Science - USA (2019-08-30)

30 AUGUST 2019 • VOL 365 ISSUE 6456 863

GRAPHIC: KELLIE HOLOSKI/

SCIENCE

can differ from pregnancy to pregnancy,

even in the same couple. All HLA-C allotypes

can bind variable dNK receptors, called killer

cell immunoglobulin-like receptors (KIRs),

which can be activating or inhibitory. Com-

binations of maternal KIR and fetal HLA-C

genetic variants that activate dNK cells may

attract EVT cells and enhance placental per-

fusion, thus improving placental invasion

and promoting fetal growth. Conversely,

combinations that inhibit dNK cells may

impede trophoblast invasion and contribute

to preeclampsia and FGR ( 6 ). Similar mecha-

nisms also seem to occur in mice ( 7 ).

Other receptor-ligand pairs between dNK

cells and trophoblasts may also be important

to human placentation, such as the inhibitory

receptors NK group 2A (NKG2A) and leuko-

cyte immunoglobulin-like

receptor B1 (LILRB1) (also

expressed on macrophages),

which bind to HLA-E and

HLA-G, respectively. It has

become apparent that, like

adaptive immune responses,

innate immunity also has

some form of memory. Dur-

ing first pregnancies, LILRB1

and the activating receptor

NKG2C (which also binds

HLA-E) may trigger epigen-

etic DNA modifications in

some dNK cells, potentially

making them more active

in subsequent pregnancies,

which are less prone to pre-

eclampsia provided the male

partner does not change ( 8 ).

Therefore, some form of im-

munological memory may

facilitate placental invasion into the decidua.

The cellular networks operating in the

decidua and placenta are starting to be re-

vealed by technological advances. RNA se-

quencing in mouse dNK cells and related

innate lymphoid cells (ILCs) has revealed

further specialization of these cells in the

decidua, where subpopulations of ILCs may

play distinct roles in trophoblast invasion

and vascular remodeling, immune defenses,

and immunological memory ( 9 ). Single-cell

RNA sequencing of the human maternal-

fetal interface has begun to identify cellu-

lar and molecular networks ( 10 ). Decidual

stromal cells produce interleukin-15 (IL-15),

the key dNK cell growth factor. Fetal EVT

cells produce transforming growth factor

b (TGFb), which promotes differentiation

of maternal Tregs. Three subpopulations of

dNK cells interact with both maternal and

fetal cells using different receptor-ligand

combinations ( 10 ). Analysis of both cell

suspensions and tissue sections by mass cy-

tometry should provide new functional and

spatial information. Organoids (three-di-

mensional cell cultures that form structures

morphologically similar to tissues in vivo)

( 11 ) of human trophoblasts cocultured with

maternal immune cells with defined genetic

backgrounds could also help to study these

complex interactions.

Maternal and fetal cells in the decidua

and placenta do not operate in isolation,

and metabolism has a considerable impact

on immune cell function. How do nutrition

and oxygen concentration influence these

interactions? Metabolomics analyses of the

human placenta have started to decipher the

differences between metabolites in maternal

and fetal tissues ( 12 ) and might eventually

integrate our growing appreciation of how

metabolism shapes immune functions. Do

microorganisms influence the immunology

of placentation, thereby contributing to preg-

nancy complications? Labor may be precipi-

tated by intrauterine infections, leading to

preterm birth. The human genome harbors

DNA from retroviruses that have helped tro-

phoblasts to fuse together to form the syn-

cytiotrophoblast, which separates the fetus

from maternal blood. Maternal microbes or

their products may influence fetal-placental

development, and some evidence suggests

that microbiota in the placenta and amniotic

fluid may colonize the fetal gut in utero ( 13 ).

However, other findings do not support the

existence of a placental microbiota ( 14 ), and

further research is needed.

