Science - USA - 03.12.2021

vaccine compared with a group of SARS-
CoV-2–recovered vaccinees generating recall
responses from preexisting immunity. These
analyses provide insights into mRNA vaccine–
induced immunological memory and may be
relevant for future vaccine strategies, includ-
ing recommendations for additional booster
vaccine doses.

Results and discussion
Cohort design

We collected 348 longitudinal samples from
61 individuals receiving either the Pfizer
BNT162b2 (N= 54) or Moderna mRNA-1273
(N= 7) SARS-CoV-2 vaccines at six time points
(Fig. 1A), ranging from prevaccination base-
line to 6 months postvaccination. This study
design allowed us to monitor the induction

and maintenance of antigen-specific immune

responses to the vaccine. Specifically, sampling

at 1, 3, and 6 months postvaccination enabled

the analysis of immune trajectories from peak

responses after the second vaccine dose through

the establishment and maintenance of immu-

nological memory. This cohort was divided into

two groups on the basis of prior SARS-CoV-2

infection (N= 45 SARS-CoV-2–naïve individu-

als;N= 16 SARS-CoV-2–recovered individuals).

Ageandsexwerebalancedinbothgroups.

Paired serum and peripheral blood mono-

nuclear cell (PBMC) samples were collected

from all individuals, allowing for a detailed

analysis of both serologic and cellular immune

memory to SARS-CoV-2 antigens. Notably, the

subjects with a prior infection allowed us to

study the dynamics of reactivating preexisting

immunity with mRNA vaccines. Though pre-

existing immunity generated by infection may

differ from that generated by vaccination, re-

sponses observed in this group may provide

insights into the boosting of vaccine-induced

immunity using additional doses of vaccine.

Antibody responses to SARS-CoV-2

mRNA vaccines

We first measured anti-spike and anti-RBD

binding antibody responses in plasma sam-

ples by enzyme-linked immunosorbent assay

(ELISA). As reported previously by our group

and others, mRNA vaccines induced robust

circulating antibody responses to the SARS-

CoV-2 spike protein and spike RBD with distinct

patterns of early response in SARS-CoV-2–naïve

and–recovered individuals (Fig. 1B) ( 16 , 34 – 36 ).

Goelet al.,Science 374 , eabm0829 (2021) 3 December 2021 2of17

(^131030131030131030)
10
30
100
300
anti−Spike IgG
FRNT50
(^131013101310)
10
30
100
300
anti−RBD IgG
FRNT50
p = 4.3e-08
p = 3.9e-07
p = 1.7e-04
p = 1.8e-04
= 0.8 = 0.62 = 0.81
p = 2.4e-08
= 0.76 = 0.62 = 0.83
p = 1e-08
E
Naive Recovered
D614GB.1.351B.1.617
.2
D614
G
B.^1 .351B.1.^6 17.2
10
100
1000
10000
FRNT50
6 Months

0.052

ns
10
100
1000
10000
(^050100150200)
Days Post−Vaccine
FRNT50
D614G Neutralization
10
100
1000
10000
0 2 4 6
Months
D614G
10
100
1000
10000
(^050100150200)
Days Post−Vaccine
FRNT50
B.1.351 Neutralization
10
100
1000
10000
0 2 4 6
Months
B.1.351
44/4516/16 33/339/9
43/45
16/16
31/33
9/9
t1/2 = 51d t1/2 = 72d
t1/2 = 45d
t1/2 = 49d
t1/2 = 63d
t1/2 = 231d
p = 0.17
Decay Rates = ns
Decay Rates = signif
p = 6.2x10-5
SARS-CoV-2 Naive
N = 45
SARS-CoV-2 Recovered
N = 16

• mRNA
  Vaccine
  Age: 36.9 [22-67]
  Sex: 21M 24F
  Age: 38.3 [23-59]
  Sex: 10M 6F
  A
  1
  10
  100
  1000
  (^050100150200)
  Days Post−Vaccine
  ug/mL
  anti−Spike IgG
  1
  10
  100
  1000
  0 2 4 6
  Months
  anti−Spike IgG
  1
  10
  100
  1000
  (^050100150200)
  Days Post−Vaccine
  ug/mL
  anti−RBD IgG
  1
  10
  100
  1000
  0 2 4 6
  Months
  anti−RBD IgG
  B
  C
  Dose 1 +2wks Dose 2
  T1 T2 T3 T4
  +1wk
  T5 T6
  3 months 6 months
  Memory
  Cellular Memory
  D
  4 8 4 8
  Serology
  45/45
  16/16
  32/32
  9/9
  45/45
  16/16
  31/32
  9/9
  p = 0.11
  Decay Rates = ns
  Decay Rates = signif
  p = 3.5x10-4
  BB
  Mean: 136 80 117 15954331372
  D614G B.1.351 B.1.617.2
  Pfizer
  Moderna
  Variant:

  ---

  
  

  ---

  
  

  ---

  
  

  ---

  
  

  t1/2 = 47d
  t1/2 = 28d
  t1/2 = 47d
  t1/2 = 42d
  t1/2 = 33d t1/2 = 48d
  Fig. 1. SARS-CoV-2 mRNA vaccines induce robust antibody responses.
  (A) University of Pennsylvania COVID-19 vaccine study design and cohort
  summary statistics. (B) Anti-spike and anti-RBD IgG concentrations over time in
  plasma samples from vaccinated individuals. (C) Pseudovirus neutralization
  titers against WT D614G or B.1.351 variant spike protein over time in plasma
  samples from vaccinated individuals. Data are represented as focus reduction
  neutralization titer 50% (FRNT 50 ) values. (D) Comparison of D614G, B.1.351, and
  B.1.617.2 FRNT 50 values at 6 months postvaccination. (E) Correlation between
  anti-spike or anti-RBD IgG and neutralizing titers (D614G = black, B.1.351 = green,
  and B.1.617.2 = orange; statistics were calculated using nonparametric Spearman
  rank correlation). Dotted lines indicate the limit of detection for the assay.
  For (B) and (C), 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. Changes in decay rate over time (linear versus
  two-phase decay) were determined on the basis of a likelihood ratio test.DDecay
  Rates indicates whether decay rates were different in SARS-CoV-2–naïve
  and–recovered groups. Statistics were calculated using unpaired [(B) and (C)]
  or paired (D) nonparametric Wilcoxon test with Benjamini-Hochberg (BH)
  correction. Blue and red values indicate comparisons within naïve or recovered
  groups. P< 0.05; P< 0.01; P< 0.001; ****P< 0.0001; ns, not significant.
  RESEARCH | RESEARCH ARTICLE