Author Correction: BNT162b2 vaccine induces neutralizing antibodies and poly-specific T cells in humans

In this article, we reported that some participants exhibited pre-existing CD4⁺ and CD8⁺ T-cell responses against the C-terminal part of the SARS-CoV-2 spike protein. However, we have discovered that the apparent pre-existing responses were actually responses against a contamination of two of the four lots of the spike protein sub-pool 2 (S pool 2, covering the C-terminal part of the spike protein) originating from the manufacturer, which were used in several assays. We identified the contaminating peptide in these two lots and have corrected some figures to remove the affected S pool 2 data. Removing these data affects only those statements relating to pre-existing T-cell responses, which we have now also removed from the manuscript. The contamination did not directly affect S pool 1 data. The overall results and conclusions remain the same.

Since baseline responses to S pool 2 were measured using contaminated lots, we assumed that positive responses were due to the contaminating peptide. Therefore, we have now reanalyzed the data and excluded all data points for participants meeting the following two criteria: (1) their samples were analyzed with one of the contaminated lots of S Pool 2 and (2) their baseline response was positive. As only a fraction of S pool 2 data was actually removed, the overall results and conclusions remain the same.

In Fig. 2 data describing the magnitude of T-cell responses, 9 of 37 participants were removed from the S pool 2 analysis in panels 2a and 2b of the revised figures. For panel 2a, this altered the number of vaccine responders shown for each dose group, but the noted poly-epitopic CD4⁺ T-cell responses are still apparent. Panel 2b shows the magnitude of responses with combined S pool 1 and S pool 2 data and the data points (11 for CD4⁺ T cells and 15 for CD8⁺ T cells) corresponding to the S pool 2 data for the 9 participants were removed. Two of 6 participants were removed from the CD8⁺ T-cell spot counts from the 30 µg data in panel 2c. The response numbers in the revised Fig. 2c remain unchanged as these are combined S pool 1 and S pool 2 responses.

Within the cytokine polarization data in Fig. 3, the lower portion of panel 3a corresponding to S pool 2 data was revised as we excluded 1 participant from the 20 µg dose cohort due to non-specific responses to the S pool 2 peptide lot in this experiment. Remaining data and conclusions in the revised figure remain unchanged.

In the revised Extended Data Fig. 5, data from 9 of 37 participants were also removed from panel 5b as this corresponds to the data in Fig. 2. The response values remain unchanged in the revised figure as they take into account the combined S pool 1 and S pool 2 responses.

We updated the pie chart in Extended Data Fig. 5d to remove 2 data points from the “S pool 1 & 2” category and 1 from the “S pool 1 only” category for CD4⁺ T cells. For the CD8⁺ T-cell chart, we removed 6 data points from the “S pool 1 only” category and 1 data point from “None”. Overall, samples from “S pool 1 only” were removed because without reliable S pool 2 data for these participants, we could not conclusively assign responses to “S pool 1 only” versus “S pool 1 & 2”.

We fully removed panel 5e from Extended Data Fig. 5 as it was based on what were previously considered pre-existing responses to S pool 2. Since we now know these are non-specific responses due to the contamination, the panel has been removed.

For Extended Data Fig. 5, panel 5f (now 5e), we previously chose ELISpot examples that were from participants with a non-specific response to S pool 2. We have updated the revised figure to show a single example from a participant that had no such response.

Three panels within Extended Data Fig. 7 were revised: 7b, 7d and 7e. In panel 7b exploring the CD4⁺ T-cell profiles, we removed 1 participant in the 20 µg dose group due to the non-specific response to S pool 2. In panel 7d, we removed data for 10 of 41 participants that showed non-specific CD8⁺ T-cell responses to S pool 2. The values above the data points were also updated as these correspond to the mean fractions per dose cohort and were thus affected by the data removal. In the revised panel 7e, we removed data from 3 of 8 participants in the CD8⁺ responses against S pool 2. The overall response kinetics shown in this figure remain unchanged.

The figures have been updated. Text in Figs. 2, 3 and Extended Data Figs. 5, 7 legends were updated, and the “Vaccine-induced T cell responses” and “Polarization of T cell responses” sections have been updated, while the fourth paragraph of the original Discussion (and citations to original refs. 35,36) has been removed. For comparison, the original figure panels are available as Supplementary Information alongside the online version of this amendment. The changes have been made to the HTML and PDF versions of the article.