Factors influencing the immunogenicity of influenza vaccines

ABSTRACT

Annual vaccination is the best prevention of influenza. However, the immunogenicity of influenza vaccines varies among different populations. It is important to fully identify the factors that may affect the immunogenicity of the vaccines to provide best protection for vaccine recipients. This paper reviews the factors that may influence the immunogenicity of influenza vaccines from the aspects of vaccine factors, adjuvants, individual factors, repeated vaccination, and genetic factors. The confirmed or hypothesized molecular mechanisms of these factors have also been briefly summarized.

KEYWORDS:

• Influenza
• vaccine
• immunogenicity
• immune response

Introduction

Influenza is an acute infectious respiratory disease, which is mainly transmitted by droplets and human-to-human close contacts. It is estimated that there are 1 billion cases worldwide annually, including 3–5 million serious cases and 290,000–650,000 deaths.¹

The most effective measure to prevent influenza infection is by taking annual vaccination. However, for some special populations, including pregnant women, infants, elderly people and chronic patients, influenza still threats their life and health due to the poor effectiveness of current influenza vaccines. Pregnancy confers an increased risk for hospitalization with influenza. It was estimated that the risk of influenza-associated hospitalization in pregnant women was significantly higher than nonpregnant ones, with an odds ratio of 2.44.^2,3 WHO recommended that pregnant women should be the prior population for influenza vaccination. For infants, a variety of complications such as pneumonia and myocarditis can be triggered after influenza infection, some studies also reported that the infection can even cause encephalopathy and encephalitis among infants.^4–6 When elderly people are infected with influenza, their conditions are usually developed rapidly and may even cause death.⁷ For people with other basic diseases, the infection with influenza viruses will not only lead to severe conditions, sometimes will also aggravate their original diseases. Thus, in order to provide a better protection for these people, it is of great significance to identify the correlative factors that may influence the influenza vaccine effectiveness (VE).

There are mainly two factors that may influence the VE. One is the matching degree of circulating strains and vaccine strains, and the other is the immunogenicity of the vaccine. Many studies have confirmed that the estimates of VE vary widely from season to season. And the VE will decrease when the vaccine strains and circulating strains are poorly matched.^8,9 A meta-analysis study showed that in poorly matched seasons, the risk ratio (RR) of being hospitalized for influenza was 2.04 (95%CI = 1.29–3.22), whereas in well-matched seasons, the RR was only 0.64 (95%CI = 0.33–1.22), suggesting the importance of effective prediction for annual influenza vaccine strains.¹⁰ On the other hand, the immunogenicity of the vaccine, which means, the vaccine’s capacity of inducing immune response after inoculation, also has an important impact on the VE. The immunogenicity of influenza vaccine mainly can be evaluated by three indicators: the geometric mean titer (GMT) of antibodies, the seroprotection rate, and the seroconversion rate. Many factors may affect the immunogenicity of influenza vaccine, including vaccine factors, adjuvants, individual factors, repeated vaccination, and genetic factors.^11–14

Effect of vaccine factors

Dosage of vaccines

The effect of vaccine dose on immunogenicity usually varies from population to population. A study showed that antibody responses to half-dose trivalent inactivated vaccine (TIV) in adults were not inferior to the full-dose vaccine, particularly for those aged between 18 and 49. This suggests that under the circumstance of vaccine shortage, half-dose vaccination can be an option for 18–49 adults. However, for people aged 50–64, the antibody responses induced by half-dose vaccine is significantly lower than that by full-dose vaccine, indicating that half-dose vaccine cannot provide enough protection for elderly people.¹⁵ A prospective cohort study conducted by Coleman et al.¹⁶ also confirmed that, for healthy adults aged 20–45, the immunogenicity was adequate in both half-dose and full-dose AS03-adjuvanted monovalent 2009 pandemic H1N1 vaccine, and the difference has no significance in statistics. Many studies have shown that, compared with standard-dose influenza vaccine, high-dose TIV can effectively improve the antibody responses and reduce laboratory-confirmed influenza-like illness for elderly people.^17–19 For children, while exploring the optimal dose of vaccination, the safety of vaccine also should be seriously considered. A study showed that infants aged 6–35 months who received full-dose TIV had higher levels of geometric mean titers against A/H1N1 than those who received half-dose TIV. No differences were found between two dose groups against A/H3N2 and B, and the safety of full-dose TIV was also demonstrated in this study.²⁰ Skowronski et al.²¹ found that for infants aged 6–11 months, full-dose TIV may induce higher HI antibody titers to all 3 vaccine components compared with half-dose, without increasing reactogenicity. In general, most current studies suggested that high-dose vaccination is correlated with better immunogenicity, many scientists also recommend high-dose vaccines for people with heavy disease burdens, such as the elderly and immunocompromised individuals.^22,23 A high-dose influenza vaccine called Fluzone High-Dose Quadrivalent which contains four times the antigen was licensed only for persons aged 65 years and older. This kind of higher-dose vaccine is well tolerated and highly immunogenic, and may provide a better protection for elderly people.

