ArticleA purine metabolic checkpoint that prevents
autoimmunity and autoinflammationGraphical abstractFAMIN p.254V FAMIN p.284RFAMIN p.254I
Immunopathology
Dendritic 
cell
T cell
Proliferation
IFN-γ
Influenza A
0
50
100 0
50
100 0
50
100
inosinehypoxanthine
inosine
inosine
hypoxanthine
hypoxanthine
MHC I
antigen
A2AR
T cell
TCR
Dendritic 
cell
F
Succinyl-
AMP
AMP
adenine
IMP
GMP
NAD+
XMP
NADH
adenosine
MTAguanine
guanosine
IMPDH
inosine
F
F
F
FF FHighlightsd A FAMIN-enabled purine metabolon in dendritic cells
restrains T cell priming
d FAMIN prevents cytoplasmic NADH/NAD+ reductive stress
that enhances priming
d Inosine generated from hypoxanthine by FAMIN inhibits T cell
priming
d FAMIN ameliorates immunopathology in influenza but
dampens tumor surveillanceSaveljeva et al., 2022, Cell Metabolism 34, 106–124
January 4, 2022 ª 2021 The Authors. Published by Elsevier Inc.
https://doi.org/10.1016/j.cmet.2021.12.009Authors
Svetlana Saveljeva, Gavin W. Sewell,
Katharina Ramshorn, ..., Allan Bradley,
Gordon Dougan, Arthur Kaser
Correspondence
ak729@cam.ac.uk
In brief
Saveljeva et al. identify a biochemical
mechanism in dendritic cells that
restrains T cell priming and prevents
immunopathology but dampens tumor
surveillance. FAMIN enables a purine
nucleotide cycle, which prevents
cytoplasmic NADH/NAD+ reductive
stress that augments antigen
presentation, and it generates inosine,
which inhibits T cell activation.ll
OPEN ACCESS
llArticle
A purine metabolic checkpoint that prevents
autoimmunity and autoinflammation
Svetlana Saveljeva,1,2,6 Gavin W. Sewell,1,2,6 Katharina Ramshorn,1,2,6 M. Zaeem Cader,1,2 James A. West,1,2
Simon Clare,3 Lea-Maxie Haag,2 Rodrigo Pereira de Almeida Rodrigues,1,2 Lukas W. Unger,1,2
Ana Bele´n Iglesias-Romero,1,2 Lorraine M. Holland,1,2 Christophe Bourges,1,2 Muhammad N. Md-Ibrahim,2
James O. Jones,2 Richard S. Blumberg,4 James C. Lee,1,2 Nicole C. Kaneider,1,2 Trevor D. Lawley,3 Allan Bradley,1
Gordon Dougan,1,5 and Arthur Kaser1,2,7,*
1Cambridge Institute of Therapeutic Immunology and Infectious Disease, Jeffrey Cheah Biomedical Centre, University of Cambridge,
Cambridge CB2 0AW, UK
2Division of Gastroenterology and Hepatology, Department of Medicine, University of Cambridge, Addenbrooke’s Hospital, Cambridge CB2
0QQ, UK
3Wellcome Trust Sanger Institute, Hinxton, Cambridge CB10 1SA, UK
4Division of Gastroenterology, Hepatology and Endoscopy, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical
School, Boston, MA 02115, USA
5Division of Infectious Diseases, Department of Medicine, University of Cambridge, Cambridge CB2 0QQ, UK
6These authors contributed equally
7Lead contact
*Correspondence: ak729@cam.ac.uk
https://doi.org/10.1016/j.cmet.2021.12.009SUMMARYStill’s disease, the paradigm of autoinflammation-cum-autoimmunity, predisposes for a cytokine storm with
excessive T lymphocyte activation upon viral infection. Loss of function of the purine nucleoside enzyme
FAMIN is the sole known cause for monogenic Still’s disease. Here we discovered that a FAMIN-enabled pu-
rine metabolon in dendritic cells (DCs) restrains CD4+ and CD8+ T cell priming. DCs with absent FAMIN ac-
tivity prime for enhanced antigen-specific cytotoxicity, IFNg secretion, and T cell expansion, resulting in
excessive influenza A virus-specific responses. Enhanced priming is already manifest with hypomorphic
FAMIN-I254V, for which
6%ofmankind is homozygous. FAMIN controlsmembrane trafficking and restrains
antigen presentation in an NADH/NAD+-dependent manner by balancing flux through adenine-guanine
nucleotide interconversion cycles. FAMIN additionally converts hypoxanthine into inosine, which DCs
release to dampen T cell activation. Compromised FAMIN consequently enhances immunosurveillance of
syngeneic tumors. FAMIN is a biochemical checkpoint that protects against excessive antiviral T cell
responses, autoimmunity, and autoinflammation.INTRODUCTION
Deorphaning an autoimmunity risk gene product unearthed an
unprecedented function at the heart of cellular metabolism,
conserved from bacteria to man. This risk gene encodes FAMIN
(also known as LACC1, C13orf31), an enzyme that unifies in a
single protein the activities of adenosine deaminase (ADA;
adenosine + H2O . inosine + NH3), purine nucleoside phos-
phorylase (PNP; inosine + phosphate [Pi] # hypoxanthine +
ribose-1-phosphate [R1P]; guanosine + Pi # guanine + R1P),
andmethylthioadenosine phosphorylase (MTAP;methylthioade-
nosine [MTA] + Pi# adenine + methyl-thioribose-1-phosphate
[MTR1P]). FAMIN’s fourth catalytic activity is that of an adeno-
sine phosphorylase (adenosine + Pi# adenine + R1P), previ-
ously considered absent from eukaryotic metabolism (Cader
et al., 2020). Adenine and ribose are primordial metabolites
from which life is thought to have emerged from prebiotic106 Cell Metabolism 34, 106–124, January 4, 2022 ª 2021 The Autho
This is an open access article under the CC BY license (http://creativebiochemistry (Ralser, 2018). They are defining constituents of
the genetic code, the energy currency, and the major cofactors
of a cell. Since purine nucleotide de novo synthesis yields
straight to nucleotides (i.e., purine monophosphates; IMP,
AMP, and GMP), ADA, PNP, and MTAP had been thought to
be the sole enzymes to supply purine nucleobases (adenine,
guanine, and hypoxanthine) from nucleosides (adenosine, gua-
nosine, inosine, and MTA) (Bzowska et al., 2000). ADA and
PNP deficiency causes severe combined immunodeficiency
(SCID) with loss of T and B lymphocytes (Giblett et al., 1972,
1975). Loss of function of FAMIN, in sharp contrast, is linked to
autoimmunity and autoinflammation, specifically to Still’s dis-
ease (also known as systemic juvenile idiopathic arthritis, sJIA),
juvenile idiopathic arthritis (JIA), and early-onset Crohn’s disease
(Al-Mayouf et al., 2020; Patel et al., 2014; Rabionet et al., 2019;
Wakil et al., 2015; Yasin and Schulert, 2018). These very rare
loss-of-function mutations aside, partial loss of activity, causedrs. Published by Elsevier Inc.
commons.org/licenses/by/4.0/).
ll
OPEN ACCESSArticleby a SNP that leads to a valine-for-isoleucine substitution at
amino acid 254 (I254V), for which 
6% of humans are homozy-
gous, increases risk of Crohn’s disease and leprosy (Barrett
et al., 2008; Zhang et al., 2009).
FAMIN loss of function is the sole known cause of autosomal-
recessive (i.e., monogenic) forms of Still’s disease. Still’s disease
affects young children; starts with daily recurring fever, rash, and
lymph node enlargement; and morphs over weeks into a debili-
tating arthritis (Yasin and Schulert, 2018). The initial phase re-
sembles periodic fever syndromes with inflammasome activa-
tion and IL-1b and IL-18 release; the later arthritic phase is
thought to be driven by pathogenic T lymphocytes. About 20%
of children with Still’s disease develop ‘‘secondary hemophago-
cytic lymphohistiocytosis’’ (HLH; also known as ‘‘macrophage
activation syndrome,’’ MAS) (Brisse et al., 2016b; Grom et al.,
2016). HLH/MAS is typically triggered by viral infections and
can occur even when Still’s disease is in remission while on IL-
1/IL-6-blocking therapeutics (Grom et al., 2016). HLH features
a cytokine storm accompanied by excessive expansion and acti-
vation of CD4+ and CD8+ T lymphocytes and hemophagocytic,
IFNg-activated macrophages. It manifests with disseminated
intravascular coagulation (DIC), acute respiratory distress syn-
drome, and multi-organ failure, and is often fatal (Brisse et al.,
2016a, 2016b). HLH/MAS is not restricted to Still’s disease and
children. For example, virus-induced HLH/MAS is caused by
Epstein-Barr virus and many other pathogens and has been
implicated in fatality from seasonal (H3N2), avian (H5N1), and
swine (H1N1/2009) influenza A virus (IAV) infections (Beutel
et al., 2011; Henter et al., 2010).
