respectively, after ~6 months (week 29). Boost-
ing with mRNA-1273 or mRNA-1273.binduced
fourfold and about sixfold increases in WA-1
andbS-specific IgG titers such that, at 2 weeks
after the memory boost, WA-1 andbS-specific
IgG GMTs were restored to peak levels (Fig. 1,
A and B). By week 37, S-specific IgG GMTs
remained >30,000 and >20,000 AUC, respec-
tively, for WA-1 andb(Fig. 1, A and B), which,
for WA-1, translated to ~1400 international
units/ml (table S1). RBD-specific responses
displayed similar kinetic trends (Fig. 1, C and
D). Together, these data show that there is
no difference in the ability of mRNA-1273 or
mRNA-1273.bvaccines to boost primary mRNA-
1273 – elicited S-binding IgG responses.
Next, we assessed the kinetics of neutral-
izing antibody responses using a lentiviral-
based neutralizing antibody assay for D614G,
the benchmark strain, as well as several glob-
ally circulating variants [b,d, P.1-g, B.1.429-
Epsilon (e), and B.1.526-Iota (i)]. Consistent
withpreviousdatainNHPsandhumans( 7 , 15 ),
mRNA-1273 elicited a variant-dependent hier-
archy of neutralizing antibody responses as
determined by a lentiviral-based pseudovirus
assay. At peak, the median inhibitory dilu-
tion (ID 50 ) GMTs for D614G was 4700, fol-
lowed by 5900, 1500, 1100, 830, and 770 for
e,g,i,d, andb, respectively, representing
onefold to sixfold decreases compared with
D614G. At the memory time point (week 24),
neutralizing antibody titers decreased to 640
ID 50 GMTs for D614G and 810, 280, 250, 740,
and 350 ID 50 GMTs fore,g,i,d, andb, re-
spectively. There was a greater fold reduc-
tion in neutralization titers from week 6 to
week 24 for D614G compared withbandd
(P< 0.0001) (Fig. 1, E and F). Similar obser-
vations were made using D614G- andbVSV–
based pseudovirus (fig. S2, A and B) or live
virus neutralization assays (fig. S2, C and
D). Last, WA-1 S-specific antibody avidity,
one measure of affinity maturation, was sig-
nificantly increased over the 6 months after
the primary mRNA-1273 vaccination series
in these NHPs (P< 0.0001) (Fig. 1, I and J).
This indicates that continued affinity mat-
uration occurs after primary vaccination,
leading to an increase in antibody quality,
and suggests that, for some variants, affin-
ity maturation of mRNA-1273–elicited neu-
tralizing antibody responses may occur over
time, even in the absence of continued anti-
gen exposure.
Six months after the primary vaccination
series, either a homologous (mRNA-1273) or
a heterologous (mRNA-1273.b) boost induced,
on average, a 12-fold increase in pseudovirus-
neutralizing titers for all the variants 2 weeks
later, resulting in, e.g., 5000 and 3000b-specific
ID 50 GMTs after mRNA-1273 or mRNA-1273.b
boosts, respectively (Fig. 1, E and F). These
responses to variants were significantly higher1344 10 DECEMBER 2021•VOL 374 ISSUE 6573 science.orgSCIENCE
ABCDEFGHIJFig. 1. Temporal serum antibody responses to SARS-CoV-2 variants.Rhesus macaques (n= 8 per group)
were immunized as described in fig. S1. (AtoD) Sera collected at weeks 6, 29, 31, and 37 were assessed
for SARS-CoV-2 WA-1 (black) andb(red) S-specific [(A) and (B)] and RBD-specific [(C) and (D)] IgG by
multiarray ELISA. (EandF) Sera collected at weeks 6, 24, and 31 were assessed for D614G (black),b, P.1-g
(green), B.1.429-e(purple), B.1.526-i(orange), and B.1.617.2-d(blue) lentiviral-based pseudovirus
neutralization. (GandH) Sera collected at weeks 12 and 37 were assessed for D614G,b, anddlentiviral-
based pseudovirus neutralization. (IandJ) Sera collected at weeks 6, 24, and 31 were assessed for WA-1
S-specific antibody avidity. For (A) to (F) and (I) and (J), lines represent GMTs and geometric error.
Arrows point to immunization time points. For (G) and (H), symbols represent individual NHPs. Dotted lines
indicate neutralization assay limits of detection. Significance was measured by pairedttest.
RESEARCH | RESEARCH ARTICLES