Delivery modes of vaccination

Influenza vaccines can be delivered in a variety of ways, primarily by traditional intramuscular (IM) injection. In 2011–2012, an intradermal (ID) influenza vaccine was first used on people aged 18–64 in the US. A randomized-controlled trial (RCT) showed that the GMT and seroprotection rate of individuals injected ID vaccine were higher than those injected with traditional IM vaccine.²⁴ Another two RCT researches confirmed that the immunogenicity of ID influenza vaccine with 9 μg hemagglutinin (HA) was comparable to IM vaccine with 15 μg HA, and the safety was acceptable.^25,26 Compared with traditional IM vaccine, ID vaccine has many advantages including: lower antigen dosage and vaccine cost; application of microinjection system, smaller needle, less pain and simpler control.²⁷ Based on the advantages mentioned above, many experts believe that the use of ID method will help to improve the coverage rate of influenza vaccine. Besides IM and ID injection, influenza vaccine also can be delivered by inhalation. This mode of inoculation could simulate the natural infection process and induce normal immune response, at the same time promote the secretion of immunoglobulin antibodies by the upper respiratory mucosal epithelium. In this way, it will provide a better and wider protection for the inoculators.²⁸ To increase the coverage of influenza vaccination and improve the VE, we still have to explore and develop more delivery methods.

Vaccine types

Currently, there are mainly three types of vaccines in the global market: inactivated influenza vaccine (IIV), live-attenuated influenza vaccine (LAIV), and recombinant influenza vaccine (RIV). Among them, IIVs are the most efficient, valuable, and low‐cost tools to effectively reduce influenza virus infections.

LAIV is a kind of nasal spray vaccine in which the virulence has been weakened but retains immunogenicity. After the inhalation with LAIV, pathogens can grow in the human body and activate the corresponding immune responses without causing any disease. This natural infection pathway can not only induce humoral immunity and cellular immunity, but also induce mucosal immunity to provide a long-term and more extensive immune protection to the body. Compared with IIV, LAIV can also extend the duration of the antibody-secreting cell (ASC) response through the long-term stimulation caused by virus replication.²⁹ However, LAIV is not recommended for children under two years of age, adults aged ≥50, pregnant women, and people with immunodeficiency due to its residual virulence. Nowadays, in terms of the comparison of VE, most studies showed that the VE of LAIV was comparable to TIV and sometimes TIV presented a higher VE in adults, while in children and adolescents, LAIV has a higher VE than TIV.^30,31 While in terms of the immunogenicity, most studies suggest that IIV induces a stronger antibody response. An observed study conducted in 2013–2014 influenza season showed that among children aged 5–17, the HI antibody responses of LAIV recipients were lower than that of IIV recipients.³² American researchers found that healthy children aged 3–17 who received IIV had higher postvaccination GMTs and seroconversion rates to all four vaccine components than LAIV recipients.³³ Another study of children with cancer confirmed that, compared with LAIV, TIV induced higher GMT against influenza A viruses (P < .001), greater seroprotection against influenza A/H1N1 (P = .01), and greater seroconversion against A/H3N2 (P = .004).³⁴ Similarly, Canadian scientists also concluded that the magnitude of seroconversion induced by IIV against influenza A virus was much greater, suggesting that LAIV was less immunogenic than IIV.³⁵ It is worth mentioning that the immunogenicity of LAIV may be underestimated by using HI titer as an evaluation indicator since the local mucosal immune responses and T cell responses elicited by LAIV cannot be measured by HI titer.^29,36 For adults, another possible reason why the immunogenicity of LAIV is underestimated is that the virus can be partly neutralized or eliminated by their previously acquired local secretory IgA (sIgA), serum IgA, serum IgG, and cytotoxic T lymphocytes. sIgA is the main humoral mediator of mucosal immunity, its antibodies can neutralize potential pathogens at the nasal entrance site before they attack the epithelial cells. While serum IgA and IgG antibodies transude from the serum to the mucus and provide protection against infection after the virus invades the mucosal epithelium.^37,38 After the infection, the protection mediated by IgA local immune response in the upper respiratory system may last for 3–6 months or longer.^29,37,39 As a result, LAIV inoculated during this period will be neutralized to a certain extent before eliciting adaptive immune responses. For most adults who have been infected with influenza, LAIV may also be partly neutralized by tissue-resident memory T cells in lungs and immunological cells which are stored in the tonsil.^40,41 Therefore, the effect of LAIV in adults cannot be objectively evaluated by serological indicators alone. Nowadays, since studies on VE and immunogenicity of LAIV are contradictory, the application of LAIV is still restrained and closely monitored in several countries. However, for young children, the priming of T cell responses and mucosal immune responses are particularly important since young children have not been through many infections so they have not acquired many specific memory B cells yet. Mucosal antibodies are essential in protecting the upper respiratory tract from infection. While T cell responses have a variety of functions, including preventing virus replication, participating in antigen presentation, inducing the differentiation of B cells and mediating the humoral immune response. The ability of LAIV to induce a robust T cell responses and mucosal IgA antibodies have been demonstrated.^42,43 Hoft et.al⁴⁴ reported that LAIV and IIV induce similar humoral responses, but only LAIV induces diverse T-cell responses in young children. Therefore, the protective potential of LAIV for children in influenza season is of importance and require further investigation.