Mice with germline deletion of Famin, or genome-edited to ex-
press one of the Still’s disease-linked loss-of-function mutations
(C284R; ‘‘Faminp.284R’’ mice), develop normally under specific
pathogen-free conditions. Similarly, mice genome-edited to ex-
press fully active (254I; ‘‘Faminp.254I’’) or partially active FAMIN
(254V; ‘‘Faminp.254V’’) are indistinguishable (Cader et al., 2016).
Famin/ and Faminp.284R mice, however, do develop more se-
vere lipopolysaccharide (LPS)-induced sepsis, evidence of
DIC, and increased plasma IL-1b levels, compared to mice ex-
pressing fully active FAMIN (Cader et al., 2016). Compromised
FAMIN activity also leads to lower reactive oxygen species
(ROS) production, decreased bacterial killing, altered NLRP3 in-
flammasome activation, and cytokine secretion in macrophages
(Cader et al., 2016; Lahiri et al., 2017), and Famin/ mice
develop more severe experimental arthritis and colitis (Kang
et al., 2020; Skon-Hegg et al., 2019). How loss of FAMIN activity,
which is abundantly expressed in macrophages and dendritic
cells (DCs) while largely absent from T cells (Heng et al., 2008),
predisposes for autoimmunity remains unknown. Particularly
elusive is via what mechanism altered core purine metabolism
due to the absence of multifunctional FAMIN, which is tethered
to the cytoplasmic surface of peroxisomes (Cader et al., 2016),
could affect immune function, since monofunctional ADA,
PNP, and MTAP are ubiquitously present.
Here we report a purine metabolon in DCs that potently re-
strains T cell priming by dampening membrane trafficking and
hence the pace of antigen uptake and presentation, and by
releasing inosine that provides an inhibitory signal via the aden-
osine A2A receptor (A2AR). Impaired FAMIN catalysis results in
excessive IAV-specific T cell responses and lung immunopa-thology, but also in enhanced tumor immune surveillance. We
describe a purely biochemical mechanismwithin DCs that exerts
fundamental control over T lymphocyte priming.
RESULTS
FAMIN activity in DCs restrains the influenza A virus-
specific T cell response
Originally aiming to gain clues into HLH predisposition, we in-
fected Famin mutant mice with a murine-adapted H3N2 IAV
strain (A/X-31) (Everitt et al., 2012). Faminp.254V and Faminp.284R
mice, which endogenously express disease-linked hypomorphic
and loss-of-function variants, respectively, developed more se-
vere disease compared to Faminp.254I mice, which express fully
active FAMIN (Figures 1A and S1A). This was associated with
more apoptosis, reflecting lung damage (Figure 1B), and
elevated plasma IFNg and IL-10 levels (Figure 1C). Anti-inflam-
matory IL-10 is produced by IFNg+ IAV-specific CD8+ T cytotoxic
type 1 cells (Tc1) in IAV-infected lungs (Sun et al., 2009), prompt-
ing us to quantify CD8+ T cells specific for NP366-374, an immuno-
dominant IAV nucleoprotein epitope. On day 7 of infection,
numbers of NP366-374-reactive CD8+ T cells in bronchoalveolar
lavage (BAL) were 4-fold higher even in Faminp.254V compared
to Faminp.254I, and higher still in Faminp.284R mice (Figures 1D
and S1B). Cytokine production by Tc1 cells is dependent on their
interaction with, and costimulation by, CD11c+ DCs infiltrating
the infected lung (Hufford et al., 2011). Deletion of Famin solely
in CD11c+ DCs (Faminflox/flox;CD11c-Cre; ‘‘FaminDDC’’) (Fig-
ure S1C) increased the numbers of NP366-374-specific and
PA224-233 (an IAV polymerase acidic protein epitope)-specific
CD8+ T cells in BAL compared to littermate Faminflox/flox mice
(‘‘FaminWT’’; Figure 1E). In contrast to germ-line variation, selec-
tive deletion in FaminDDC mice did not increase immunopa-
thology (Figure S1D). Interestingly, the excessive IAV-specific T
lymphocyte response in Faminp.284R and FaminDDC lungs was
associated with somewhat higher expression of IAV M protein
compared to their respective controls (Figures 1F and S1E).
We concluded that reduced or absent FAMIN activity in DCs re-
sulted in exaggerated hyperinflammatory IAV-specific CD8+
T cell responses that did not augment control of the viral
infection.
FAMIN in DCs restrains priming of class I and class
II-restricted antigens
Intrigued by the selective increase in T cell responses emanating
from FAMIN deficiency in DCs, we focused our study on whether
and how FAMIN controls T cell priming and turned to ovalbumin
(OVA) as a model antigen. Baseline percentages of splenic and
lymph node CD4+ and CD8+ T lymphocytes, and splenic
cDC1s and cDC2s, were indistinguishable between Faminp.254I,
Faminp.254V, and Faminp.284R mice (Table S1). Splenic CD11c+
DCs from FaminDDC mice primed naive OVA257-264-specific
OT-I T lymphocytes for increased expansion and IFNg secretion
compared to those primed by FaminWT DCs, irrespective of
whether they were pulsed with OVA257-264 peptide, OVA protein,
or necrotic fibroblasts expressing a non-secreted OVA (bm1T-
OVA; Sancho et al., 2009) requiring cross-presentation (Figures
2A–2C). Antigen-specific cytotoxicity, IFNg, and granzyme B
release were higher when naive OT-I T cells had been primedCell Metabolism 34, 106–124, January 4, 2022 107
A1 2 3 4 5 6 7
85
90
95
100
105
W
ei
gh
t (
%
)
Days
Faminp.254V
Faminp.254I
Faminp.284R
** *
IF
N
γ 
(p
g/
m
L)
Faminp.254V
Faminp.254I
Faminp.284R
500
1000
1500 *
0
D
C
D
8+
 N
P
36
6-
37
4 
 (×
10
4  c
el
ls
)
Faminp.254V
Faminp.254I
Faminp.284R
0
10
20
30 *
FaminWT
Famin∆DC
0
5
10
15
20 **
C
D
8+
 N
P
36
6-
37
4 
 (×
10
4  c
el
ls
)
0
5
10
15
20
25 *
C
D
8+
 P
A
22
4-
23
3  (
×1
04
 c
el
ls
)
FaminWT
Famin∆DC
E
0
10
20
30
40
TU
N
EL
+  (
ce
lls
/a
re
a)
*
*
Faminp.254V
Faminp.254I
Faminp.284R
B
Faminp.254V
Faminp.254I
Faminp.284R
0
5
10
15
20
25
IL
-1
0 
(n
g/
m
L)
*
*
C
F
Faminp.254V
Faminp.254I
Faminp.284R
In
flu
en
za
 M
/A
ct
B
0
5
10
15
20
*
Figure 1. FAMIN deficiency augments T cell responses to influenza A virus
(A) Percentage weight loss of Faminp.254I, Faminp.254V, Faminp.284R mice, which are engineered to express fully active, hypomorphic, and inactive FAMIN,
respectively, following infection with influenza A virus (IAV) H3N2, strain A/X-31 (n = 6/11/6).
(B) Average number of TUNEL+ cells in lungs of Faminp.254I, Faminp.254V, and Faminp.284R mice 7 days post-infection (n = 6/9/6).
(C) Plasma IFNg and IL-10 levels 7 days after IAV infection (n = 6/11/6). Note IL-12p70 was below the detection limit.
(D and E) Absolute numbers of IAV NP366-374 tetramer+ CD8+ T cells in Faminp.254I, Faminp.254V, and Faminp.284R (D), or CD3+CD8+NP366-374 and PA224-233
tetramer+ cells in FaminWT and FaminDDC (E) bronchoalveolar lavage fluid (BAL) 7 days after infection (n = 6/11/6; 15/14).
(F) IAV M protein gene expression in lung tissue of Faminp.254I, Faminp.254V, and Faminp.284R mice 7 days after infection (n = 6/11/6).
