Goelet al.,Science 374 , eabm0829 (2021) 3 December 2021 9 of 17
Gated on Non-Naive CD4+ T
A
Activation
Markers
24 hours
CD4+ T
CD8+ T
Quantify
Frequency/Phenotype
Stimulate with Spike
Peptide Megapools
CD40LCD107a
CD200IFN-
4-1BB
Flow Cytometry
A
B
0.058 0.059 0.039 0.66
0.039 0.032 0.015 0.20
Gated on Non-Naive CD8+ T
Pre-Immune Post-Vaccine
Pre-Immune Post-Vaccine
CD200
CD40L
IFN-
4-1BB
Unstimulated + Peptide Pool Unstimulated + Peptide Pool
Unstimulated + Peptide Pool Unstimulated + Peptide Pool
* **
CM
EM1
EM2
EM3
EMRA
3m6m
% of Peak Response
Total CD4 Durability
(% of Peak Magnitude)
= 0.5 ,p= 0.0062
0.0
0.3
0.6
0.9
0.00 0.25 0.50 0.75
% EM1 at 1 Month
6 Month Durability
AIM+ cTfh AIM+ Th1 AIM+ Th2 AIM+ Th17 AIM+ Th1/17
Naive
Recovered
(^0) 0.50.75 (^1360) 0.50.75 (^1360) 0.50.75 (^1360) 0.50.75 (^1360) 0.50.75 1 36
0.001
0.010
0.100
1.000
0.001
0.010
0.100
1.000
Months Post−Vaccine
% of nnCD4+
CD4+ Helper Subsets
**
****ns
**
0.053
- ns
nsns
AIM+ CM AIM+ EM1 AIM+ EM2 AIM+ EM3 AIM+ EMRA
Naive
Recovered
(^0) 0.50.75 (^1360) 0.50.75 (^1360) 0.50.75 (^1360) 0.50.75 (^1360) 0.50.75 1 36
0.001
0.010
0.100
0.001
0.010
0.100
Months Post−Vaccine
% of nnCD4+
CD4+ Memory Subsets
*ns
ns
nsns
***ns
Naive Recovered
(^050100150200050100150200)
0.001
0.010
0.100
1.000
Days Post−Vaccine
% of nnCD4+
AIM+ CD4 T (CD200+ CD40L+)
0.01
0.03
0.10
0.30
0 2 4 6
Months
% of nnCD4+
CD4+ AIM+
Naive Recovered
(^050100150200050100150200)
0.001
0.010
0.100
1.000
Days Post−Vaccine
% of nnCD8+
AIM+ CD8 T (4 of 5 markers)
0.001
0.010
0.100
0 2 4 6
Months
% of nnCD8+
CD8+ AIM+
41/41 37/39 28/31
36/41 27/39 13/31
14/15 8/13 3/9
10/15 3/13 2/9
t1/2 = 47d t1/2 = 187d
SARS-CoV-2 Naive
N = 42
SARS-CoV-2 Recovered
N = 16
- mRNA
Vaccine
C
D
CXCR5
SSC-A
CXCR3
CCR6
CD4+ Helper Subsets
cTfh
Th17 Th1/17
Th2 Th1
IJ
E
FG
H
CCR7+: CM
CCR7-: EM1
CCR7+: EM2
CCR7-: EM3CCR7-: EMRA
Naive
CD45RA
CD27
CD4+ Memory Subsets
t1/2 = 27dt1/2 = NA
p = 0.33
Decay Rates = ns
Decay Rates = ns
p = 0.11
AIM+ Cells
Pfizer
Moderna
ns
ns
ns
ns
**
ns
ns
ns
nsns ns ns
nsns
ns
- ns ns nsns
ns
ns - ns
* ns
nsns
ns
** nsns
ns
0.099 ns
nsns
< 0.05
< 0.01
P value
−1
−0.5
0
0.5
1
Fig. 5. SARS-CoV-2 mRNA vaccines generate durable memory T cell
responses.(AandB) Experimental design (A) and gating strategy (B) for
quantifying the frequency of SARS-CoV-2–specific CD4+and CD8+T cells by
AIM assay. For CD4+T cells, antigen specificity was defined on the basis of
coexpression of CD40L and CD200. For CD8+T cells, antigen specificity was
defined on the basis of expression of at least four of five activation markers,
as indicated in (A). (CandD) Frequencies of AIM+CD4+T cells (C) and AIM+
CD8+T cells (D) over time in PBMC samples from vaccinated individuals.
Data were background subtracted using a paired unstimulated control for each
time point and are represented as a percentage of non-naïve CD4+or CD8+
T cells. Black triangles indicate time of vaccine doses, fractions above plots
indicate the number of individuals above their individual baseline at memory time
points, and summary plots show mean values with the 95% confidence interval.
Decay rates were calculated using a piecewise linear mixed-effects model
with censoring.DDecay Rates indicates whether decay rates were different
in SARS-CoV-2–naïve and–recovered groups. (E) AIM+CD4+T cell memory
subsets were identified on the basis of surface expression of CD45RA, CD27,
and CCR7. (F) Frequencies of AIM+CD4+T cell memory subsets over time.
(G) Correlation matrix of memory subset skewing at peak (1-month) response
with total AIM+CD4+T cell durability at 3 and 6 months. Durability was measured
as the percentage of peak response maintained at memory time points for
each individual. (H) Correlation between percent of EM1 cells at peak response
and 6-month durability. (I) AIM+CD4+T helper subsets were defined on the basis
of chemokine receptor expression. (J) Frequencies of AIM+CD4+T helper
subsets over time. For (F) and (J), lines connect mean values at different time
points. Dotted lines indicate the limit of detection for the assay. Statistics
were calculated using unpaired nonparametric Wilcoxon test with BH correction
for multiple comparisons. Correlations were calculated using nonparametric
Spearman rank correlation. *P< 0.05; *P< 0.01; P< 0.001; ****P< 0.0001;
ns, not significant.
RESEARCH | RESEARCH ARTICLE