RIV is a synthetic vaccine that reassembles the genes of influenza viruses onto some other viruses, plasmids, bacteria,, and cells. Eggs are not required in the whole producing process of RIV. In this way, the manufacturing process might be faster than that of egg-based vaccines. In 2013, a trivalent RIV called FluBlok made from Protein Sciences was first licensed by the FDA. Afterward in 2017, a quadrivalent version was licensed to use on adults 18 years and older in the United States. RIV contains purified recombinant HA protein antigens which use the baculovirus as an expression vector. Its immunogenicity and safety have also been demonstrated in recent years. The HA dose in FluBlok was three times higher than that in IIV without a significant increase in side-effects.⁴⁵ An earlier phase 1 clinical trial found that the HI titer of trivalent RIV (RIV3) was lower than that of TIV in children aged 6 to 59 months. The serologic responses to FluBlok were higher in the older group (children aged 36–59 months) than the younger group, but were still somewhat lower compared to TIV.⁴⁶ A randomized trial conducted in subjects aged 6 to 17 years showed that the GMT and seroconversion rate of RIV to H1N1 and H3N2 were higher than that of IIV. No vaccine-related adverse events were found in both two vaccines.⁴⁷ Baxter et al.⁴⁸ observed higher antibody responses to influenza A virus, but similar responses to influenza B virus in FluBlok group in comparison to IIV group in healthy adults. The superior immunogenicity of RIV (especially to A/H3N2) has also been reported in the elderly≥65 years.⁴⁹ In a word, RIV was safe, immunogenic, and even effective under the condition of antigenic mismatch between the vaccine antigens and circulating viruses, suggesting the potential of RIV to provide a better cross-protection.⁵⁰ With the continuous innovation of technology on vaccines, more and more emerging RIVs such as NanoFlu from Novavax and Supemtek from Sanofi will be available soon.

Effects of adjuvants

The addition of adjuvant is a practical means of increasing the immunogenicity and effectiveness of vaccines. Aluminum adjuvant is one of the most classical influenza vaccine adjuvants. As a kind of adsorbent, aluminum adjuvant may strongly absorb antigens into precipitation and slowly release it to prolong the reaction time and fully activate the humoral immune response, while its ability to induce cellular immune response is poor.⁵¹ In order to optimize the immunogenicity of influenza vaccines, various new adjuvants such as oil-in-water emulsions including AS03 and MF59 have emerged. Their role is to induce immune cells to secrete cytokines and chemokines, promote transition from monocytes to dendritic cells and recruit monocytes and neutrophils from the blood to migrate to the vaccination site. Thus, the immune response can be enhanced.^52–54 An RCT study designed to evaluate the immunogenicity of an inactivated 2009 H1N1 influenza vaccine showed that in both healthy adults and elderly people, the GMT of subjects in the AS03-adjuvanted group was statistically higher than that in the control group. 61% of adjuvant free subjects and 81% of AS03 adjuvanted subjects reached seroprotective levels, respectively.⁵⁵ A study of children aged 30 to 96 months showed that compared with recipients who only received TIV, MF59-adjuvanted TIV vaccination produced 6.9 to 8.0-fold higher antibody responses against A/H3N2 and B strains.⁵⁶ Another study aimed to evaluate the immune responses to A/H1N1 vaccination also reported that the application of MF59 as an adjuvant may significantly upregulate the speed and level of antibody response.⁵⁷ The superior immunogenicity of MF59-adjuvanted influenza vaccines was also proved by a systematic review and meta-analysis contained 31 studies.⁵⁸ Many researches have confirmed that influenza vaccines adjuvanted with AS03 or MF59 are generally well tolerated and displayed an acceptable safety profile similar to the nonadjuvanted ones.^53,59–61 A MF59 added influenza vaccine named FLUAD was already approved in the United States for persons aged 65 years and older in 2015. Now, its quadrivalent version was also approved in 2020. Studies of some other emerging adjuvants were also performed. As a potential adjuvant for influenza vaccines, CpG can activate immune cells such as T cells, B cells, and NK cells to produce larger amounts of cytokines to enhance the body’s specific and nonspecific immune response. Researchers found that the use of CpG7909 adjuvant may maintain its immunogenicity with lower dose to save the vaccine cost. The CpG7909 adjuvanted vaccine was also found to have similar frequency and intensity of most adverse reactions to nonadjuvanted ones, and both of them were well tolerated.⁶² The mechanisms of cytokine adjuvants such as interferon depend on their original regulatory functions in the immune system. In principal, adjuvanted influenza vaccines are safe and well-tolerated, though they may increase some local site symptoms like redness and pain at the injection site. Except for FLUAD, other licensed influenza vaccines with adjuvants like Prepandrix, Orniflu, Panflu and so on are still in use under different conditions.⁶³ To cope with different influenza seasons, studies on practical-adjuvanted influenza vaccines still need to be further explored.