Data represented as mean ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001 (repeated-measures one-way ANOVA, one-way ANOVA, or unpaired two-tailed
Student’s t test where appropriate). See also Figure S1.
ll
OPEN ACCESS Articleby FaminDDC than by FaminWT splenic DCs (Figures 2D and 2E).
CD8a+ conventional DCs type 1 (cDC1) preferentially prime naive
CD8+ T cells, and CD11b+ cDC2 preferentially CD4+ T cells (Du-
rai and Murphy, 2016). Exaggerated cytotoxic T lymphocyte
(CTL) responses were similarly observed when primed by bone
marrow (BM)-derived Famin/ compared to Famin+/+ cDC1
(Figure S2A), hence extending to DCs immunologically distinct
from splenic DCs (Naik et al., 2005). BM-derived cDC1 from
Faminp.284R and Famin/ mice also primed OT-I T cells for
higher proliferation and IFNg secretion compared to those
from Faminp.254I and Faminp.254V mice (Figure S2B). Restimula-
tion of cDC1-OT-I T cell co-cultures after 72 h (Figure 2F), or after
further differentiation over 6 days into antigen-specific T effector
(TE) and T effector memory (TEM) cells via IL-2 and IL-15 (Fig-
ure 2G), respectively, resulted in highest IFNg secretion when
priming was provided by Faminp.284R and Famin/ cDC1, inter-108 Cell Metabolism 34, 106–124, January 4, 2022mediate by Faminp.254V, and lowest by Faminp.254I cDC1. Hence,
CD8+ T cell responses increased with decreasing FAMIN activity
in DCs, and the enhanced priming effect persisted when further
differentiated into TE and TEM cells.
To investigate this in vivo, OT-I T cells were adoptively trans-
ferred into FaminDDC and FaminWT mice followed by intraperito-
neal immunization with ovalbumin, and CTL activity was as-
sessed 4 days later. OVA257-264-specific cytotoxicity, IFNg, and
granzyme B release of splenic T cells were strikingly higher in
FaminDDC compared to FaminWT mice (Figure 2H). Increased
CTL activity was intrinsic, as OT-I proliferation in vivowas similar
between genotypes upon adjuvant-free priming (Figure S2C).
Hence, lack of FAMIN activity in DCs enhanced their ability to
prime CD8+ T cell responses to a model antigen in vivo.
The ability of FAMIN-impaired DCs to prime exaggerated anti-
gen-specific T cell responses extended to MHC II-restricted
AOVA257-264
0
2000
4000
6000
IF
N
γ 
(n
g/
m
L)
FaminWT
Famin∆DC
*
ovalbumin
0
500
1000
1500
2000
IF
N
γ 
(n
g/
m
L)
*
FaminWT
Famin∆DC
– + – +
D
0.0 0.5 1 1.5 2
0
20
40
60
80
E/T ratio
C
yt
ot
ox
ic
ity
 (
%
)
FaminWT
Famin∆DC
*
*
F
1 10 20
0
10
20
30
40
E/T ratio
C
yt
ot
ox
ic
ity
 (%
)
*
*
*
0
200
400
600
800
E/T ratio
G
ra
nz
ym
e 
B
 (p
g/
m
L)
*
*
1 10 20
FaminWT
Famin∆DC
H
FaminWT
Famin∆DC
0
1000
2000
3000
IF
N
γ 
(n
g/
m
L)
Faminp.254V
Faminp.254I
Faminp.284R
OVA257-264
– +
***
*
1 10 20
0
5
10
15
20
E/T ratio
IF
N
γ 
(n
g/
m
L)
*
**
*
FaminWT
Famin∆DC
0
1
2
3
4
*
OVA323-339
Famin+/+
Famin–/–
IF
N
γ 
(n
g/
m
L)
– +
J
Famin–/–
C
FaminWT
Famin∆DC
–
bm1 T OVA
0
200
400
600
800 *
IF
N
γ 
(n
g/
m
L)
+
B
1.1
1.2
1.3
1.4
1.5
1.6
1.7
P
ro
lif
er
at
io
n 
in
de
x
*
OVA257-264
FaminWT
Famin∆DC
2.0
2.2
2.4
2.6
2.8
P
ro
lif
er
at
io
n 
in
de
x
*
FaminWT
Famin∆DC
bm1 T OVA
1.00
1.25
1.50
1.75
2.00
P
ro
lif
er
at
io
n 
in
de
x
*
ovalbumin
FaminWT
Famin∆DC
0.5 1 2
0
10
20
30
40
50
E/T ratio
G
ra
nz
ym
e 
B
 (n
g/
m
L)
* *
*
0.5 1 2
0.0
0.5
1.0
1.5
2.0
2.5
E/T ratio
IF
N
γ 
(n
g/
m
L)
*
FaminWT
Famin∆DC
FaminWT
Famin∆DC
E
anti-CD3/CD28
0
5
10
15
20
25
IF
N
γ 
(n
g/
m
L)
Faminp.254V
Famin−/−
Faminp.254I
Faminp.284R
TE
0
20
40
60
80
100
anti-CD3/CD28
Faminp.254V
Famin−/−
Faminp.254I
Faminp.284R
TEM
IF
N
γ 
(n
g/
m
L)
– + – +
*
**
**
**
*
***
G
0
200
400
600
800
IL
-2
 (p
g/
m
L)
*
OVA323-339
Famin+/+
Famin–/–
– +
K
anti-CD3/CD28
0
20
40
60
80
100
IF
N
γ+
 (%
)
*
Famin+/+
Famin–/–
– +
L
FaminWT
Famin∆DC
2.0
2.5
3.0
3.5
P
ro
lif
er
at
io
n 
in
de
x
*
M
CD4+ OT-II T cells CD4+ OT-II T cells CD4+ OT-II T cells CD4+ OT-II T cells
Pr
ol
ife
ra
tio
n 
in
de
x
Famin+/+
Famin–/–
1.0
1.5
2.0
2.5 *
I
CD4+ OT-II T cells
Figure 2. DC FAMIN restrains CD4+ and CD8+ T cell responses
(A–C) IFNg release and OT-I T cell proliferation indices after 72 h of co-culture with splenic FaminWT or FaminDDC CD11c+ DCs pulsed with OVA257-264 peptide (A),
ovalbumin (B), or UV-irradiated bm1 T OVA mouse embryonic fibroblasts (C) (n = 3).
(legend continued on next page)
ll
OPEN ACCESSArticle
Cell Metabolism 34, 106–124, January 4, 2022 109
ll
OPEN ACCESS ArticleCD4+ T cells. Famin/ splenic CD11c+ DCs, pulsed with OVA323-
339, primed for increased proliferation of syngeneic naive OVA323-
339-specific OT-II T cells when compared to Famin+/+ DCs (Fig-
ure 2I). The levels of IFNg and IL-2 were 4- and 8-fold higher,
respectively, in supernatants from co-cultures with Famin/
compared to Famin+/+ DCs (Figures 2J and 2K). The proportion
of intracellular IFNg+ OT-II T cells upon restimulation increased
from 40% ± 8% to 75% ± 6% when they had been primed by
OVA323-339-pulsed Famin/ compared to Famin+/+ BM-derived
cDC2 (Figure 2L). The proportion of IL-4+, IL-17+, and Foxp3+
OT-II T cells remained below 1% (Figure S2D). Naive CD4+ OT-II
T cells adoptively transferred into FaminDDC mice exhibited
increased proliferation upon intraperitoneal OVA immunization
compared to those transferred into FaminWT mice (Figure 2M).
Altogether, impaired FAMIN in DCs increased antigen-specific
T cell priming via both class I and II in vitro and in vivo.
FAMIN controls DC metabolism and tunes antigen
uptake and presentation without a transcriptional
signature
Despite the profound differences in their priming activity, Famin
itself and genomically adjacent Ccdc122 were the sole differen-
tially expressed genes (DEGs) in Famin/ compared to Famin+/+
BM-derived cDC1 analyzed by RNA sequencing (RNA-seq; Fig-
ure 3A). A comparison of cDC1 from Faminp.254I and Faminp.254V
mice did not reveal a single DEG (Figure S3A), and only 56 up-
and 32 downregulated transcripts between Faminp.284R and
Faminp.254I cDC1 (Figure S3B; Table S2). Among those DEGs
were only four (Fcgr1, Tlr7, Ikbkg, and Lcn2) encoding immune
mediators, and no enrichment for gene ontology processes
linked to DC activation was observed (data not shown). Famin
genotype did not affect protein expression of ADA, PNP, and
MTAP (Figure S3C), which share catalytic activities with
FAMIN. Liquid chromatography-mass spectrometry (LC-MS)
demonstrated a marked reduction in purine nucleotides from
Faminp.254I via Faminp.254V to Faminp.284R cDC1 (Figures 3B
and 3C; Table S3), raising the possibility of a bona fide biochem-
ical mechanism controlling T cell priming.
