The presented protocols describe how to perform a hemagglutination inhibition assay to quantify influenza-specific antibody titers from serum samples of influenza vaccine recipients. The first assay determines optimal viral antigen concentrations by hemagglutination. The second assay quantifies influenza-specific antibody titers by hemagglutination inhibition.
Antibody titers are commonly used as surrogate markers for serological protection against influenza and other pathogens. Detailed knowledge of antibody production pre- and post-vaccination is required to understand vaccine-induced immunity. This article describes a reliable point-by-point protocol to determine influenza-specific antibody titers. The first protocol describes a method to specify the antigen amounts required for hemagglutination, which standardizes the concentrations for subsequent usage in the second protocol (hemagglutination assay, HA assay). The second protocol describes the quantification of influenza-specific antibody titers against different viral strains by using a serial dilution of human serum or cell culture supernatants (hemagglutination inhibition assay, HI assay).
As an applied example, we show the antibody response of a healthy cohort, which received a trivalent inactivated influenza vaccine. Additionally, the cross-reactivity between the different influenza viruses is shown and methods to minimize cross-reactivity by using different types of animal red blood cells (RBCs) are explained. The discussion highlights advantages and disadvantages of the presented assays and how the determination of influenza-specific antibody titers can improve the understanding of vaccine-related immunity.
Infection with influenza virus is associated with considerable morbidity, mortality, and high healthcare costs1,2,3,4. In particular, elderly, newborns, pregnant women, and patients with chronic disease are at risk for more severe clinical outcomes. Therefore, vaccination against circulating influenza virus strains is the primary measure to decrease the burden of disease in these high-risk populations. The increase of the individual immune response after vaccination, e.g., influenza-specific antibodies above a protective threshold, reduces the individual risk of infection and in general the likelihood of viral transmission within a population5. A detailed understanding of the vaccine-induced humoral immune response in different populations and across various age groups is a key element to answer important clinical questions6,7,8,9, such as: Why do some elderly patients have infections despite previous vaccination? What is a "good" and "sufficient" vaccine-induced protection? How often should a vaccine be applied to an immunosuppressed patient to reach protective titers? What is the most effective dosage? What is the impact of a novel adjuvant on post-vaccination antibody titers? The measurement of the vaccine-specific antibody production may help answer these important questions and improve vaccination outcomes.
The quantification of virus-specific antibody titers can be performed with various immunological methods. This includes solid-phase10 or bead-based ELISA11 assays, the HI assay12, and neutralizing assays13. ELISA-based methods allow the screening of relatively large amounts of serum samples against various antigens. Also, pathogen-specific Immunoglobulin (Ig)M and IgG can be separately explored. Although the characteristics of an antigen, e.g., the linear amino acid sequence or virus-like particle may influence the binding of antibodies, the spectrum of potential epitopes is very broad and does not provide information on whether an antibody response has functional relevance.
In contrast, the neutralization assay determines the potential of antibodies to functionally inhibit the infection of cells and therefore reflects the neutralization potential. However, this method is very labor intensive, requires culturing of specific cell lines and live viruses, and therefore, it is time-consuming, expensive, and requires special equipment.
This article describes a step-by-step the World Health Organization (WHO)-based HI protocol12 to quantify influenza-specific antibody titers. Hemagglutination is a characteristic effect of some viruses leading to the agglutination of erythrocytes. The inhibition of this effect with patient sera allows the measurement of inhibitory antibody concentrations, which reflects a neutralizing effect.
We have modified the workflow of the WHO-protocol to allow a more efficient handling of multiple samples at the same time and thereby reducing the required time. The first protocol describes the determination of the agglutination potential of a particular influenza antigen. In doing so, the correct influenza antigen concentration is determined for the second protocol. This part should be repeated with every new viral antigen, as well as each batch of blood.
The second protocol describes the determination of influenza-specific antibody titers. The presented protocols are optimized for the investigation of influenza virus and human serum samples however, it can also be applied for mouse serum samples or cell-culture supernatants from stimulated immune cells, e.g., virus-specific B-cells. Results can be determined as absolute measured titers. In many vaccine studies, the geometric mean titers and the 95% confidence interval are shown for each particular population. For interpretation, seroprotection or seroconversion are often used to describe the susceptibility of a population to a certain virus. Seroprotection is defined as a titer of ≥1:40, and seroconversion as a more than 4-fold titer increase with achievement of seroprotective titers between two time points (most commonly pre-vaccination and 30 days post-vaccination are used).