Effect of individual factors

Gender

Gender is one of the important factors which may influence the immune responses of influenza vaccine. In Engler’s study,¹⁵ women of all ages had been identified with significant higher GMT than man after vaccination, whatever the dose or strain is. Other studies have found that vaccines can induce higher levels of CD4 ⁺ lymphocytes and Th1 cytokines in women than men.^64,65 In the light of this phenomenon, some researchers believe that the reason why women’s responsiveness is higher than that of men can be attributed to the difference in the sex steroid hormone levels. This kind of hormone can directly binding to intracellular receptors in immune cells, such as monocytes, B cells and T cells.^66,67 Furma et al.⁶⁸ found that men with elevated serum testosterone level and high expression of lipid metabolism-related genes often conduct the lowest antibody responses to TIV, suggesting that testosterone could inhibit the immune response after influenza vaccination. For women, half of the activated genes in T cells have estrogen response elements in their promoters, indicating that estrogens can directly mediate the expression of these genes to influence the activity of T cells.⁶⁹ In addition, estradiol can also stimulate the production of antibodies by B cells at physiological concentrations.⁷⁰ American researchers found that after the vaccination, compared with adult males, adult females developed higher vaccine-induced immune responses to the influenza vaccine. And this immune response is positively associated with the concentrations of estradiol. The mice model also showed that the antibody responses were increased in females by estradiol and decreased in males by testosterone.⁷¹ In female mice, the ovariectomy may negatively modulate the antibody response to influenza vaccine, but the immune response can be restored after the administration of estrogen.⁷²

Age

Individuals of different ages also differ in their ability to respond to influenza vaccines. Young children, especially newborns, usually have a higher burden of influenza infection and a relatively weak responsiveness to vaccines. It has been reported that children aged 2 to 5 years have a higher immune response to TIV than children under 2 years of age.⁷³ Sasaki et al.⁷⁴ also found that IgG circulating ASC and serum antibody responses were relatively low in the younger children compared to older children and adults, suggesting that the immune responses induced by vaccines were different in various age groups. The immune system of infants is not fully developed, their magnitude and activity of antigen-presenting cells (APCs), immune cells and cytokines are significantly lower than adults and older children.⁷⁵ During infancy, the peripheral B cell pool is characterized predominantly by transitional and naive B cells.⁷⁶ Due to the lack of memory B cells generated from antigenic stimulation, infants are less responsive to vaccines compared with adults and older children. For newborns, the neonatal immunological milieu is polarized toward Th2-type immunity with dampening of Th1-type immunity. Compared with older children, infants’ functional deficiencies of APCs in innate immune system and impaired humoral immunity may result in poorer antibody responses, both in quantity and quality.^77,78 Besides, researchers concluded that the low abundance of intestinal microbiota may also account for the relatively weak immunity of infants.⁷⁹ Old people are also part of the vulnerable populations. Current evidence suggests that influenza vaccines are less effective in older people than in young adults, and the protection period induced by influenza vaccines in the elderly is also relatively short.^11,80,81 A retrospective analysis of 31 studies by Goodwin et al.⁸¹ showed that the estimate of clinical vaccine efficacy in the elderly (aged 58–104) was significantly lower than that of in young adults (17–53% vs. 70–90%), as well as the seroconversion rate, seroprotection rate and the GMT level. In addition, the serological responses after the vaccination of the elderly over 75 years old is even worse than those under 75 years old. According to the data of 2015–2016 influenza epidemic season in Poland, a study reported that the GMT level and seroprotection rate of children aged 0–4 and the elderly over 65 were both lower, suggesting the susceptibility of these two age groups to influenza.⁸² In a retrospective study in Scottish, based on data from nine influenza seasons, the VE was 57.1% [95% confidence interval (CI) = 31.3–73.3%] for all age groups and 59.6% (95% CI = 21.9–79.1%) for people under 65 years old. For old people over the age of 65, the VE was only 18.8% (95% CI = −103.7–67.6%), which was much lower than that in the non-elderly groups.⁸³ Unlike young children, the underlying cause of the low responsiveness to vaccines in old people is age-dependent decrease in immunological competence, often referred to as ‘‘immunosenescence’’.⁸⁴ Aging has a series of negative effect on a variety of molecules, cells, and structures in the body. With increased age, there may be a degeneration of thymus, decrease in phagocytosis, downregulation of antigen presentation, deficiency of B cells and T cells, imbalance of Th1/Th2 cytokines, and progressive deterioration of innate and adaptive immune responses.^85–87 Panda et al.⁸⁸ found that the production of immune-related factors such as TNF-α, IL-6, IL-12, and IFN-α in the elderly (>65 years old) was significantly lower than that in the young adults (21–30 years old), suggesting that aging is associated with the imbalance of cytokine levels and may reduce the immune responses. The effects of immunosenescence have also been demonstrated by numerous studies in mouse model.^89,90 In a word, it is of great significance for the public health to popularize the coverage of influenza vaccination among the elderly and provide them with a more optimized vaccination strategy due to their reduced capacity for antiviral infection and poorer responsiveness to vaccines.