Priming of naive T cells entails T cell receptor (TCR) binding to
a peptide-MHC complex on a professional antigen-presenting
cell, which is then fine-tuned by costimulatory molecules and
secreted mediators (Blum et al., 2013; Cantrell, 2015). A pulse
with a fluorescent AF647-ovalbumin conjugate (AF647-OVA) as
model antigen demonstrated higher uptake, particularly at earlier
time points, in Faminp.284R compared to Faminp.254I splenic DCs,(D and E) Specific cytotoxicity against OVA257-264-pulsedwild-type splenocytes (D
FaminWT or FaminDDC splenic DCs pulsed with OVA257-264 (n = 3).
(F) IFNg secretion from re-stimulated (OVA257-264 for 5 h) OT-I T cells after 72 h
cDC1 pulsed with OVA257-264 (n = 3).
(G) IFNg secretion after 5 h anti-CD3/CD28 re-stimulation of OT-I T cells that had b
Faminp.254V, Faminp.284R, and Famin/ BM-derived cDC1 pulsed with OVA257-26
(H) OVA257-264-specific cytotoxicity, granzyme B, and IFNg secretion of splenocy
naive OT-I T cells and immunized with ovalbumin 72 h earlier (n = 3).
(I) Proliferation indices of OT-II T cells 96 h after priming with OVA323-339-pulsed
(J and K) IFNg (J) and IL-2 (K) in supernatants of OT-II cells 96 h after priming wi
(L) Percentage IFNg+ OT-II T cells after restimulation with anti-CD3/CD28, follow
(M) Proliferation indices of OT-II T cells adoptively transferred into FaminDDC and
Data represented as mean ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001 (one-wa
Figure S2.
110 Cell Metabolism 34, 106–124, January 4, 2022with intermediate levels in Faminp.254V cells (Figure 3D). Higher
AF647-OVA uptake in Faminp.284R and Faminp.254V compared
to Faminp.254I genotypes was observed in both cDC1 and
cDC2 splenic DCs, as well as in BM-derived cDCs (Figures
S3D and S3E). This suggested that in the absence of FAMIN ac-
tivity, antigen uptake was accelerated and available for presen-
tation via MHC class II for CD4+ T cells and cross-presentation
via MHC class I for CD8+ T cells, the latter requiring endo-
some-to-cytosol transfer (Blander, 2018). This can be measured
using endocytosed b-lactamase in DCs that are pre-loaded with
a cytosolic probe that loses its FRET signal upon b-lactamase
cleavage, when the latter gains access to the cytosol (Cebrian
et al., 2011). Compared to Faminp.254I cells, Faminp.284R splenic
DCs exhibited increased probe cleavage, especially at the
earliest time point, demonstrating increased endosome-to-
cytosol transfer (Figure 3E). The peptide repertoire presented
on surfaceMHC I is continuously optimized by peptide exchange
in the endoplasmic reticulum (ER) (Williams et al., 2002).
Increased staining with monoclonal antibody 25-D1.16, which
recognizes OVA257-264 bound to H-2Kb (Porgador et al., 1997),
directly demonstrated increased peptide presentation on
Faminp.284R and Faminp.254V compared to Faminp.254I splenic
DCs after a pulse with OVA257-264 (Figures 3F and S3F). The dif-
ference between Famin genotypes in peptide:MHC I complexes
was again most pronounced early after the OVA257-264 pulse,
indicative of FAMIN controlling the pace of the process. Famin
genotype did not affect total surface MHC I and II expression
(Table S4; Figure S3G). Hence, loss of FAMIN activity led to
faster-paced endosomal antigen uptake, transfer to cytosol,
and peptide exchange and presentation on MHC I.
FAMIN can promote flux through a cycle that interconverts IMP
to succinyl-AMP (S-AMP), AMP, and back to IMP via sequential
activities of adenylosuccinate synthase (ADSS), adenylosuccinate
lyase (ADSL), and AMP deaminase (AMPD) (Figure 3G) (Cader
et al., 2020). In skeletal muscle and macrophages, the IMP–S-
AMP–AMP cycle promotes energy metabolism and is referred to
as purine nucleotide cycle (PNC) (Cader et al., 2020; Lowenstein
and Tornheim, 1971). Cellular levels of IMP, S-AMP, and AMP
decreased from Faminp.254I and Faminp.254V to Faminp.284R BM-
derived cDC1 (Figures 3B and S3H). Tracing [13C16] palmitate,
we observed decreased flux into Krebs cycle metabolites a-keto-
glutarate, succinate, and malate in Faminp.254V and Faminp.284R
compared to Faminp.254I BM-derived DC1s (Figure S3I). We also
detected decreased onward flux into aspartate (which enters
the IMP–S-AMP–AMP cycle) in Faminp.284R DCs. A similar pattern), and IFNg and granzymeB release (E) of OT-I T cells that had been primedwith
of priming with Faminp.254I, Faminp.254V, Faminp.284R, or Famin/ BM-derived
een differentiated into TE and TEM cells, following 72 h of priming by Famin
p.254I,
4 (n = 3).
tes of FaminDDC and FaminWT mice that had been adoptively transferred with
splenic Famin+/+ and Famin/ DCs (n = 3).
th OVA323-339-pulsed splenic Famin+/+ and Famin/ DCs (n = 3).
ing priming with OVA323-339-pulsed BM-derived cDC2 7 days earlier (n = 3).
FaminWT mice 72 h after immunization with ovalbumin (n = 3).
y ANOVA or unpaired two-tailed Student’s t test where appropriate). See also
Faminp.254I Faminp.254V Faminp.284R
A
F
Z score
-2 -1 0 1 2
1x10
-4
0.01
1.0
-15 -10 -5 0 5 10 15
FD
R
-a
dj
us
te
d 
p-
va
lu
e Famin
Ccdc122
1x10-6
0.001
0.1
1x10
-5
1x10-7
Decreased expression 
in Famin–/– versus Famin+/+
Increased expression 
in Famin–/– versus Famin+/+
Log2 fold change
Faminp.254V
Faminp.254I
Faminp.284R
0
20
40
60
Time (minutes)
**
** *
1
2
3
4
5
(R
es
po
ns
e 
ra
tio
)
*
*
Faminp.254I
Faminp.284R
D
B
Time (minutes)
E
N
IMP Succinyl-AMP AMPAspartate
0
5×105
1×106
1.5×106
1×107
2×107
3×107
4×107
5×107
P
ea
k
ar
ea
***
***
***
***
***
***
**
*
Faminp.254V
Faminp.254I
Faminp.284R
– 15 30 60 90
ATP
ADP
AMP
GTP
GDP
GMP
IMP
guanine
guanosine
F
Succinyl-
AMP
AMP
AMPD
ADSS
aspartate + GTP
GDP + Pi
ADSL
fumarate
NH3
adenosine
adenine
F
inosine
F
hypoxanthine
F
SAM SAH
dcSAM
MTA
F
malate
IMP
GMP
GMPR
GMPS
+ glutamate
ATP + glutamine
IMPDH
NAD+
NADPH + H+
XMP
H+ + NADH NADP+ + NH3
AMP + PPi
F
MTAP
ADA
PNP
PNP
C
0
10
85
90
95
100
Time (minutes)
H
-2
Kb
-O
VA
25
7-
26
4  (
%
)
15 30 60 120–
Faminp.254V
Faminp.254I
Faminp.284R
0
10
20
30
40
E
C
A
R
(m
pH
/m
in
)
Faminp.254I
Faminp.254V
Faminp.284R
***
Faminp.254I
Faminp.254V
Faminp.284R
BM-derived cDC1BM-derived cDC1
1 25 49 73 97 121
0
20
40
60
80
100
120
Time (minutes)
O
C
R
(p
m
ol
es
/m
in
)
Oligo FCCP Rot + Ant
H I
G
J
20000
25000
30000
35000
40000
C
yt
op
la
sm
ic
pH
ro
do
(R
FU
)
*
***
Faminp.254I
Faminp.254V
Faminp.284R
Splenic DCs
0
P
ea
k
ar
ea
5×107
4×107
3×107
2×107
1×107
***
**
Unlabelled malate
Faminp.254V
Faminp.254I
Faminp.284R
0
20
40
60
80
Fr
ac
tio
na
l i
nc
or
po
ra
tio
n 
(%
 o
f t
ot
al
)
***
Faminp.254V
Faminp.254I
Faminp.284R
M
0
4×107
La
ct
at
e
(p
ea
k
ar
ea
)
***
***
3×107
5×107
2×107
1×107
Faminp.254I
Faminp.254V
Faminp.284R
Splenic DCs
K
C
yt
op
la
sm
ic
 a
nt
ig
en
 tr
an
sf
er
A
F6
47
-O
V
A
+  (
%
)
13C4
15N1
Faminp.254V
Faminp.254I
Faminp.284R
0
20
40
60
80
100
Fr
ac
tio
na
li
nc
or
po
ra
tio
n
(%
of
to
ta
l)
Aspartate M+5
L
-5 0 5
0.01
0.1
1.0
UDP Hypotaurine
UTP ATP
GDPADP
FD
R
-a
dj
us
te
d 
p-
va
lu
e
Decreased metabolite abundance 
in Faminp.284R versus Faminp.254I
Increased metabolite abundance 
in Faminp.284R versus Faminp.254I
Log2 fold change
*** **
13C4 Malate (M+4)
Figure 3. FAMIN controls DC metabolism and tunes antigen uptake and presentation without a transcriptional signature
(A) Differentially expressed genes between Famin/ and Famin+/+ BM-derived cDC1s (n = 4; GEO: GSE126473).