Both protocols are easy to use and they can be adapted to a broad range of research questions. In particular, they can be used to determine reliably and quickly the antibody titers against various other viruses with the capacity for hemagglutination, such as measles, polyomaviruses, mumps, or rubella14,15,16.
The study protocols were approved through the local ethical review board (www.EKNZ.ch) and written informed consent was obtained from all participants.
1. Serum Collection
2. Preparation of Antigens
CAUTION: Five different antigens are used (see Table of Materials). Prepare antigens in a Biosafety Level 2 (BSL-2) laboratory.
3. Preparation of Cholera Filtrate
NOTE: Cholera filtrate is used as a receptor destroying enzyme (RDE) according to the WHO protocol12. This removes innate inhibitors from the serum which would interfere with the assay17.
4. HA Assay
NOTE: To ensure that the HI assays are comparable between several plates, the same amount of virus particles must be used for each plate. The HA assay (also called HA titration) is performed to quantify the virus particles necessary for hemagglutination, and is recorded in HA units. A "unit" of hemagglutination is an operational unit dependent on the method used for HA titration and is not a measurement of an absolute amount of virus. Thus, an HA unit is defined as the amount of virus needed to agglutinate an equal volume of a standardized RBC suspension. According to the WHO, the standard amount used for the HI assay is 4 HA units per 25 µL. For an illustration of the principle of the HA assay see Figure 1.
Figure 1: Principle of hemagglutination and hemagglutination inhibition. No hemagglutination occurs in a negative control situation without viruses and antibodies (left column), and erythrocytes hemagglutinate only in the presence of influenza virus (middle column). However, when the hemagglutinin of the influenza virus is blocked by virus-specific antibodies then no hemagglutination can occur (right column). Please click here to view a larger version of this figure.
NOTE: The RBCs used are dependent on the type of influenza virus in the assay (Table 1). Further, for various types of 96-well micro titer plates, the incubation time as well as the appearance of the non-agglutinated cells differ (Table 2).
Influenza antigen | A/California/7/09 (H1N1) | A/Switzerland/9715293/2013 (H3N2) | A/Texas/50/2012 (H3N2) | B/Brisbane/60/08 | B/Massachusetts/02/2012 | ||
RBC species | Chicken | Guinea Pig | Guinea Pig | Turkey | Turkey |
Table 1: Influenza antigens and corresponding species of RBCs. According to the manufacturer's instructions (NIBSC).
RBC species | Chicken | Turkey | Guinea pig | Human type O |
Concentration of RBCs (v/v) | 0.75% | 0.75% | 1% | 1% |
Type of microtiter plate | V bottom | V bottom | U bottom | U bottom |
Incubation time, RT | 30 min | 30 min | 1 hour | 1 hour |
Appearance of non-agglutinated cells | Button* | Button* | halo | halo |
Table 2: Assay conditions with different species of RBCs. According to the WHO protocol. (* flows when tilted).
Figure 2: Plate design of the HA assay. The HA titration is performed in duplicates. No antigen was added to the control rows. Also see Figure 4 for the determination of the best antigen concentration. Please click here to view a larger version of this figure.
Figure 3: Agglutination patterns of avian and mammalian RBCs. V-shaped micro titer plates are used when working with avian RBCs. The readout is performed in a tilted plate position, and non-agglutinated RBCs start to run down forming a tear-like shape. U-shaped microtiter plates are used when working with mammalian RBCs. The readout is then performed in a non-tilted position, and non-agglutinated RBCs form a small halo. Please click here to view a larger version of this figure.
Figure 4: Readout of the HA titration with avian RBCs to determine the titer of 4 HA units. The optimal antigen amount required for hemagglutination is measured by the hemagglutination assay (antigen titration assay). The last well where complete hemagglutination occurs is the HA titration endpoint and contains 1 HA unit. Because of the 2-fold dilutions of the antigen, two wells ahead of the HA titration endpoint, the titer corresponds to 4 HA units. Please click here to view a larger version of this figure.
5. HI Assay
NOTE: The work-flow of the protocol has been optimized to allow a more efficient handling of multiple samples at the same time, by using PCR tube stripes and a thermo cycler (see below).