Body mass index (BMI)

Sagawa et al.⁹¹ found that for people aged over 65 with a BMI<18.5 and a weight loss ≥5% within 6 months had a poor immune response to influenza vaccine. Another study of 48 diabetic patients showed a dose–response relationship between BMI index and serological response to vaccination. The proportion of subjects whose specific antibody level to A/H1N1 after vaccination increased ≥4 times in patients with BMI<22.1, BMI = 22.1–23.8 and BMI>23.9 were 62, 69, and 100%, respectively, and the difference was still statistically significant (P = .008) after adjusting for confounding factors.⁸⁰ Except for low BMI, high BMI and obesity are also risk factors for low responsiveness to influenza vaccination. It has been reported that compared with individuals with normal body weight, obese individuals (BMI>30) have decreased CD8 + T cell activation, decreased expression of functional proteins, and a steeper decline in vaccine antibody over time, indicating that obesity may impair the body’s ability to induce protective antibodies.⁹² American researcher Honce⁹³ also found that obesity may not only lead to weakened immune responses, but also is associated with increased risk of influenza infection and serious complications. This perspective has also been validated in mouse models of some related studies.^94,95 Based on the study of peripheral blood mononuclear cells cultured in vitro, researchers concluded that obesity is related to the deficiencies in activation and function of CD4 + T cells and CD8 + T cells. Compared with healthy weight individuals, obese people had lower levels of activation markers such as CD69, CD28, CD40 ligand, and IL-12 receptor expressed by CD4 + T cells and CD8 + T cells, and lower levels of functional biomarkers including IFN-γ and granzyme B.⁹⁶ In addition, differences in the richness and diversity of gut microbiome between obese and healthy people may also contribute to their different responses to vaccines.⁹⁷

Unhealthy condition

Health status is also a factor that affects the body’s immunity. The comparison of immune response to influenza vaccines in people with different health conditions are listed in Table 1. A study of 70 patients with hematological malignancies found that only 39.3% of patients with B-cell malignancies (P < .001), 45.5% with allogeneic stem cell transplant recipients (P < .001) and 85.0% of patients with chronic myeloid leukemia (P = .086) were found to produce protective antibodies, compared with healthy controls whose antibody titers were all reached protective level after vaccination.⁹⁸ A prospective study on immunocompromised children with autoimmune diseases, AIDS, congenital immunodeficiency, or prematurity from Meier et al.⁹⁹ showed that the GMT and seroprotection rate of immunocompromised children were lower than that of immunocompetent children. For most organ transplant recipients, they can only induce moderate immune responses after the vaccination. According to a study conducted in 66 stable renal transplant recipients and 19 healthy volunteers, researchers found that compared with healthy volunteers, postvaccinal humoral responses to A/H1N1 and A/H3N2 strains were less frequent in transplanted patients (H1N1: 9.5% vs. 57.9%, P < .001; H3N2: 21.5% vs. 57.9%, P = .004)¹⁰⁰; Another study conducted in Germany reported that the influenza vaccine does not provide a protective immune response in the majority of kidney transplant recipients¹⁰¹; Duchini et al.¹⁰² also found that liver transplant recipients were less responsive to A/H3N2 influenza vaccine than healthy adults. For diabetics, some studies have shown that their immune responses to influenza vaccine are comparable to those of normal people.^103,104 While analyzing the postvaccinal antibody level and seroprotection rate of 44 patients with Duchenne muscular dystrophy (DMD) and 41 healthy healthcare workers, Japanese researchers concluded that there is no significant difference between these two groups.¹⁰⁵ For AIDS patients, study showed that influenza vaccine may induce them to produce antibodies comparable to those of healthy people, suggesting the applicability of influenza vaccine for AIDS patients.¹⁰⁶

Table 1. Immune response to influenza vaccines in people with different health conditions