(B) Heatmap of purine nucleotide levels in Faminp.254I, Faminp.254V, and Faminp.284R cDC1s (n = 6/5/6); for details, see Table S3.
(legend continued on next page)
ll
OPEN ACCESSArticle
Cell Metabolism 34, 106–124, January 4, 2022 111
ll
OPEN ACCESS Articlein labeled Krebs cycle metabolites was present after a pulse with
[13C6] glucose, where onward flux (via pyruvate dehydrogenase)
into [13C2] aspartate decreased, and onward flux (via pyruvate
carboxylase or malic enzyme) into [13C3] aspartate increased in
FAMIN-impaired DCs (Figure S3I). In contrast, flux from
[13C5
15N2] glutamine into Krebs cycle metabolites and aspartate
increased in FAMIN-impaired cDC1s (Figure S3I). Overall, this
was consistent with perturbed fatty acid oxidation (FAO) and lipid
carbon channeling into the PNC, as well as compensatory
changes in glutamine metabolism, mirroring key observations in
FAMIN-deficient macrophages (Cader et al., 2020). Conse-
quently, cDC1s’ oxygen consumption rate (OCR), reflecting
oxidative phosphorylation (OXPHOS), decreased from Faminp.254I
via Faminp.254V to Faminp.284R genotypes (Figure 3H). The extra-
cellular acidification rate (ECAR) was also lower in Faminp.284R
compared to Faminp.254I BM-derived cDC1s (Figure 3I), and
secretion of lactate, with which protons (H+) are co-exported,
correspondingly declined from Faminp.254I to Faminp.254V and
Faminp.284R DCs (Figure 3J). Corresponding observations were
made in splenic DCs (Figure S3J), in which cDC2 predominate
over cDC1 (Table S1). The cytoplasmic pH (pHc) of cDC1 (data
not shown) and splenic DCs becamemore acidic as FAMIN activ-
ity decreased (Figure 3K). This demonstrated that FAMIN pro-
motes DCs’ energymetabolism and prevents cytoplasmic acidifi-
cation. By consuming aspartate and releasing its carbons as
fumarate, which can be hydrated to malate, the IMP–S-AMP–
AMP cycle can affect electron (e) transfer between cytoplasm
and mitochondria, which ensues via the malate-aspartate shuttle
(Borst, 2020; Cader et al., 2020). The aspartate pool supplying the
IMP–S-AMP–AMP cycle was inaccessible by exogenously sup-
plied [13C4
15N1] aspartate (Figure 3L), similar to most cells in cul-
ture (Birsoy et al., 2015). Fractional incorporation of exogenously
provided [13C4] malate was strikingly higher in Famin
p.284R
compared to Faminp.254I and Faminp.254V BM-derived DCs, as
levels of unlabeled malate were conversely lowest in Faminp.284R
and highest inFaminp.254I cells (Figures 3MandS3K). These differ-
ences in cellular malate, encompassing cytoplasmic and mito-(C) Differential metabolite abundance between Faminp.254I and Faminp.284R BM-
ionization modes) highlighted in red and annotated (n = 5).
(D) Percentage AF647-OVA+ Faminp.254I, Faminp.254V, or Faminp.284R splenic CD1
(E) Endosome-to-cytosol transfer in b-lactamase-loaded Faminp.254I and Faminp
(F) Percentage splenic CD11c+ DCs staining positive for OVA257-264 bound to H-
(G) Schematic depiction of the IMP–S-AMP–AMP cycle and IMP–XMP–GMP cycle
synthase; ADSL, adenylosuccinate lyase; AMPD, AMP deaminase; IMPDH, IMP
poxanthine guanine phosphoribosyltransferase; APRT, adenine phosphoribosyltra
MTA, methylthioadenosine; MTAP, MTA phosphorylase; SAM, S-adenosylmethio
(H) Oxygen consumption rate (OCR) of Faminp.254I, Faminp.254V, and Faminp.284R
rotenone plus antimycin A (Rot + Ant) (n = 16–18, from 3 mice per genotype).
(I) Basal extracellular acidification rate (ECAR) of Faminp.254I, Faminp.254V, and Fa
(J) Secreted lactate levels in supernatants of Faminp.254I, Faminp.254V, and Faminp.
per genotype).
(K) Cytoplasmic pH (pHc) of Famin
p.254I, Faminp.254V, and Faminp.284R splenic DC
(L) Fractional labeling of cellular aspartate as [13C4
15N1] isotopomer in Famin
p.2
[13C4
15N1] aspartate (n = 6, from 3 mice per genotype).
(M) Fractional labeling of [13C4] malate and levels of unlabeled (M+0) malate follow
malate (n = 9/9/5, from 3/3/2 mice per genotype).
(N) Total levels of aspartate, IMP, succinyl-AMP, and AMP following a 3 h pulse o
from 3/3/2 mice per genotype).
Data represented as mean ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001 (one-wa
Figure S3.
112 Cell Metabolism 34, 106–124, January 4, 2022chondrial pools, corroborated that energy metabolism is perva-
sively perturbed in FAMIN-impaired DCs. Exogenous malate
resulted in marked differences in levels of aspartate, IMP, S-
AMP, and AMP between Famin genotypes (Figure 3N). Altogether
this pointed, in analogy tomacrophages (Cader et al., 2020), to the
IMP–S-AMP–AMP cycle as an immediate biochemical effector of
FAMIN catalysis.
Adenine-guanine nucleotide interconversion paces
antigen uptake and T cell priming
We therefore asked whether IMP–S-AMP–AMP cycling restrains
DC antigen uptake. Halting the IMP–S-AMP–AMP cycle with L-
alanosine and hadacidin, IMP- and aspartate-analog inhibitors
of ADSS (Guicherit et al., 1994), respectively, indeed increased
AF647-OVA uptake in Faminp.254I BM-derived cDC1 to levels
observed in Faminp.284R cells (Figures 4A and 4B). In contrast,
blocking ADSS in Faminp.284R cDC1 did not further increase
AF647-OVA uptake (Figures 4A and 4B). Increased antigen up-
take by L-alanosine, conditional on Famin genotype, was simi-
larly observed in BM-derived cDC2 (Figure S4A). 6-thio-IMP, a
metabolite of clinically used immunomodulators 6-mercaptopu-
rine (6-MP) and azathioprine (Hanauer et al., 2019; Tiede et al.,
2003), has been reported to inhibit ADSS (Atkinson et al.,
1964). 6-MP phenocopied L-alanosine and hadacidin on
Famin-dependent AF647-OVA uptake in splenic DCs (Figure 4C),
affecting cDC1 and cDC2 subsets equally (Figure S4B). Trans-
fection of splenic DCs with Adss small interfering RNA (siRNA)
increased AF647-OVA uptake in Faminp.254I cells to levels
observed in control-transfected Faminp.284R cells, while not
further increasing uptake in the latter (Figure S4C). Inhibition of
AMPD with Cpd3 (Admyre et al., 2014) recapitulated effects of
ADSS inhibition (Figure 4D), corroborating that FAMIN-enabled
IMP–S-AMP–AMP cycling restrains antigen uptake. We next as-
sessed whether increased antigen uptake upon blocking the
IMP–S-AMP–AMP cycle translates into enhanced T cell priming.
