Nature - USA (2020-05-14)

Nature | Vol 581 | 14 May 2020 | 205

The splenic nerve releases norepinephrine. In response to adrenergic
stimulation, non-neural elements such as choline acetyl transferase
(ChAT)-expressing T cells can synthesize and release acetylcholine
in the spleen^1 ,^2. By intravenous infusion of norepinephrine or ace-
tylcholine during immunization, and by transduction of B cells with
membrane-anchored acetylcholinesterase, we found that developing
SPPCs were likely to be exposed to acetylcholine and subjected to its
direct influence in vivo (Supplementary Text 1 and Extended Data Fig. 2).
We therefore further tested the expression of acetylcholine recep-
tors by B cells. There are two classes of acetylcholine receptor (AchR):
G-protein-coupled muscarinic receptors and pentameric, nicotinic
ligand-gated channels. B-lineage cells did not express substantial
amounts of muscarinic AchRs or nicotinic AchRs containing the α7
subunit, which is highly expressed by myeloid cells^3 ,^4 ; however, they
abundantly expressed nicotinic AchRs containing the α9, β1, and β4 sub-
units (Fig. 2a). Next, we used the CRISPR–Cas9 gene editing technique
to generate mice lacking the Chrna9, Chrnb1 and Chrnb4 genes on the B6
background. No homozygous germline Chrnb1 knockout was obtained,
potentially indicating embryonic lethality (data not shown). However,
both Chrna9−/− and Chrnb4−/− mice were viable and backcrossed with
wild-type B6 mice for three generations (Extended Data Fig. 3a–d).
Upon interbreeding, Chrnb4−/− mice had only infrequent litters, which
contained fewer pups than normal (data not shown). Chrna9−/− mice
were fertile, born with the expected Mendelian frequency, and con-
tained normal B cell and T cell compartments (Extended Data Fig. 3e, f
and data not shown). Using a synthetic analogue of acetylcholine
(Extended Data Fig. 3g), we measured the acetylcholine-binding capac-
ity of B-lineage cells from wild-type and Chrna9−/− mice. SPPCs exhibited
a higher capacity to bind acetylcholine than GC or total B cells, and an
α9 deficiency led to a substantial reduction in the acetylcholine-binding
capacity of SPPCs (Extended Data Fig. 3h, i).
Bone marrow cells from Chrna9−/− and μMT (B cell-deficient) mice
were mixed at a 20:80 ratio and transplanted into radiation-treated
wild-type mice to create chimaeras in which all B cells were deficient
in the α9 nicotinic AchR subunit, whereas other haematopoietic cells
were largely intact. Two weeks after immunization of these chimaeric
mice with NP-KLH, Chrna9−/− B cells exhibited a significant defect in

generating SPPCs as compared to wild-type B cells, whereas the mag-

nitude of the GC response was not altered (Fig. 2b, c). In addition, we

tested μMT mice directly infused with Chrna9−/− B cells in three experi-

ments and with Chrnb4−/− B cells in two experiments; in this setup, B

cells deficient in either α9 or β4 AchRs produced fewer SPPCs after

immunization with NP-KLH than did wild-type control cells (Extended

Data Fig. 4a, b). Therefore, acetylcholine promotes SPPC formation

via nicotinic AchRs that contain the α9 subunit and probably the β4

subunit. To investigate whether this acetylcholine effect depends on

splenic nerve activity, we conducted splenic denervation on μMT mice

before reconstitution with wild-type or Chrna9−/− B cells (Extended

Data Fig. 4c). Whereas the loss of α9 AchRs on B cells led to a reduc-

tion in SPPCs in sham-operated mice, this reduction was abrogated in

mice without intact splenic nerves (Fig. 2d, e). Together, these data

indicate that splenic nerve activity promotes SPPC formation through

an acetylcholine-induced process in B cells.