Figure 5: Plate design and workflow of the HI assay. Five time points of two people can be measured on one plate. The HI titer ranges from 8 to 1,024. An anti-serum of the used antigen served as a positive control and a back titration was performed to check if the antigen dilution equals 4 HA units. The serial dilution of the serum sample is shown for 2 individual vaccine recipients. Please click here to view a larger version of this figure.
Figure 6: Readout of the HI assay with avian RBCs. The pre- and post-vaccination induced influenza specific antibody response is determined by HI assay. In this example, person one has higher HI titers than person two. Both persons show an antibody response after vaccination; 180 days after the vaccination the antibody titers of both persons are decreased again. Please click here to view a larger version of this figure.
Pre- and post-vaccination induced antibody response against Influenza A H3N2
The vaccine-induced antibody response was assessed in 26 healthy volunteers who received an inactivated trivalent subunit influenza vaccine containing Influenza A/H1N1/California/2009, A/H3N2/Texas/2012, and B/Massachusetts/02/2012 prior to the 2014/2015 influenza season. Figure 6 shows a representative example of 2 vaccine recipients. Interestingly, during that particular influenza season, A/H3N2/Texas/2012 was not circulating, and instead the season included the somewhat different viral strain: A/H3N2/Switzerland/2013. The viral hemagglutinin of A/H3N2/Texas/2012 and A/H3N2/Switzerland/2013 show 97% sequence identity and differ in only eleven amino acids (see Table 4), whose positions are highlighted in Figure 7.
Figure 7: Hemagglutinin comparison of A/H3N2 influenza strains. We compared the hemagglutinin of the viral strains A/Texas/50/2012 and A/Switzerland/9715293/2013. Since there are no hemagglutinin crystal structures of these strains, we used the crystal structure of the highly similar hemagglutinin of the influenza strain A/Victoria/361/201118, which shows 98% sequence identity with the Texas strain and 95% sequence identity with the Switzerland strain. The amino acid positions in which the Texas and Switzerland strains differ are highlighted. Please click here to view a larger version of this figure.
We observed a cross-reactive immune response for the viral strains A/H3N2/Switzerland/2013 and A/H3N2/Texas/2012. HI titers against Influenza A/H3N2/Switzerland/2013 were significantly lower in terms of geometric mean titers and induced seroprotection (Figure 8A) in comparison to Influenza A/H3N2/Texas/2012 (Figure 8B).
Figure 8: Geometric mean antibody-titers of healthy donors. The geometric mean antibody-titers (GMTs) of 25 healthy donors pre- and post-vaccination are determined using two different antigens. The mean titers of A/H3N2/Switzerland/2013 (A) and A/H3N2/Texas/2012 (B) are shown. An immune response due to the vaccination can be observed as increasing titers after the vaccination (d7-d60), compared to the GMTs before the vaccination (d0). 180 days after the vaccination, the GMTs decrease again. Of note, only A/H3N2/Texas/2012 (which was in the vaccine) reaches protective titers. Bars indicate geometric mean titers, and whiskers indicate the 95% confidence intervals. The dashed line indicates the seroprotection threshold. The % of seroprotected people (titer >1:40) is shown in the graph. Please click here to view a larger version of this figure.
After vaccination, the antibody titers against A/H3N2/Texas/2012 increased in most subjects; although the A/H3N2/Switzerland/2013 strain was not present in the vaccine, the titer against A/H3N2/Switzerland/2013 increased in some subjects as well. Figure 9 shows the correlation between both titers over all time points with an R2 of 0.745 for a linear regression model. As one would expect, the induction of the antibody response against A/H3N2/Switzerland/2013 was less potent.
Figure 9: Cross-reaction between A/H3N2 influenza strains. The A/H3N2/Texas titers of every individual and time point are plotted against the corresponding A/H3N2/Switzerland titers. A linear regression model shows an R2 of 0.745. Please click here to view a larger version of this figure.
Hemagglutination potential is based on the type of blood used
The viral hemagglutinin shows different species-dependent potential to hemagglutinate erythrocytes. This species-dependent effect also impacts the hemagglutination inhibition assay. To improve the specificity of measured anti-viral titers, we evaluated the best suited type of erythrocytes for five viral antigens (Influenza B/Brisbane/60/2008 and B/Massachusetts/02/2012, Influenza A/H1N1/California/2009, A/H3N2/Texas/2012, and A/H3N2/Switzerland/2013) to achieve the maximum hemagglutination but also the lowest cross-reactivity. We used positive control sera from the National Institute for Biological Standards and Control (NIBSC) against each antigen to perform these assays.