Compared to control-silenced OVA-pulsed splenic Faminp.254I
DCs, those silenced for Adss, Adsl, or Ampd2/Ampd3 primedderived cDC1s; identifiable differential LC-MS features (positive and negative
1c+ DCs following uptake of AF647-OVA for indicated times (n = 3).
.284R cDC1s (n = 6/6, from 3 mice per genotype).
2Kb, after incubation with OVA257-264 for indicated times (n = 3).
, with relationship to FAMIN products and substrates. ADSS, adenylosuccinate
dehydrogenase; GMPS, GMP synthase; GMPR, GMP reductase; HPRT, hy-
nsferase, ADA, adenosine deaminase; PNP, purine nucleoside phosphorylase;
nine; dcSAM, decarboxylated SAM.
BM-derived cDC1s. Basal OCR followed by oligomycin A (Oligo), FCCP, and
minp.284R BM-derived cDC1s (n = 16–18, from 3 mice per genotype).
284R splenic DCs after 3 h incubation in OptiMEMmedium (n = 8–9, from 3mice
s measured using pHrodo indicator probe (n = 12, from 3 mice per genotype).
54I, Faminp.254V, and Faminp.284R BM-derived cDC1 following a 3 h pulse with
ing a 3 h pulse of Faminp.254I, Faminp.254V, and Faminp.284R cDC1s with [13C4]
f Faminp.254I, Faminp.254V, and Faminp.284R cDC1s with [13C4] malate (n = 9/9/5,
y ANOVA or unpaired two-tailed Student’s t test where appropriate). See also
IH
10
15
20
25
***
A
F6
47
-O
V
A
+  (
%
)
Splenic DCs
GMPS
Psicofuranine–
Faminp.254V
Faminp.254I
Faminp.284R
J L
0
10
20
30
40 *
IF
N
γ 
(n
g/
m
L)
si ctrl 
si Gmpr2
si Gmps
si Impdh1/Impdh2
Faminp.254V splenic DC-OT-I T cells
K
0
0.2
0.4
0.6
0.8 *
IF
N
γ 
(n
g/
m
L)
si ctrl 
si Gmpr2
si Gmps
si Impdh1/Impdh2
Faminp.254I splenic DC-OT-I T cells
45
*
**
Splenic DCs
15
20
25
30
35
40
*
A
F6
47
-O
V
A
+  (
%
)
IMPDH
MPA–
Faminp.254V
Faminp.254I
Faminp.284R
0.0
12
13
14
15
16
Fr
ac
tio
na
li
nc
or
po
ra
tio
n
(%
of
to
ta
l)
GMP
15N1
Faminp.254I
Faminp.254V
Faminp.284R
***
15N2 15N3 15N4 15N5
0
10
20
30
*
IF
N
γ 
(n
g/
m
L)
si ctrl 
si Gmpr2
si Gmps
si Impdh1/Impdh2
Faminp.284R splenic DC-OT-I T cells
M N
0-103 103 104 105
1.2
1.4
1.6
1.8
P
ro
lif
er
at
io
n 
in
de
x
*
si ctrl 
si Gmpr2
si Gmps
si Impdh1/Impdh2
Faminp.254I splenic DC-OT-I T cells
0
20
40
60
80
100
N
or
m
al
iz
ed
 T
o 
M
od
e
CFSE
O
1.6
1.7
1.8
1.9
2.0
P
ro
lif
er
at
io
n 
in
de
x
1.5
1.6
1.7
1.8
1.9
2.0
P
ro
lif
er
at
io
n 
in
de
x
si ctrl 
si Gmpr2
si Gmps
si Impdh1/Impdh2
si ctrl 
si Gmpr2
si Gmps
si Impdh1/Impdh2
Faminp.284R splenic DC-OT-I T cellsFaminp.254V splenic DC-OT-I T cells
P
C
BM-derived cDC1
D
Faminp.254V
Faminp.254I
Faminp.284R
A
F6
47
-O
V
A
+  (
%
)
Splenic DCs
A
F6
47
-O
V
A
+  (
%
)
BM-derived cDC1
Faminp.254V
Faminp.254I
Faminp.284R
A
A
F6
47
-O
V
A
+  (
%
)
30
40
50
60
70
80
50
60
70
80 **
20
25
30
35
Faminp.254V
Faminp.254I
Faminp.284R
ADSS
L-alanosine–
ADSS
6-MP–
AMPD
Cpd3–
BM-derived cDC1
Faminp.254V
Faminp.254I
Faminp.284R
B
30
40
50
60
70
80
A
F6
47
-O
V
A
+  (
%
)
ADSS
Hadacidin–
* *
+ Cpd3- Cpd3G
AMP
ADP
ATP
GMP
GDP
GTP
XMP
IMP
Succinyl-AMP
Z score
-2 -1 0 1 2
**
***
***
Faminp.254I splenic DC-OT-I T cells
1.7
1.8
1.9
2.0
2.1
2.2
P
ro
lif
er
at
io
n 
in
de
x
si ctrl 
si Adsl
si Adss
si Ampd2/Ampd3
0-103 103 104 105
0
20
40
60
80
100
N
or
m
al
iz
ed
 T
o 
M
od
e
CFSE
E F
0
20
40
60
80
100 *
IF
N
γ 
(n
g/
m
L)
si ctrl 
si Adsl
si Adss
si Ampd2/Ampd3
Faminp.254I splenic DC-OT-I T cells
***
Figure 4. Adenine-guanine nucleotide interconversion paces antigen uptake and T cell priming
(A–D) Percentage AF647-OVA+ cDC1s (A, B, and D) or splenic DCs (C) following incubation with AF647-OVA for 30 min in the presence of L-alanosine (A),
hadacidin (B), 6-mercaptopurine (6-MP) (C), or Cpd3 (D) (n = 3–6, 3mice per genotype; none of the treatments affected cell viability; please note control panels are
shared between B and D).
(legend continued on next page)
ll
OPEN ACCESSArticle
Cell Metabolism 34, 106–124, January 4, 2022 113
ll
OPEN ACCESS Articlenaive OT-I T cells for increased proliferation, with IFNg elevated
upon Adss knockdown (Figures 4E, 4F, S4D, and S4E). Consis-
tent with fatty acid carbon entering the IMP–S-AMP–AMP cycle
(Cader et al., 2020; Figure S3I), silencing the rate-limiting enzyme
of FAO, Cpt1a, in Famin+/+ DCs elevated their OT-I priming po-
tency to levels observed in Famin/ cells, while not augmenting
it further in the latter (Figure S4F). Hence, knockdown of IMP–S-
AMP–AMP cycle enzymes in FAMIN-sufficient DCs, or blocking
the upstream supply of carbon entering the cycle, phenocopied
enhanced priming due to compromised FAMIN.
Membrane trafficking, including endocytosis and processes
leading to antigen presentation, is controlled by proteins regu-
lated byGTP/GDP binding (Kirschner andMitchison, 1986; Sten-
mark, 2009). Guanine nucleotides are synthesized from IMP via
xanthosine monophosphate (XMP), catalyzed by IMP dehydro-
genase (IMPDH) and GMP synthase (GMPS; Figure 3G). GMP
reductase (GMPR) converts GMP back to IMP (Figure 3G) (Hed-
strom, 2012), which, together with the IMP–S-AMP–AMP cycle,
interconverts adenine and guanine nucleotides. Cellular purine
nucleotide levels all declined with compromised FAMIN activity,
except XMP, which trended higher (Figure S4G). XMP also
increased, and GMP and AMP decreased, upon AMPD inhibition
in Faminp.254I cDC1s (Figure 4G; Table S5). This suggested
increased flux through IMPDH consequent to impaired FAMIN
or IMP–S-AMP–AMP blockade. The IMPDH inhibitor mycophe-
nolate (MPA), an immunosuppressant effective in arthritis and
transplantation (Broen and van Laar, 2020), reduced AF647-
OVA uptake in Faminp.284R and Faminp.254V splenic DCs (Fig-
ure 4H), as did the GMPS inhibitor psicofuranine (Udaka and
Moyed, 1963) (Figure 4I), suggesting increased flux through
GMPS, too. Increased flux through GMPS was evident by
[15N1] GMP labeling increasing from Famin
p.254I to Faminp.284R
DCs after a pulse with [13C5
15N2] glutamine (Figure 4J), from
which the amide nitrogen is transferred to form GMP (Figures
3G and S4H) (Tesmer et al., 1996). No GMP isotopomers with
[15N2] or higher were observed, which would have indicated de
novo purine nucleotide synthesis (Figure S4H). We therefore
tested whether increased flux through IMPDH and GMPS medi-
ated enhanced priming. OVA-pulsed splenic DCs with Impdh1/
Impdh2 knockdown (Figures S4D and S4E) primed naive OT-I
T cells for lower IFNg secretion compared to control-silenced
DCs (Figures 4K–4M), and lower proliferation, which declined
from the different levels primed by Faminp.254I and Faminp.254V
DCs (Figures 4N–4P). OT-I T cell IFNg secretion and proliferation
also trended lower when primed by Gmps-silenced splenic DCs(E and F) Proliferation indices, CFSE overlays (E), and IFNg released (F) from OT-I
48 h after nucleofection with ctrl or Adsl, Adss, and Ampd2/Ampd3 siRNAs (n =
(G) Heatmap of purine nucleotide levels in Faminp.254I cDC1s, treatedwith Cpd3 or
Table S5.
