Determinants of S.

aureus carriage in the developing

infant nasal microbiome
Emma Accorsi1, Eric A. Franzosa1,2, Tiffany Hsu1,2, Regina J. Cordy3, Ayala Maayan-Metzger4,5, Hanaa Jaber5,
Aylana Reiss-Mandel5, Casey DuLong1, Marc Lipsitch1, Gili Regev-Yochay4,5, Curtis Huttenhower1,2
1
Harvard T. H. Chan School of Public Health; 2Broad Institute; 3Wake Forest University; 4Sackler School of Medicine; 5Sheba Medical Center

Staphylococcus aureus is a leading cause of healthcare- and community-asso- Microbiome drivers of infant S. Infant nasal microbiomes
aureus phenotypes mature over the first year, but
ciated infections and can be difficult to treat due to antimicrobial resistance.
About 30% of individuals carry S. aureus asymptomatically in their nares, a risk
factor for later infection, and interactions with other species in the nasal micro-
remain distinct from mothers
Infant Metadata Infant Metadata
biome likely modulate its carriage. It is thus important to identify ecological or Infant Species Infant Functions
functional genetic elements within the maternal or infant nasal microbiomes Acinetobacter unclassified
that influence S. aureus acquisition and retention in early life. We recruited 36 Infants had a weak upward trend in alpha diversity over time.
Haemophilus influenzae
mother-infant pairs and profiled a subset of monthly longitudinal nasal samples
They rapidly diverged from their species composition at birth,

108 Functions (ECs)

from the first year after birth (n=284) using shotgun metagenomic sequencing. Moraxella catarrhalis

The infant nasal microbiome was highly variable, particularly within the first 1-2 Staphylococcus aureus but the rate of change slowed over time indicating stabilization
months. It was weakly influenced by maternal nasal microbiome composition, Staphylococcus epidermidis toward a more mature microbiome. Infants were more similar to
but primarily shaped by developmental and external factors (e.g. daycare). In- their own mother than to unrelated mothers at month 1 (PER-
Streptococcus parasanguinis
fants displayed distinctive patterns of S. aureus carriage, positively associated
with Acinetobacter species, Streptococcus parasanguinis, Streptococcus sali- Streptococcus salivarius MANOVA, p=0.005), although infant composition was distinct
varius, and Veillonella species and inversely associated with maternal Dolosig- Veillonella unclassified from maternal composition at all months except 8 (p<0.05).
ranulum pigrum. Furthermore, we identified a gene family, likely acting as a tax- Infant 1) Self−dissimilarity: birth 2) Self−dissimilarity: last timepoint

Alpha−diversity (Gini−Simpson index)

onomic marker for an unclassified species, that was significantly anticorrelated Infant Metadata Infant Metadata 1.00

Infant beta−diversity (Bray−Curtis dissimilarity)

Mother
Mother Species Mother Functions
with S. aureus in infants and mothers. In gene-content based strain profiling, 0.8
0.75
Cutibacterium acnes
infant S. aureus strains were more similar to maternal strains. This improved

84 Functions (ECs)
0.50
understanding of S. aureus colonization is an important first step toward devel- Dolosigranulum pigrum 0.6

opment of novel, ecological therapies for controlling S. aureus carriage. Micrococcus luteus 0.4 0.25

Propionibacterium unclassified 0.2

0.00
Staphylococcus epidermidis
0.0
3) Mothers 4) Other infants by S. aureus status
Streptococcus gordonii 0 1 2 3 4 5 6 7 8 9 101112 M 1.00

Infants display striking patterns

5.0+ Time (months)
0.75

SA gain

SA culture

SA cult or seq
Breastfeeding

Daycare

Time

SA early acquisition

SA ever acquisition

SA persistent carriage

SA sequencing

Daycare

Time

SA gain

% SA positive time points

SA culture

SA cult or seq
SA early acquisition

SA ever acquisition

SA persistent carriage

SA sequencing
Comparison of infants with:
2.5

of S. aureus carriage
Self at birth 0.50
-sign(β)*
log10(qval) Self at most recent
0.0 preceding timepoint 0.25
Related mother
Metadata Variables 0.00
−2.5 Unrelated mothers