Reference	Study population	Country	Sample size	Season	GMT (95%CI)	Seroprotection /seroconversion rate (%) (95%CI)	Test strains
de Lavallade H. et al.(2010)^98	B-cell malignancies vs. recipients of allogeneic hematopoietic stem cell transplant vs. chronic myeloid leukemia vs. healthy adults	UK	28 vs.22 vs.20 vs.24	2009	17.7 (8.7-35.7) vs. 41.8 (14.5-120.3) vs. 100.4 (54.2-186.0) vs. 362.0 (216.4-605.5)*	39.3 (21.2-57.4) vs. 45.5 (24.6-66.3) vs. 85.0 (69.4-100) vs.100*	A/H1N1+ AS03 adjuvant
Meier S. et al. (2011)^99	Immunocompromised children vs. immunocompetent children	Switzerland	9 vs.51	2009-2010	237(61-921) vs. 350(251-487)*	89(52-100) vs. 98(90-100)*	A/H1N1
Candon S. et al.(2009)^100	Recipients of renal transplant vs. healthy adults	France	66 vs.19	2005-2006	573 ± 417 vs. 821 ± 341^a*	9.5% vs. 57.9% ^b *	A/H1N1
					127 ± 230 vs. 139 ± 111 ^a*	21.5% vs. 57.9% ^b *	A/H3N2
					174 ± 159 vs. 184 ± 106 ^a*	19.7% vs. 26.3% ^b *	B
Brakemeier S. et al.(2012)^101	Renal allograft recipients vs. healthy adults	Germany	58 vs.20	2009-2010	96.5 (72.1–120.9) vs. 406.4 (366.3–445.7)*	34.5% vs. 91 %*	A/H1N1+ AS03 adjuvant
Duchini A. et al.(2001)^102	Liver Transplant Recipients vs. patients With Cirrhosis vs. healthy adults	USA	20 vs. 14 vs. 9	1999-2000	52 (24-109) vs. 61 (30-125) vs. 65 (28-149)	25% vs. 43% vs. 56%	A/H1N1
					56 (27-115) vs. 216 (95-489) vs. 321 (144-710)*	15% vs. 21% vs. 89%*	A/H3N2
					43 (23-81) vs. 119 (48-292) vs. 92 (37-225)	30% vs. 69% vs. 56%	B
Diepersloot R. J. et al.(1987)^103	Diabetic patients vs. healthy adults	Netherlands	145 vs.24	1987	-	64% vs. 66%^c	A/H1N1
			110 vs.20		-	85% vs. 90%^c	A/H3N2
			126 vs.24		-	57% vs. 50%^c	B
el-Madhun A. S. et al.(1998)^104	Juvenile diabetic patients vs. healthy juveniles	Norway	5 vs.5	1998	211 vs. 139	100% vs. 100%	A/H1N1
					184 vs. 211	100% vs. 100%	A/H3N2
					121 vs. 211	80% vs. 80%	B
Saito T. et al. (2015)^105	Duchenne muscular dystrophy patients vs. healthy healthcare workers	Japan	44 vs. 41	2009	75 vs. 69	70% vs. 73%	A/H1N1
Agrati C. et al. (2012)^106	HIV patients vs. healthy healthcare workers	Italy	67 vs. 65	2012	109.9 vs.106.9	100% vs. 98.4%	A/H1N1+ MF59 adjuvant

Seroprotection rate = percentage of subjects with post-vaccination titer≥40; Seroconversion rate = percentage of subjects with either a pre-vaccination HI titer <1:10 and post-vaccination HI titer≥1:40 or a pre-vaccination HI titer≥1:10 and a minimum 4-fold increase in post-vaccination HI titer; “-”: Data not shown; ^a Mean antibody titers (x ± SD); ^b Percentage of subjects with post-vaccination titer≥1:50; ^c Percentage of subjects with post-vaccination titer>1:100; *P < .05;

Abbreviations: GMT = geometric mean titer; CI = confidence interval; HI = hemagglutination inhibition.

Habits and customs

Living habits such as smoking, drinking, staying up late, diet, and so on have a great impact on the immune system. Godoy et al.¹⁰⁷ found that among the elderly over the age of 65, smoking may greatly reduce the VE and was associated with the increased rates of influenza-related hospitalizations (OR = 1.32, 95%CI = 1.04–1.68). The VE was 21% in current smokers/ex-smokers and 39% in nonsmokers. In Spain, a multicenter case–control study conducted in old people also came to a similar conclusion, researchers in this study suggested that those with a history of smoking should be recommended as one of the target populations for routine influenza vaccination. Among adults over 18 years of age, current smokers had a higher risk of influenza-related hospitalization than ex-smokers (OR = 2.18 vs. 1.73), indicating that the health effects of smoking do not disappear, but will decrease over time.¹⁰⁸ Horvath et al.¹⁰⁹ found that the activity of NK cells and the levels of interferon-γ in smokers were inhibited after influenza vaccination. Experimental animal model studies have also shown that smoking impairs the respiratory immune system, leads to a decrease in cytokines such as interferon and CD4 + T cells, dampens antiviral signaling in small airway epithelial cells, and thus downregulates the body’s ability to defend against pathogens.^110–113 Besides smoking, long term and high alcohol consumption is also a risk factor for suppressing the immunity and increasing the body’s susceptibility to pathogens,¹¹⁴ while moderate alcohol consumption was associated with reduced inflammation and better immune responses to some vaccines.¹¹⁵ Researchers found that alcohol may influence the immune system by regulating the balance and interaction of host microbiome.¹¹⁶ At present, although the evidence on the association of lifestyle habits with the immune responses to influenza vaccination is not sufficient, existing studies still suggests that it is necessary to take the habits and customs into consideration while analyzing the effectiveness and immunogenicity of influenza vaccines.