(H and I) Percentage AF647-OVA+ splenic DCs of indicated genotypes after inc
independent experiments in H; n = 6, 3 mice per genotype in I).
(J) Fractional incorporation into indicatedGMP isotopomers following a 3 h pulse o
7/6/6, 3 mice per genotype).
(K–M) IFNg released from OT-I T cells co-cultured for 72 h with ovalbumin-puls
nucleofection with Gmpr2, Gmps, Impdh1/Impdh2, or ctrl siRNAs (n = 6, 3 mice
(N–P) Proliferation indices from OT-I T cells co-cultured for 72 h with Faminp.254I (N
after nucleofection with ctrl or Gmpr2, Gmps, Impdh1, and Impdh2 siRNA (n = 6
Data represented as mean ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001 (one-wa
Figure S4.
114 Cell Metabolism 34, 106–124, January 4, 2022(Figures 4K–4P). This demonstrated that increased flux via
IMPDH and GMPS in FAMIN-impaired DCs was responsible
for their increased antigen uptake and T cell priming.
IMPDH-dependent NADH/NAD+ redox state controls the
pace of antigen uptake and MHC I recycling
Bypassing IMPDH and GMPSwith exogenous guanine increased
AF647-OVA uptake in Faminp.254I and Faminp.254V splenic DCs,
phenocopying enhanced antigen uptake of Faminp.284R DCs (Fig-
ure 5A). In the latter, guanine did not further augment uptake (Fig-
ure 5A). Prima vista, this suggested that guanine nucleotide pools
may control antigen uptake. Increased antigen uptake, however,
was at odds with decreased GTP and GDP levels in FAMIN-
impaired DCs (Figure 3B). This implied a byproduct of the IMP–
XMP–GMP cycle, rather than guanine nucleotide pool size, may
be responsible for increased membrane trafficking. IMPDH re-
duces NAD+ to NADH + H+ (Figure 3G). Inhibition of IMPDH
rescued cytoplasmic acidification in Faminp.284R DCs (Figure 5B),
adding further evidence for enhanced flux through IMPDH. An
altered pHc can affect vesicular trafficking (Heuser, 1989; Korol-
chuk et al., 2011; Walton et al., 2018). ADSS inhibition in
Faminp.254I DCs enhanced AF647-OVA uptake (Figure 4A) without
causing cytoplasmic acidification (Figure 5C), arguing against pHc
changes accounting for altered membrane trafficking. IMPDH in-
hibition did not rescue OCR or ECAR deficits in Faminp.284R DCs
(Figures 5D and 5E), confirming that compromised OXPHOS
and glycolysis are not directly responsible for exaggerated anti-
gen uptake. As total cellular NAD(H) integrates protein-bound
and free forms across cytoplasmic and mitochondrial pools with
their distinct redox states, we measured the secreted lactate/py-
ruvate ratio to deduce the cytosolic free NADH/NAD+ ratio
(Goodman et al., 2020; Krebs, 1967; Williamson et al., 1967). Fa-
minp.284R splenic DCs exhibited a markedly higher lactate/pyru-
vate ratio than Faminp.254I cells (Figure 5F), implying increased
cytosolic NADH/NAD+. In contrast, the secreted b-hydroxybuty-
rate/acetoacetate ratio, reflecting mitochondrial free NADH/
NAD+, remained unchanged (Figure S5A). To investigate whether
re-oxidation of cytoplasmic NADH rescues exaggerated antigen
uptake, we provided pyruvate as external e acceptor that regen-
erates NAD+ via lactate dehydrogenase (LDH) (Figure 5G).
Pyruvate indeed rescued increased AF647-OVA uptake in
Faminp.284R splenic DCs (Figure 5H). Four-carbon a-ketobutyrate
(AKB) is an alternative substrate for regenerating NAD+ from
NADH via LDH (Figure 5G) (Sullivan et al., 2015). AKB is primarily
used as e acceptor and not as carbon substrate in otherT cells co-cultured for 72 h with Faminp.254I splenic DCs pulsed with ovalbumin
6, 3 mice per siRNA).
vehicle for 18 h (n = 5); normalized peak integrals and fold changes are shown in
ubation with mycophenolic acid (MPA) (H) or psicofuranine (I) (n = 18, from 3
f Faminp.254I, Faminp.254V, and Faminp.284R cDC1swith [15N2
13C5] glutamine (n =
ed Faminp.254I (K), Faminp.254V (L), or Faminp.284R (M) splenic DCs 48 h after
per group).
), Faminp.254V (O), and Faminp.284R (P) splenic DCs pulsed with ovalbumin 48 h
, 3 mice per siRNA).
y ANOVA or unpaired two-tailed Student’s t test where appropriate). See also
NAD+
H
A
F6
47
-O
V
A
+  (
%
)
A
20
30
40
50
***
**
Splenic DCs
guanine–
12
14
16
18
20
22
24
A
F6
47
-O
V
A
+  (
%
)
Faminp.254I
Faminp.284R
pyruvate AKB–
Splenic DCs
I
15
20
25
30
35
pyruvate AKB–
Faminp.254I
Faminp.254I + Cpd3
AMPD
A
F6
47
-O
V
A
+  (
%
)
Splenic DCs
***
***
** ***
***
Faminp.254V
Faminp.254I
Faminp.284R
4000
4500
5000
5500
6000
6500
C
yt
op
la
sm
ic
pH
ro
do
( R
FU
)
Faminp.254V
Faminp.254I
Faminp.284R
IMPDH
MPA–
B
**
**
***
BM-derived cDC1
20000
30000
40000
50000 ***
**
***
*
C
yt
op
la
sm
i c
pH
ro
do
(R
FU
)
Faminp.254V
Faminp.254I
Faminp.284R
ADSS
L-alanosine–
C
Splenic DCs
D
0
20
40
60
80
*** **
IMPDH
MPA–
Faminp.254I
Faminp.284R
E
C
A
R
(m
pH
/ m
in
)
27 62 88 113
0
20
40
60
80
100
120
140
Time (min)
O
C
R
(p
m
ol
es
/m
in
)
BM-derived cDC1
Faminp.254I
Faminp.284R
Faminp.254I + MPA
Faminp.284R + MPA
E
BM-derived cDC1
F
NADH
NAD+
α-hydroxybutyrate
Pyruvate
Lactate
α-ketobutyrate
+ AKB
Faminp.284R
Faminp.284R
J K
BM-derived cDC1
0 10 20 30
0
10
20
30
40
Time (min)
*
BM-derived cDC1
0 10 20 30
0
10
20
30
40
50
Time (min)
M
H
C
Ir
ec
yc
le
d
(%
)
*
M
H
C
Ir
ec
yc
le
d
(%
)
Oligo FCCP Rot + ant
1.0
1.2
1.4
1.6
1.8
2.0
La
ct
at
e/
 p
yr
uv
at
e 
ra
tio
**
G
Faminp.254I
Faminp.284R
Splenic DC
supernatants
Faminp.254I
Faminp.284R
Glucose
NADH
Pyruvate
NADH
NAD+
LDH LDH
Figure 5. IMPDH-dependent NADH/NAD+ redox state controls the pace of antigen uptake and MHC I recycling
(A) Percentage AF647-OVA+ splenic DCs of indicated genotypes following incubation with guanine (n = 6, 3 mice per genotype).
(B) pHc of Famin
p.254I, Faminp.254V, and Faminp.284R cDC1s in the presence of MPA (n = 8–12, 3 mice per genotype).
