chicken eggs are needed ; 2) The manufacture of influenza vaccines is limited to influenza virus strains that replicate well in eggs, but some vaccine strains such as high pathogenic avian influenza H5N1 which can kill the embryos could not be produced in this way ; 3) The vaccine strain that grows in embryonated chicken eggs sometimes undergoes antigenic variations, resulting in a less effective vaccine; 4) last but not least, the vaccine has potential safety concerns in individuals with egg allergies [5, 26]. Considering all this defects of traditional egg-based vaccine, efforts have been made to develop alternative vaccine production methods, among which the present influenza subunit and VLP vaccines are produced in a much more superior way.
Nowadays influenza subunit and VLP vaccines have been produced in bacterial , mammalian [4, 39, 54], and recombinant baculovirus (rBV) expression systems [11, 52]. Owing to the development of protein purification technology, as well as the commercially available purification systems [24, 42], it is now much easier to purify proteins expressed in prokaryote or eukaryote systems. So under these favorable backgrounds, influenza subunit and VLP vaccine technology is developed from strength to strength. This paper is focused on the recent development of subunit (for viral vector vaccine such as intranasal influenza vaccine designed by Vaxin, the active ingredient of which was the protein subunit, is also classified as subunit vaccine) and VLP vaccines against influenza A virus. According to the difference in the expression of influenza antigen, subunit vaccines can be further divided into HA subunit vaccines, NA subunit vaccines, NP subunit vaccines and M2 subunit vaccines . As to VLP vaccine, which is produced by co-expressing of influenza M1 in combination with HA, NA, and/or M2 protein, is also discussed in the following text.
The current licensed influenza vaccine provides an economical and effective means to reduce the impact of an influenza infection. However, as discussed above, conventional egg-based influenza vaccines have a number of limitations. All the limitations call for a new vaccine production method, through which the influenza vaccine can be produced in a much shorter time period and in a much easier scalable way. The recombinant influenza vaccine, taking the advantage of recombinant DNA technologies, can be produced in such a novel way.
The influenza subunit and VLP vaccines (Table 1) offer the following potential advantages compared with the conventional egg based vaccines. Firstly, selection or adaptation of influenza virus strains for production at high levels in eggs is not required; meanwhile the cloning, expression and manufacture of recombinant influenza virus can be accomplished within a short period of time (probably less than 8 weeks) much less than the traditional egg based vaccine production (about 6 months) ; Secondly, the influenza subunit and VLP vaccines are produced in a scalable, reproducible, and quality-stable, as well as low bio-burden way, which results in lower cost and quicker response in the epidemic season with a new epidemic or pandemic strain of influenza virus , thus providing sufficient vaccine for inoculators ; Thirdly, puriflcation procedures for influenza subunit and VLP vaccines is relatively easier which does not include influenza virus inactivation or organic extraction procedures, thus avoiding additional safety concerns because of residual harmful chemicals in the vaccine. Perhaps what is the most important, from a clinical perspective, vaccine produced in this way is highly purifled and does not contain ovalbumin or other antigenic proteins present in eggs [11, 23]. Besides, if the universal influenza vaccine would be produced, it will leave out the trouble of adjusting the vaccine strain each year, and will have the possible reservation of large amount of vaccine against the emerging pandemic human flu. Though the universal vaccine is still in the early stage of research, this concept and further manufacturing of product deserve the expectation. All those advantages fully demonstrated the great potential for the production of recombinant influenza vaccines to prevent influenza infection in epidemic even pandemic seasons.
Table 1. Summary of the potential subunit and VLP vaccines against influenza
Above all, our comprehensive review on the recent developments of subunit and VLP vaccines against human influenza virus would help to get insight into the current situation of influenza vaccines, and suggest the future design and development of novel influenza vaccines.