Microbiome

More and more evidence has proven the important role of microbiome in modulating human immunity.^117,118 Zimmermann et al.¹¹⁹ found that the adaptive immune responses were positively associated with the relative abundance of the phylum actinobacteria and firmicutes but negatively correlated with the phylum Proteobacteria and Bacteroidetes. As for the antibody responses to influenza vaccines, Thomas Hagan et al.⁹⁷ found that compared with the control group, the antibiotic group had lower levels of neutralizing antibodies to A/H1N1, indicating that the application of antibiotics may impair the immunity by reducing the richness and diversity of gut microbiome. Another study showed that the serological responses to TIV of germ-free or antibiotic-treated mice were impaired, but can be restored by taking a strain of E. coli orally.¹²⁰ Toll-like receptor 5 (TLR5) is an important immune-related protein molecule, which is expressed in bone marrow-derived cells such as mononuclear macrophages. Many studies have confirmed the positive correlation between the expression of TLR5 and the antibody responses.^120–123 Jason et al.¹²⁰ found that compared with wild-type mice, TIV-specific IgG, and IgM antibody responses were significantly reduced in Tlr5^−/- mice. Bacterial flagellin is the only ligand of TLR5, thus, the microbiome may influence the immune responses to influenza vaccine by regulating the expression of TLR5. Emerging evidence proved that the immunogenicity of influenza vaccines can be improved by using flagellin as an adjuvant.^124–128 In addition to the TLR5-mediated pathway, the microbiome may also directly or indirectly affect the antibody responses through a variety of ways, such as hormones and immune-related metabolites.¹²⁹ To further explore the specific mechanisms of microbiome on immunity, other related factors should also be considered.

Effect of repeated vaccination

For people with low immunity, repeated vaccination in the same influenza season may improve the protective effect of the influenza vaccine to some extent. In the study of 70 patients with hematological malignancies mentioned above, researchers also found a significant increase in the seroprotection rates from 39% to 68% (P = .008) in patients with B-cell malignancies and from 45% to 73% (P = .031) in allogeneic SCT recipients after the second vaccination, suggesting that the booster immunization could significantly improve the antibody level of patients.⁹⁸ Another study also showed that the second vaccination may upregulate the immunogenicity of influenza vaccine in the elderly aged over 61 years old. The seroprotection rate of subjects was 79.1% 21 days after the first dose, but rose to 93.3% 14 days after the second dose (35 days after the first dose).¹³⁰ However, unlike repeated vaccination during the same flu season, continuous annual vaccination may lead to reduced VE and immunogenicity of vaccines. When encountering a virus strain similar but not identical to the one previously infected, the immune system tends to induce a strong anamnestic response only to the original-infected strain, and may also diminish the production of memory cells stimulated by the subsequent one.¹³¹ Studies reported that prior year vaccination may reduce the effector B-cell responses to new TIV immunization.^132,133 In an influenza vaccination cohort study of pregnant women, researchers found that higher baseline antibody levels (P < .001) and prior year influenza vaccination (P = .03) were both statistically associated with reduced odds of seroconversion.¹³⁴ Thompson et al.¹³⁵ also concluded that repeated annual vaccination was related to the decrease of geometric mean fold change ratios (GMRs) to A/Perth/16/2009-like virus, suggested that previous vaccination history may have a negative impact on the immune responses of the follow-up vaccination. Surender Khurana et al.¹³⁶ also confirmed that, regardless of the vaccine platform, the antibody-affinity maturation to HA of all three influenza virus strains can be reduced by repeated annual influenza vaccination. The specific mechanism of how preexisting immunity affects immunogenicity is still unclear and requires further exploration.

Effect of genetic factors

For individuals with basically the same individual factors such as age and sex, even if they are vaccinated with the same vaccine, the immune responses generated following the vaccination are often different, indicating that genetic factors play an important role in this process. Single nucleotide polymorphisms (SNPs), as one of the most common types of heritable variation in humans, has been shown to have a significant impact on the immune responses to vaccines.^137–139

SNPs in HLA

HLA is a polygenic and polymorphic complex located on the short arm of chromosome 6, which plays an important role in antigen recognition. After the invasion of pathogens, it is first absorbed and degraded into polypeptides by APC, and then expressed on the APC surface in the form of peptide-HLA (p-HLA) complex, which transmits the information of antigen to CD4 + T cells to initiate the whole immune response process.^140,141 A number of studies have reported the relationship between SNPs in HLA and the immune responses to influenza vaccination: As early as 1976, researchers found that the HLA-A Type W16 allele was associated with lower levels of influenza-specific antibodies.¹⁴² An RCT study conducted in 185 elderly people from Moss et al.¹⁴³ found that compared with the elderly who fail to reach the seroprotection level, the frequencies of HLA-DRB1*04:01 allele and HLA-DPB1*04:01 allele were higher in those who reached the seroprotection level. British researchers found an increase in the frequency of HLA-DRB1* 0701 (P = .016) and a decrease in the frequency of HLA-DQB1*0603-9/14 (P = .045) in non-responders who received TIV.¹⁴ Based on the analysis of immune responses of healthy Caucasian men aged 18–40 years after influenza vaccination, Poland et al.¹⁴⁴ concluded that HLA-A*1101 (P = .0001), HLA-A*6801 (P = .09), HLA-B*3503 (P = .02), HLA-B*1401 (P = .06), HLA-C*0802 (P = .05), HLA-DRB1*1104 (P = .04), HLA-DRB1*1601 (P = .02) and HLA-DQB1*0502 (P = .03) alleles were correlated with a higher median of specific antibodies to H1N1, while the HLA-DRB1* 1303 allele (P = .04) was associated with lower antibody titers to H1N1. It has also been reported that the HLA-DR7/4, DQB1*0302 genotype (OR = 5.15; 95%CI = 1.94–13.67; P = .001) of non-Hispanic white children and the HLA-DR7/Y (including DRB1 11, DRB1 13 and DRB1 14) genotype (OR = 5.84; 95%CI = 1.68–20.28; P = .006) of Hispanic children were correlated with low responsiveness to TIV vaccination.¹⁴⁵ Currently, the specific mechanisms by which gene variants in HLA influence the immune responses are not clear. Some scholars speculate that individuals with different HLA genotypes may produce HLA molecules with various antigen affinity through transcription and translation, thereby affecting the efficiency of antigen presentation.