For Influenza B, we could observe that the B/Massachusetts/02/2012 induced antibody response does not provide protection against B/Brisbane/60/2008. In contrast, antibodies against B/Brisbane/60/2008 showed cross-reactivity against B/Massachusetts/02/2012 at a 4-fold lower titer across different erythrocytes (see Table 3). Of interest, guinea pig blood did not properly hemagglutinate with Influenza B. Turkey blood did best in showing the potential to hemagglutinate and the highest titers with relative low cross-reactivity apart from the previously mentioned A/H3N2/Texas and /Switzerland strains.
Turkey | Guinea Pig | Chicken | Human type O | |
B/Brisbane | 1024 | – | 1024 | 1024 |
B/Massachusetts | 1024 | 384 | 768 | 1024 |
A/H3N2/Switzerland | 1024 | 1024 | – | 1024 |
A/H3N2/Texas | 1024 | 1024 | 512 | 1024 |
A/H1N1/California | 1024 | 1024 | 768 | 768 |
Table 3: Positive control titers against the respective influenza HA antigen across different species.
No | A/H3N2/Texas/2012 strain | A/H3N2/Switzerland/2013 strain | Position |
1 | Asparagine (N) | Alanine (A) | 128 |
2 | Alanine (A) | Serine (S) | 138 |
3 | Isoleucine (I) | Arginine (R) | 140 |
4 | Arginine (R) | Glycine (G) | 142 |
5 | Asparagine (N) | Serine (S) | 145 |
6 | Phenylalanine (F) | Serine (S) | 159 |
7 | Glycine (G) | Valine (V) | 186 |
8 | Proline (P) | Serine (S) | 198 |
9 | Serine (S) | Phenylalanine (F) | 219 |
10 | Asparagine (N) | Aspartate (D) | 225 |
11* | Lysine (K) | Arginine (R) | 326 |
*not shown in the crystal structure in Figure 8, because the hemagglutinin was cutted at residue 325. |
Table 4: List of different amino acids of hemagglutinin between A/H3N2/Texas/2012 and A/H3N2/Switzerland/2013 strains
Quantification of pre- and post-vaccination influenza virus specific antibody titers is an important tool necessary for vaccine studies. Based on the surrogate measures of protection against virus infection, such as seroprotection (>1:40) or seroconversion (4-fold titer increase), vaccination strategies can be optimized9. Using the provided protocols can determine: (i) the hemagglutination potential of a particular virus, and (ii) the antibody titers for a virus of interest.
Modification and Troubleshooting:
This protocol is based on the WHO standard12. We modified the protocol by using PCR tube strips for serum preparations (see step 5). This modification helped to significantly reduce the workload and to increase the throughput of the assay. Further, we reduced the antigen amount by one fourth in the antigen titration step, which is cost effective overtime. A lower amount of serum (10 µL) can be used for the RDE treatment, which helps especially when the sample amount is limited (e.g., mouse sera). The back titration and positive control are included in the antibody measurement plate to serve as a proper internal control and to monitor the aging of erythrocytes.
In addition, we used different erythrocyte concentrations than those in the WHO standard12 to set the optimal size of the erythrocyte clot for a good visual readout. To guarantee this we suggest checking the erythrocyte concentration before the assay. Although, we have not optimized this part in our protocol, methods such as absorbance measurement with OD or cell counting could be used.
If the RDE is not completely inactivated, RBCs can be desialylated and reverse HA-positive wells when hemagglutination is measured at room temperature. Although we never observed this problem, in this case, we suggest performing the HI at 4 °C, since RDE activity is significantly lower at 4 °C. However, performing the HI assay at 4 °C is slower.
Limitations of the Technique:
A few critical aspects of the HI assay include the following points: Interestingly, the hemagglutination is strongly dependent on the particular type of erythrocyte (e.g., turkey or guinea pig RBC). The optimal type of blood should be tested before a particular virus strain is evaluated and the same type of blood should be used throughout the assay. Although the induction of cross-reactivity between viruses may generate an immunological advantage in the case of a slightly new virus19,20, this may cause some problems from a diagnostic point of view due to low specificity. Therefore, cross reactivity between similar viruses should be carefully addressed and discussed in studies. By choosing erythrocytes from a specific species, the amount of cross-reactivity can be somewhat lowered.
