Comparison of different cell substrates on the measurement of human influenza virus neutralizing antibodies

Abstract

Eight cell lines were systematically compared for their permissivity to primary infection, replication, and spread of seven human influenza viruses. Cell lines were of human origin (Caco-2, A549, HEp-2, and NCI-H292), monkey (Vero, LLC-MK2), mink (Mv1 Lu), and canine (MDCK). The influenza viruses included seasonal types and subtypes and a pandemic virus. The MDCK, Caco-2, and Mv1 Lu cells were subsequently compared for their capacity to report neutralization titers at day one, three and six post-infection. A gradient of sensitivity to primary infection across the eight cell lines was observed. Relative to MDCK cells, Mv1 Lu reported higher titers and the remaining six cell lines reported lower titers. The replication and spread of the seven influenza viruses in the eight cell substrates was determined using hemagglutinin expression, cytopathic effect, and neuraminidase activity. Virus growth was generally concordant with primary infection, with a gradient in virus replication and spread. However, Mv1 Lu cells poorly supported virus growth, despite a higher sensitivity to primary infection. Comparison of MDCK, Caco-2, and Mv1 Lu in neutralization assays using defined animal antiserum confirmed MDCK cells as the preferred cell substrate for influenza virus testing. The results observed for neutralization at one day post-infection showed MDCK cells were similar (<1 log(2) lower) or superior (>1 log(2) higher) for all seven viruses. Relative to Caco-2 and Mv1 Lu cells, MDCK generally reported the highest titers at three and six days post-infection for the type A viruses and lower titers for the type B viruses and the pandemic H9N2 virus. The reduction in B virus titer was attributed to the complete growth of type B viruses in MDCK cells before day three post-infection, resulting in the systematic underestimation of neutralization titers. This phenomenon was also observed with Caco-2 cells.

Figure 1. Sensitivity of eight cell lines to primary infection by A/Uruguay/716/07.

Cell lines Mv1 Lu, MDCK, Caco-2, A549, Vero LLC-MK2, NCI-H292, and HEp-2, were infected with the same serial 10-fold dilution of A/Uruguay/716/07 at ∼10,000 FFU per well (FFU based on infectivity in MDCK cells). Plates were incubated at 33°C for 18 to 22 h post-infection, then stained using strain-specific sheep polyclonal anti-HA antibody.

Figure 2. Composite analysis of replication and spread of seven different influenza types/subtypes in eight different cells.

The A549, Caco-2, HEp-2, LLC-MK2, MDCK, Mv1 Lu, NCI-H292, and Vero cells were infected with the same serial 10-fold dilutions (for virus nomenclature see Materials and Methods) to limiting dilution starting at approximately 10,000 FFU per well (FFU based on infectivity in MDCK cells). Composite data are for FFA foci number on Day 1 (D1) and HA expression at Day 3 (D3) and 6 (D6) post-infection. The results represent the average of 3 independent assays. Culture medium was supplemented with TPCK-Trypsin at 0.5 µg/ml. The FFA foci number is the average number of foci over 3 independent assays and indicates the row used for counting foci. Each row (a-g) represents a 10-fold dilution of virus and rows below the foci count would contain progressively fewer infectious particles until the dilution limit is reached. Rows above the countable foci had too many foci to accurately count and row H was used as blank. Average subjective scores (see Materials and Methods) were determined (values of 0, 0.5, 1, 2, and 3) and color coded as white (0 no HA expression), cream (0.1–1.9), orange (2–2.4) and red (2.5–3.0).

Figure 3. Trypsin-independent replication and spread of B/Florida/04/06 in MDCK, Caco-2, and Mv1 Lu cells.

Cell lines MDCK, Caco-2, and Mv1 Lu were infected with the same serial 10-fold dilution of B/Florida/04/06 starting at approximately 10,000 FFU per well (FFU based on infectivity in MDCK cells). Cells were supplemented with 0.1 µg/ml or no trypsin (as indicated) and were incubated at 33°C for 18 to 22 h post-infection and stained using strain-specific sheep polyclonal anti-HA antibody. The wells shown above are from a dilution that contained approximately 10 (MDCK cells) or 40 (Caco-2 and Mv1 Lu cells) infectious particles on Day 1, 3, and 6 post-infection.

Figure 4. NA activity in virus infected control spent supernatants at day 3 and 6 post-infection.

Approximately 100 FFU of each virus type/subtype (see Materials and Methods for virus nomenclature) were used in neutralization assays. Spent culture medium supernatants from un-neutralized virus control wells were harvested on Day 3 (gray) and Day 6 (black) and tested for NA activity. The raw fluorescence signals are plotted for each type/subtype by cell line (MDCK, Caco-2, and Mv1 Lu). The NA activity presented is the average of 3 test days. The solid black line indicates the cell control threshold for determining whether virus replication was reportable above background.

Figure 5. Neutralization of seven different influenza virus types and subtypes in MDCK, Caco-2, and Mv1 Lu cells (by cell line).

Approximately 100 FFU of each virus type/subtype (see Materials and Methods for virus nomenclature) was added to serial 2-fold dilution of normal sheep antiserum (nSheep), sheep anti-HA (aHA), sheep anti-NA (aNA), or ferret polyclonal antiserum (pre- [fPre] or post-immunization [fPost] with the specified virus strain) and incubated at room temperature for 30 min. The serum treated virus was transferred onto MDCK (panel A), Caco-2 (panel B) and Mv1 Lu (panel C) cells. Neutralization titers were determined by FFA on Day 1 (blue solid diamonds, Titer-FFA) and NA activity on Day 3 (grey open diamonds; Titer-NA-3) and Day 6 (solid grey diamonds; Titer-NA-6). Titers were determined based on the log₂ dilution that reduced the foci number or NA activity greater than or equal to 50% of that observed in virus control wells. Titers at the baseline (0.0) represent non-reportable titers, as virus growth was insufficient to make an estimate. The Day 1 and Day 3 titers are the average of 3 assays and the Day 6 titers are the average of 6 assays. The nSheep serum had a titer of <6.3 when assayed with minimal dilution. The titer reported in this chart is higher due to the starting dilution used in assays that was identical to the dilution used for the hyperimmune anti-HA antisera.

Figure 6. Neutralization of seven different influenza virus types and subtypes in MDCK, Caco-2, and Mv1 Lu cells (by days).

Neutralization titers in MDCK (blue solid triangles), Caco-2 (gray open diamonds), and Mv1 Lu (gray solid diamonds) cells were determined by FFA at Day 1 (panel A), and NA activity at Day 3 (panel B) and Day 6 (panel C). Approximately 100 FFU of each virus type/subtype were added to serial 2-fold dilution of normal sheep antiserum (nSheep), sheep anti-HA (aHA), sheep anti-NA (aNA), or ferret polyclonal antiserum (pre- [fPre] or post-immunization [fPost] with the specified virus strain) and incubated at room temperature for 30 min. Serum treated virus was then transferred onto cells and titers were determined by NA activity. Titers were determined based on the log₂ dilution that reduced the NA activity greater than or equal to 50% of that observed in virus control wells. The nSheep serum had a titer of <6.3 when assayed with minimal dilution. The titer reported in this chart is higher due to the starting dilution used in assays that was identical to the dilution used for the hyperimmune anti-HA antisera.