Once cellular and molecular networks

that regulate placentation are identified, the

challenge will be to understand how these

are altered in pregnancy complications. Are

we ready to manipulate the human immune

system during pregnancy to prevent com-

plications? Despite the precedent of immu-

notherapy to prevent Rh disease, it remains

too risky to interfere with the maternal im-

mune system. Currently, anti-inflammatory

and immunosuppressive drugs such as as-

pirin, hydroxychloroquine, and steroids are

used to improve the outcome of preeclamp-

sia, antiphospholipid syndrome (in which

antibodies cause thrombosis and, in turn,

potentially miscarriage), and chronic in-

flammation of the placental villi, or even to

improve fertility. The benefit of these treat-

ments is generally minimal and probably

not specific. For example, aspirin may help

treat preeclampsia because it reduces sys-

temic endothelial inflammatory responses,

but its use does not correct the primary de-

fect at the maternal-fetal interface. There

is optimism about immune checkpoint in-

hibitors (ICIs) to treat patients with can-

cer. These antibodies block

inhibitory receptors and

can reawaken antitumor

immune responses to can-

cer cells, thereby improving

survival. Subsets of dNK

cells and T cells in the de-

cidua express inhibitory re-

ceptors that can be targeted

by ICIs, and trophoblasts

express ligands for some of

these receptors ( 10 ). Treat-

ing pregnant mice with ICIs

causes fetal loss. However,

three patients with mela-

noma treated with ICIs sur-

vived and had successful

pregnancy outcomes [one

case reported in ( 15 )], il-

lustrating the possibility of

immunotherapy to target

the maternal-fetal interface.

Researchers are just starting to crack the

immunological code of pregnancy and may

one day intervene to improve outcomes. j

REFERENCES AND NOTES

1. I. Brosens, R. Pijnenborg, L. Vercruysse, R. Romero, Am. J.
   Obstet. Gynecol. 204 , 193 (2011).


2. A. Erlebacher, Annu. Rev. Immunol. 31 , 387 (2013).


3. S. A. Robertson et al., Front. Immunol. 10 , 478 (2019).


4. M. Panduro, C. Benoist, D. Mathis, Annu. Rev. Immunol. 34 ,
   609 (2016).


5. A. A. Ashkar, J. P. Di Santo, B. A. Croy, J. Exp. Med. 192 , 259
   (2000).


6. A. Moffett, F. Colucci, J. Clin. Invest. 124 , 1872 (2014).


7. J. Kieckbusch, L. M. Gaynor, A. Moffett, F. Colucci, Nat.
   Commun. 5 , 3359 (2014).


8. M. Gamliel et al., Immunity 48 , 951 (2018).


9. I. Filipovic et al., Nat. Commun. 9 , 4492 (2018).


10. R. Vento-Tormo et al., Nature 563 , 347 (2018).


11. M. Y. Turco et al., Nature 564 , 263 (2018).


12. J. M. Walejko, A. Chelliah, M. Keller-Wood, A. Gregg, A.
    Edison, Metabolites 8 , 10 (2018).


13. M. C. Collado, S. Rautava, J. Aakko, E. Isolauri, S. Salminen,
    Sci. Rep. 6 , 23129 (2016).


14. M. C. de Goffau et al., Nature 572 , 329 (2019).


15. M. Burotto et al., Semin. Oncol. 45 , 164 (2018).

ACKNOWLEDGMENTS

Supported by Wellcome Trust grant 200841/Z/16/Z.

10.1126/science.aaw1300

KlR CSF1R

HLA-C

HLA-E

NKG2A

Decidual

natural

killer cell

Extravillous

trophoblast

KLRB1

TIGIT

LILRB1

CD155

Lymphotactin

CSF1

HLA-G

Macrophage Dendritic cell

XCR1

Regulatory T cell

CLEC2D, C-type lectin domain family 2 member D; CSF1, colony-stimulating factor 1; CSF1R, CSF1 receptor; HLA, human

leukocyte antigen; KIR, killer cell immunoglobulin-like receptor; KLRB1, killer cell lectin-like receptor B1; LILRB1,

leukocyte immunoglobulin-like receptor B1; NKG2A, natural killer cell receptor G2A; TIGIT, T cell immunoreceptor with Ig

and ITIM domains; XCR1, lymphotactin receptor; CD155, cluster of diferentiation 155.

Decidua

Spiral artery

Villous

tree

CLEC2D

SCIENCE sciencemag.org

Some maternal-fetal interactions in the decidua

Trophoblast cells transform uterine arteries into spiral arteries. Extravillous trophoblast

cells bud off the placental villi to invade the decidua, where they interact with maternal

immune cells through ligand-receptor pairs.

Published by AAAS