Activated CD4 T cells can express ChAT and secrete acetylcho-

line in response to norepinephrine^1 , and ChAT-expressing T cells

may translate noradrenergic signalling by the splenic nerve into

acetylcholine-dependent promotion of SPPC formation. We

reconstituted T cell-deficient mice with CD4 T cells isolated from

ChAT-IRES-Cre:Rosa26Ai14 reporter mice (Extended Data Fig. 5a). No

Ai14+ T cells were detected in the spleen before immunization with

NP-KLH, but 8–10 days after NP-KLH immunization the frequency of

Ai14+ CD4 T cells was about 0.5%. These Ai14+ cells were of a CD44hi

activated phenotype and expressed Chat mRNA (Extended Data

Fig. 5b). On tissue sections, many of these cells were in regions rich

in tyrosine hydroxylase (TH)-positive nerve fibres and aggregates of

plasma cells (Extended Data Fig. 5c). Next, we used CD4 T cells from

ChAT-IRES-Cre:Rosa26DTR or control Rosa26DTR mice to reconstitute

T cell-deficient mice. One week after immunization with NP-KLH, these

mice were given intraperitoneal injections of 50 μg kg−1 diphtheria

abSinglets

FSC-A

FSC-H

Livecells

7AAD

CD138

B lineage

CD19

CD138

IgDTHCD3

ShamDenervated

0.8

1.0

0.6

0.7

2.5 2.5

2.7 2.6

GL7

FA

S

B220

CD138

cdShamDenervated

Day 7

Day 13

P = 0.7

GC (%)

0

2.5

5.0

7.5

Day 7Day 13

SDSD

SPPC (%)

0

0.5

1.0

1.5

Day 7Day 13

SDSD

P = 0.9 P = 0.3P = 0.001

Fig. 1 | Splenic denervation reduces the abundance of plasma cells after
immunization with T-dependent antigen. a, Images of sham-operated (top)
or denervated (bottom) spleens 6 weeks after surgery, representative of three
experiments. Green, TH staining (nerve fibres); blue, IgD staining (follicle); red,
CD3ε (T cell zone). Scale bar, 200 μm. b, Gating scheme for quantifying GC and
SPPC frequencies. c, d, Representative contour plots (top) and summary data
(bottom) of frequencies of FashiGL7hi GC B cells (c) and CD138+ SPPCs (d) 7 or
13 days after immunization. Summary data pooled from five independent
experiments; each symbol indicates one mouse, lines indicate means. In two
experiments both day 7 and day 13 were examined (red, blue), and in the other
three only day 13 was examined (grey). Two-tailed unpaired t-test. S,
sham-operated; D, denervated.

ad

b

GC (%)

0

5

10

15

SPPC (%)

0

1

2

3

4

5

GC

SPPC

TotalB cells

CHRNA1CHRNA3CHRNA5CHRNA7CHRNA10CHRNDCHRNB1CHRNB3CHRM1CHRM2CHRM4CHRNA2CHRNA4CHRNA6CHRNA9CHRNECHRNGCHRNB2CHRNB4CHRM1CHRM3CHRM5Marker

+/+ –/– c

e

Sham

Denervated

15.5 3.39

9.32 10.3

5.35 5.23

GL7

FA

S

2.26 1.53

B220

CD138

+/+ –/– +/+ –/–

B220

CD138

+/+ –/–

SPPC (%)

Sham Denervated

0

5

10

15

20

+/+–/–+/+–/–

P = 0.5 P = 0.0003

P = 0.0008

P = 0.0497

P = 0.3

Fig. 2 | B cell-intrinsic responsiveness to acetylcholine underlies plasma

cell formation promoted by splenic nerve activities. a, Expression of

nicotinic acetylcholine receptor subunits in sorted total B cells (7A AD−CD19+),

GC B cells (7A AD−CD19+GL7+Fas+) and SPPCs (7A AD−CD138+). Arrowheads,

correct α7, β1, α9, and β4 amplicon sizes. b, c, Representative contour plots (b)

and summary data (c) of percentage SPPC and GC in bone-marrow chimaeras

reconstituted with 80% μMT and 20% Chrna9+/+ (+/+) o r Chrna9−/− (−/−)

bone-marrow cells, 13 days after immunization. Data pooled from four

independent experiments. Each symbol indicates one mouse; lines denote

means. d, e, Representative contour plots (d) and summary data (e) of

percentage SPPC in sham-operated or denervated μMT mice that were

reconstituted with intravenously infused Chrna9+/+ or Chrna9−/− mature B cells.

Each symbol indicates one mouse; lines denote means. Two-tailed unpaired

t-tests (c, e).