Science - USA (2022-05-27)

reported in different studies. This is likely be-
cause of challenges in separating low-biomass
bacteria in fetal intestines and amniotic fluid
from environmental contaminants during tis-
sue preparation ( 12 , 13 ). Multiple studies that
have controlled for environmental contami-
nants have failed to find evidence of fetal bac-
terial colonization in the absence of infection
( 14 , 15 ). Conversely, in fetuses in which low-
abundance bacteria have been reported, tran-
scriptional adaptation in epithelial cells ( 13 )
and the presence of memory-phenotype T cells
have been interpretated as evidence in favor of
prenatal microbial colonization ( 16 , 17 ). How-
ever, microbial components and metabolites
transferred directly across the placenta, or in-

directly through maternal IgGs that cross the
placenta, are sufficient to prime fetal intestinal
epithelial cells in mice ( 11 )andcanprobably
also induce memory T cell differentiation.
Fetal T cells are hyperresponsive to foreign and
endogenous antigens ( 5 , 18 , 19 ). Such T cell
hyperresponsiveness itself could be considered
evidence for antigen scarcity in utero. This
is because T cells in general require sub-
threshold stimulation through their T cell
receptors and tune their thresholds for acti-
vation accordingly ( 20 ). An environment devoid
of triggering antigens would be expected to
induce lower thresholds for activation and
T cell hyperresponsiveness, as seen in fetuses.
Similar models have been put forward for
other immune cells, such as natural killer
cells ( 21 ).
An additional role of transplacental trans-
fer of microbial components could be to pro-
vide antigens to fetal T cells such that their
thresholds of activation can be tuned down in
preparation for postnatalmicrobialcoloniza-

tion to avoid immunopathology and permit

microbial colonization at birth and thereafter.

Trained immunity is an additional mechanism

by which fetal innate immune cells can be

prepared for postnatal exposures by epigenetic

adaptation ( 21 ). For example, monocytes of

HIV-exposed but uninfected fetuses adapt

and respond more vigorously to Toll-like re-

ceptor ligands after birth compared with HIV-

unexposed controls ( 22 ). Similar adaptive

changes have been reported in children born

to mothers infected by malaria during preg-

nancy ( 23 ). Through these multiple mecha-

nisms, fetal immune systems influenced by

maternal exposures to be prepared for the

challengingbalancingactofpermittingmicro-

bial colonization yet preventing dangerous in-

fections after birth.

Birth and first encounters

Labor is initiated by hormonal factors that are

triggered by physical pressure from the fetus

onto the cervix, but also by inflammatory sig-

nals and the breakdown of feto-maternal tol-

erance. Inflammatory stimuli such as infections

can trigger such loss of tolerance and initiate

preterm labor ( 24 ), and specific strains of

vaginal bacteria have been implicated ( 25 ).

Moreover, fetal cells triggered against mater-

nal antigens can secrete cytokines that lead to

uterine contractions, illustrating the bidirec-

tional nature of feto-maternal tolerance and its

role in preterm labor ( 26 ). Once membranes

rupture, a baby is exposed to microbes through

the birth canal. In vaginal delivery, fecal and

vaginal microbes from the mother seed the

newbornskinandmouth,whereasbabies

born through Cesarean section are populated

mainly by skin microbes, leading to substan-

tial differences in microbial composition early

after birth ( 27 ). Infants continue to receive

microbes from their parents and other sources,

but microbes seeded from the mother during

vaginal labor and originating in the intestine

are optimally adapted to survive in the infant

gut and will predominate over time, illustrating

the niche selection that occurs after birth ( 28 ).

Breastfeeding promotes specific colonizing

microbes in the newborn intestine, including

Bifidobacteria, which are adapted to metabo-

lize HMOs. Antibiotic exposures reduce over-

all microbial diversity and delay the normal

development of the intestinal microbiome ( 29 ),

and such early-life antibiotic exposure is as-

sociated with increased risks of developing

allergy ( 30 ), asthma ( 31 ), type 1 diabetes ( 32 ),

and obesity ( 33 ) later in childhood. Mecha-

nisms underlying thesealtered developmental

trajectories are discussed below.

Immune reactions to microbiota

In the first postnatal days, a strong systemic

immune response occurs in infants, which is

dominated by innate immune cells such as

monocytes expanding in the blood and the ele-

vation of circulating proinflammatory cytokines,

but also by endogenous regulatory factors such

as IL-1 receptor antagonist ( 34 ). Preterm babies

accumulate many disease-associated environ-

mental exposures, such as antibiotics, infections,

and cesarean delivery. They also have elevated

inflammatory responses in cord blood compared

with term children, likely related to prenatal

conditions triggering preterm delivery in the

first place. After birth, preterm and term in-

fants undergo similar changes in response to

postnatal exposures that are similar to all

newborns, leading to a marked convergence

of immune cell and protein profiles during

the first weeks of life in both preterm and term

infants ( 35 ). Despite these similar adaptive

changes during the first weeks of life, very

preterm babies (<31 weeks of gestation) remain

at much elevated risk (odds ratio ~3.6 versus

term infants) of asthma ( 36 ) and other immune-

mediated diseases.

After birth, cells of the immune system,

particularly those at barrier surfaces, must

recognize colonizing microbes and determine

whethertotolerateorresistthem.Theestab-

lishment of host-microbe mutualism is critical

for health, and several processes work to pro-

mote this. Children lacking Tregs (e.g., those

with immunodysregulation polyendocrinopathy

enteropathy X-linked syndrome) typically pre-

sent early in life with severe inflammation

involving both the intestines and skin after

microbial colonization ( 37 ). In mice, a distinct

population of FoxP3+Tregs are induced in the

thymus during the perinatal period that are

essential for maintaining self-tolerance and

preventing autoimmunity later in life ( 38 ). Ad-

ditional experiments in mice have shown that

Brodin,Science 376 , 945–950 (2022) 27 May 2022 2of

A Fetal immune system and layered hematopoiesis B Transplacental transfer of IgG and

microbial components

16 20 36 Birth Adulthood

Gestational week

Maternal IgG

carrying microbial

components

Passive immunity

Priming of fetal

epithelial cells for

postnatal colonization

Tuning/training of fetal

immune cells?

Feto-maternal

tolerance

Fe t a l

hematopoiesis

Resistance

to pathogens

Adult

hematopoiesis

Variable transition period

IFN-1–triggered

switch (infection?)

Fig. 1. Prenatal preparation.(A) Layered hematopoiesis occurs with different stem cell populations that
give rise to cells poised toward feto-maternal tolerance early during pregnancy but gradually switch over
toward pathogen resistance at the end of gestation in preparation for microbial colonization. The transition
period is gradual in nature ( 7 ), and one trigger for this switch from fetal to adult hematopoiesis is a surge
in IFN-I before birth ( 8 ). ( B) Transplacental transfer of maternal IgG, and possibly microbial antigens,
prepares fetal immune systems and epithelia for microbial colonization at birth and thereafter.

ILLUSTRATION: KELLIE HOLOSKI/

SCIENCE