SNPs in cytokines

Cytokines act as cell-signaling molecules between immune cells and play an important role in both innate and adaptive immunity. Human Functional Genomics Project (HFGP) showed that for approximately 70% of peripheral blood mononuclear cells, the effect of genetic factors on their cytokine responses varies from 0.15 to 0.75.¹⁴⁶ Many studies have confirmed that SNPs outside HLA region, such as those on cytokines, also have important effects on the immune responses to influenza vaccination.^144,147 The production of specific antibodies is a process mediated by lymphocytes and requires the assistance of T helper cells (Th). Cytokines secreted by Th1 cells, such as tumor necrosis factor-α (TNF-α), interleukin-2 (IL-2) and interferon-γ (IFN-γ), are mainly involved in cellular immunity and delayed type hypersensitivity. These cytokines play an important role in the body’s inflammatory responses caused by intracellular microbial infection and cellular immunity. While cytokines secreted by Th2 cells, such as IL-4, IL-5 and IL-10, are mainly involved in humoral immunity and regulate the immune responses against extracellular microbial and parasitic infections, as well as allergic reactions. Poland et al.¹⁴⁴ found that IL-6 rs1800796 GG genotype and IL-6 rs2069861 AA genotype were associated with higher antibody titers, while IL-12B rs3212227 CC genotype, TNF receptor superfamily member 1A (TNFRSF1A) rs4149621 G allele, and IL-1 receptor-1 (IL1R1) rs3732131 GG genotype were correlated with lower antibody levels. A study performed from Canadian researchers reported that IL-28B rs8099917 TG+GG genotype was associated with the increase of seroconversion rate after influenza vaccination (OR = 1.99, P = .038), and proposed that IL-28B is a key regulatory factor to maintain the balance of Th1/Th2.¹⁴⁷

SNPs in immune-related molecules

The immune system is a complicated system, which requires regulations from a series of immune-related molecules. Genetic variants in these molecules may also influence the immune responses. Leptin is a kind of proteohormone secreted by adipose cells, which can regulate the activity of energy metabolism, reproductive development and immune system. A study from the Mayo Clinic concluded 8 SNPs in leptin and its receptor were statistically correlated with the specific antibodies induced by influenza vaccination.¹⁴⁸ Since there is a strong association between the leptin levels and BMI, this study also suggests that obesity may also affect the immunity. Another study on interferon-induced transmembrane protein-3 (IFITM3) showed that the seroconversion rate of subjects with IFITM3 rs12252 CC genotype on day 14 after influenza vaccination was 48.6%(17/35), which was much lower than that of subjects with TT genotype (78.6%, 22/28, P = .015), indicating that TT genotype may provide a better immune protection for its carriers.¹⁴⁹ Another study conducted in 171 healthy volunteers with 3 different IFITM3 rs12252 genotypes from Na Lei et al.¹⁵⁰ showed that, the seroconversion rate to H1N1, H3N2, and B of IFITM3 rs12252 CC carriers were all lower than that of CT and TT carriers. Subsequently, they also confirmed that IFITM3 deletion attenuated the antibody response to influenza vaccination in mouse model. Heme oxygenase (HO) is an enzyme involves in the physiological activity of various cytokines and immune cells. A cohort of 147 healthy human subjects carried from Cummins et al.¹⁵¹ showed that HMOX1 rs743811 G allele (P = .017) and HMOX2 rs2160567 G allele (P = .014) were both related to poorer vaccine immune response.

The expression and function of HLA, cytokines and other immune-related molecules are all regulated by their gene variants. Nowadays, although a lot of SNPs have been confirmed to affect the immune responses to influenza vaccination, few studies have been conducted at the molecular level. Besides, further genome-wide studies should also be carried out identify more immune-related variants.

Conclusions

Vaccination is the most effective way to prevent and control the prevalence of influenza. However, due to the limited research progress of broad-spectrum universal influenza vaccine and relatively few cases of severe infection, influenza vaccination still has a narrow coverage in the population. Many “vulnerable groups” are still facing a high risk of infection in influenza seasons. In order to provide better protection for people and provide more clues for future research, it is of great significance to fully analyze the relevant factors that may affect the specific antibody responses, to identify low and non-responders to influenza vaccination, to actively develop more efficient new vaccines, and to improve the individualized vaccination program.

Competing interests

None declared.

Additional information

Funding

This work was supported by the Special Funds of the National Natural Science Foundation of China [Grant No. 82041043].