(A) 36 mother-infant pairs gave (A) Infants same SA status 0 1* 2 3 4 5 6 7 8 9 10 11 12 0 1 2 3 4 5 6 7 8 9 10 11 12

Infant Acinetobacter unclas., S. parasanguinis, S. salivarius, and Time (months)
nasal swabs monthly over the Infants different SA status

Veillonella unclas. and maternal D. pigrum were positively asso-

first year after birth. Culture test-
ciated with infant S. aureus carriage.
ing for S. aureus was performed
on all samples and a subset Infant species Mother species Infant ECs Mother ECs
Strain genotypes show similar-
(n=284) were profiled with shot-
gun metagenomic sequencing.
Acinetobacter unclass. ity in mother & infant S. aureus
Early

Corynebacterium accolens
Infant S. aureus strains were more similar to those of their own
Corynebacterium propinquum
mothers, compared to unrelated mothers or other infants.
(B) The percent of positive time C. pseudodiphtheriticum
Overall diversity Infants & mothers: pair Infants: subject
points after S. aureus acquisi- Cutibacterium acnes
C. accolens
C. propinquum
**
Distance metric:
Mother−infant pairs

tion was not significantly differ- Dolosigranulum pigrum

C. pseudodiph. *
Overall diversity
D. pigrum
ent between early and late ac-
Haemophilus influenza
M. catarrhalis Intra-pair
Inter-pair
quirers, likely due to the small P. acnes
Moraxella catarrhalis
S. aureus * Intra-infant
sample size (n=28).
Propionibacterium unclass.
S. epidermidis Inter-infant

Species
Lorem ipsum
Staphylococcus aureus Intra-pair/inter-infant Infants: SA ever SA+ infants: SA early Intra-phenotype
C. accolens (SA ever/never)
Staphylococcus epidermidis
C. propinquum Inter-phenotype
(C) Identification of S. aureus by
Late

S. mitis/oralis/pneumoniae
C. pseudodiph. (SA ever/never)
D. pigrum * Intra-phenotype
culture and sequencing showed Streptococcus parasanguinis M. catarrhalis *** (SA early/late)
P. acnes *** Inter-phenotype
strong, although not complete, Veillonella unclass. S. aureus
(SA early/late)
S. epidermidis *
concordance. 0 0.25 0.5 0.75 1 0 0.25 0.5 0.75 1 0 0.25 0.5 0.75 1 0 0.25 0.5 0.75 1
Response variable (infant)

0 0.25 0.5 0.75 1 0 0.25 0.5 0.75 1 0 0.25 0.5 0.75 1

Infant species Mother species Infant ECs Mother ECs
Jaccard distance
S. aureus gain

Acknowledgements
Never

S. aureus loss

(B) (C) S. aureus culture

S. aureus rel. abundance

1.00 1.00
The authors express their gratitude to all partici-
% pos time points post-

S. aureus cult+seq

0.75 0.75
0 1 2 3 4 5 6 7 8 9 10 11 12
0 0.25 0.5 0.75 1 0 0.25 0.5 0.75 1 0 0.25 0.5 0.75 1 0 0.25 0.5 0.75 1 pants for generously providing their time and sam-
acquisition

Months since birth

Infant species Mother species Infant ECs Mother ECs ples, and to the NIAID of the NIH (grants
R21AI112991 to CH and T32AI007535 to EA), the
0.50 0.50
Infant culture Shotgun S. aureus early acquistion
Negative metagenomic
0.25 0.25 Positive
sequencing
performed
S. aureus ever acquistion
Chief Scientist, Ministry of Health, Israel (grant
Missing
Infant only
0 0.25 0.5 0.75 1 0 0.25 0.5 0.75 1 0 0.25 0.5 0.75 1 0 0.25 0.5 0.75 1 3-00000-5622 to GRY), and the Israel Science
AUC (95% CI)
Foundation (grants 1590/09, and 1658/15 to GRY)
0.00 0.00 Mother culture Mother only
Early (20) Late (8) Pos (91) Neg (117) Negative Both Many species in the infant microbiome were predictable (11/13
S. aureus acquisition S. aureus culture test
Positive
using species, 9/13 using ECs), but S. aureus was consistently for funding.
http://huttenhower.sph.harvard.edu
Missing
Negative (117)
difficult to predict, although performance improved using ECs.