Significance with Respect to Existing Methods:
The HI is a well-established gold standard method providing highly reproducible and reliable results. Other techniques such as ELISA may detect non-neutralizing antibodies, whereas the HI only detects antibodies which bind to the HA stem loop and thereby correlate with neutralization.
Critical Steps Within the Protocol:
The most critical steps include the serum treatment with RDE to inactivate unspecific inhibitors and binding to the HA of the virus. Another critical step is to control for hemolysis of the erythrocytes as they age over time.
Future Applications:
The protocol may be used for other viruses with hemagglutination potential. Although we have only shown results on human sera samples here, the assay can also be used to measure antibody titers in mouse sera or in cell culture supernatant with stimulated B-cells (data not shown). In summary, the HI allows a rapid and reproducible assessment of vaccine-induced antibody titers.
The authors have nothing to disclose.
none.
25 ml Disposable Multichannel Pipette Reservoirs | Integra | 4312 | |
8-well PCR tubes | Brand GMBH | 781332 | For serum aliquots |
96-well microtiter plate, U-shaped | TPP | 92097 | For HI assay when using mammalian RBCs |
96-well microtiter plate, V-shaped | Corning Costar | 3897 | For HI assay when using avian RBCs |
Aqua ad iniect. Steril | Bichsel AG | 1000004 | For preparing influenza antigen and cholera filtrate solutions |
Chicken RBC (10%) | Cedarlane | CLC8800 | 10% suspension of chicken red blood cells in Alsever's solution |
Cholera filtrate | Sigma-Aldrich | C8772 | Used as receptor destroying enzyme (RDE) |
Dulbecco's PBS | Sigma-Aldrich | D8537 | For diluting the serum samples, RBCs and antigens |
Eppendorf Multichannel pipette, 12-channel, 10-100 µl | Sigma-Aldrich | Z683949 | |
Eppendorf Multichannel pipette, 8-channel, 10-100 µl | Sigma-Aldrich | Z683930 | |
Guinea Pig RBC (10%) | Cedarlane | CLC1800 | 10% suspension of guinea pig red blood cells in Alsever's solution |
Influenza Anti-A/California/7/09 HA serum | NIBSC | 14/134 | Used as positive control at the HI assay |
Influenza Anti-A/Switzerland/9715293/2013-like HA serum | NIBSC | 14/272 | Used as positive control at the HI assay |
Influenza Anti-A/Texas/50/2012-Like HA Serum | NIBSC | 13/178 | Used as positive control at the HI assay |
Influenza Anti-B/Brisbane/60/2008-HA serum | NIBSC | 13/254 | Used as positive control at the HI assay |
Influenza Anti-B/Massachusetts/02/2012 HA serum | NIBSC | 13/182 | Used as positive control at the HI assay |
Influenza antigen A/California/7/09 (H1N1)(NYMC-X181) | NIBSC | 12/168 | Inactivated, partially purified A/California/7/09 (H1N1)(NYMC-X181) virus (ca. 46µgHA/ml) |
Influenza antigen A/Switzerland/9715293/2013 (NIB88) | NIBSC | 14/254 | Inactivated, partially purified A/Switzerland/9715293/2013 (NIB88) virus (ca. 55µgHA/ml) |
Influenza antigen A/Texas/50/2012 (H3N2)(NYMCX-223) | NIBSC | 13/112 | Inactivated, partially purified A/Texas/50/2012 (H3N2)(NYMCX-223) virus (ca. 74µgHA/ml) |
Influenza antigen B/Brisbane/60/2008 | NIBSC | 13/234 | Inactivated, partially purified B/Brisbane/60/2008 virus (ca. 42µgHA/ml) |
Influenza antigen B/Massachusetts/02/2012 | NIBSC | 13/134 | Inactivated, partially purified B/Massachusetts/02/2012 virus (ca. 35µgHA/ml) |
Serum-Tubes | S-Monovette, Sardstedt | 01.1601.100 | For serum extraction with clotting activator |
Single Donor Human RBC, Type 0 | Innovative Research | IPLA-WB3 | Suspension of single donor human red blood cells in Alsever's solution (ca. 26%) |
Turkey RBC (10%) | Cedarlane | CLC1180 | 10% suspension of turkey red blood cells in Alsever's solution |
Phosphate Buffered Saline (PBS) | Gibco |