(C) pHc of Famin
p.254I, Faminp.254V, and Faminp.284R splenic DCs in the presence of L-alanosine (n = 9, 3 mice per genotype).
(D) OCR of Faminp.254I and Faminp.284R BM-derived cDC1s in the presence of MPA or vehicle control. Basal OCR and OCR following addition of oligomycin A
(Oligo), FCCP, and Rot + Ant (n = 7–14, 3 mice per genotype).
(E) Basal ECAR of Faminp.254I and Faminp.284R BM-derived cDC1in the presence of MPA or vehicle control (n = 16–18, 3 mice per genotype).
(F) Ratio of secreted lactate to pyruvate, reflective of free cytosolic NADH/NAD+, in supernatants of Faminp.254I and Faminp.284R splenic DCsmatured overnight in
RPMI-1640/10% FBS (n = 9, from 3 mice per genotype).
(G) Schematic depicting the lactate dehydrogenase (LDH) reaction, in which pyruvate is converted to lactate with regeneration of NAD+; a-ketobutyrate acts as an
alternative electron acceptor and is converted to a-hydroxybutyrate.
(H) Percentage AF647-OVA+ splenic DCs of indicated genotypes following incubation with pyruvate or a-ketobutyrate (AKB) overnight and replenished for the
time of the assay (n = 6, 3 mice per genotype).
(I) Percentage AF647-OVA+ Faminp.254I splenic DCs following overnight incubation with pyruvate or AKB alone or in presence of Cpd3 (n = 6, 3 mice per
genotype).
(J) Percentage of MHC I recycled in Faminp.254I and Faminp.284R BM-derived cDC1s over time (n = 6, 3 mice per genotype).
(K) Percentage of MHC I recycled in Faminp.284R BM-derived cDC1s at indicated times following overnight incubation with AKB or control (n = 3–5, 3 mice per
genotype).
Data represented as mean ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001 (one-way ANOVA or unpaired two-tailed Student’s t test where appropriate). See also
Figure S5.
ll
OPEN ACCESSArticle
Cell Metabolism 34, 106–124, January 4, 2022 115
A0
10
20
30
40
Famin+/+
Famin–/–
IF
N
γ 
(n
g/
m
L)
**
B
Enrichment plot: Oxidative phosphorylation
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
E
nr
ic
hm
en
ts
co
re
(E
S
)
Rank in ordered dataset
0 2000 4000 6000 8000 10000
-0.4
-0.2
0.0
0.2
0.4
0.6
zero cross at 4628
(S
ig
na
l2
N
oi
se
)
R
an
ke
d
lis
tm
et
ric
E
nr
ic
hm
en
ts
co
re
(E
S
)
Enrichment plot: Hallmark Myc Targets V2
Rank in ordered dataset
C
0 2000 4000 6000 8000 10000
0.0
0.1
0.2
0.3
0.4
0.5
-0.4
-0.2
0.0
0.2
0.4
0.6
zero cross at 4628
0
20
40
60
80
Famin+/+
Famin–/–
***
IF
N
γ 
(n
g/
m
L)
D 3 kDa filtered 
supernatant
0
10
20
30
40
IF
N
γ 
(n
g/
m
L)
*
*
Faminp.254V
Faminp.254I
Faminp.284R
1.2
1.4
1.6
1.8
2.0
P
ro
lif
er
at
io
n 
in
de
x
**
***
Faminp.254V
Faminp.254I
Faminp.284R
0-103 103 104 105
CFSE
0
20
40
60
80
100
N
or
m
al
iz
ed
 T
o 
M
od
e
E CD8+ OT-I T cells
0
2
4
6
8
10 ***
***
IF
N
γ 
(n
g/
m
L)
Faminp.254V
Faminp.254I
Faminp.284R
F CD4+ OT-II T cells
Log2 fold change
p-
va
lu
e
-1 0 1
1×10-4
0.001
0.01
0.1
1.0
Inosine
Propionylcarnitine
Dihydrothymine Deoxycytidine
Decreased levels 
in Famin–/– versus Famin+/+
Increased levels 
in Famin–/– versus Famin+/+
G
0
1×105
2×105
3×105
4×105
5×105
In
os
in
e 
(p
ea
k 
ar
ea
)
Famin+/+
Famin–/–
OptiMEM
H
**
Log2 fold change
FD
R
-a
dj
us
te
d 
p-
va
lu
e
-1 0 1
1×10-9
1×10-8
1×10-7
1×10-6
1×10-5
1×10-4
0.001
0.01
0.1
1.0
Creatinine
Carnitine
CreatinePropionylcarnitine
Cytidine
Inosine
Cytosine
Ornithine
Butyrylcarnitine
Lactate
Decreased levels 
in Faminp.284R versus Faminp.254I
Increased levels 
in Faminp.284R versus Faminp.254I
Inosine standard
Famin+/+
Famin–/–
I
0
1×105
2×105
3×105
4×105
***
***
In
os
in
e 
(p
ea
k 
ar
ea
)
Faminp.254V
Faminp.254I
Faminp.284R
OptiMEM
J
1.30
1.35
1.40
1.45
1.50
P
ro
lif
er
at
io
n 
in
de
x
0
5
10
15
Inosine, nM
IF
N
γ 
(n
g/
m
L)
L M
0.1 1 10 1000
Inosine, nM
0.1 1 10 1000
N
0.4
0.6
0.8
1.0
1.2
D
iv
is
io
n 
in
de
x
*
***
*
Spiked-in-inosine, nM
0.1 1 2.5 50
Faminp.254I
Faminp.284R
n.s.
O
0
5
10
15
20
0.1 1 2.5 5
Faminp.254I
Faminp.284R
IF
N
γ 
(n
g/
m
L)
0
**
***
n.s.
CGS21680
0
200
400
600
800
*
*IF
N
γ 
(n
g/
m
L)
Faminp.254V
Faminp.254I
Faminp.284R
***
**
1.5
2.0
2.5
3.0
P
ro
lif
er
at
io
n 
in
de
x
**
***
P
Faminp.254V
Faminp.254I
Faminp.284R
SCH58261– CGS21680 SCH58261–
K
Q
Spiked-in-inosine, nM
(S
ig
na
l2
N
oi
se
)
R
an
ke
d
lis
tm
et
ric
0 2 4 6
0
50
100
P
ea
k 
in
te
ns
ity
Time (min)
Figure 6. The FAMIN catalytic product inosine released from DCs dampens T cell activation during priming
(A) IFNg secretion from naive OT-I T cells activated by anti-CD3/CD28 in presence of 24 h supernatants of Famin/ or Famin+/+ CD11c+ splenic DCs (n = 3).
(B and C) Gene set enrichment analysis (GSEA) of RNA-seq dataset of naive OT-I T cells activated by anti-CD3/CD28 for 24 h in the presence of Famin/ or
Famin+/+ splenic DC supernatant; data depict enrichment of the Hallmark gene sets ‘‘oxidative phosphorylation’’ (B) and ‘‘Myc Targets V2’’ (C).
(legend continued on next page)
ll
OPEN ACCESS Article
116 Cell Metabolism 34, 106–124, January 4, 2022
ll
OPEN ACCESSArticlemetabolic pathways. AKB precisely phenocopied the rescue of
antigen uptake achieved by pyruvate, decreasing AF647-OVA up-
take to equally low levels in Faminp.254I and Faminp.284R splenic
DCs (Figure 5H). AKB reduced the secreted lactate/pyruvate ratio
as expected (Figure S5B), while the pHc difference was retained
between Faminp.284R and Faminp.254I splenic DCs (Figure S5C),
consistent with cytoplasmic acidification not accounting for
increased antigen uptake. Augmentation of baseline OCR was
markedly different between pyruvate and AKB (Figures S5D and
S5E). This was consistent with only pyruvate entering mitochon-
drial oxidation (Figure 5G), which we confirmed by tracing [13C3]
pyruvate and [13C4] AKB into Krebs cycle metabolites in Fa-
minp.254I and Faminp.284R cDC1s (Figures S5F–S5I). Importantly,
pyruvate and AKB also rescued increased AF647-OVA uptake in
Faminp.254I splenic DCs, in which the IMP–S-AMP–AMP cycle
was halted by Cpd3 (Figure 5I). AF647-OVA uptake serves as
proxy of one specialized endocytic pathway, but membrane traf-
ficking is involved in all the different routes of antigen uptake, pro-
cessing, and (cross-)presentation (Alloatti et al., 2016; Blander,
2018). Considering whether FAMIN-mediated redox control of
membrane trafficking is a more general principle, we asked
whether FAMIN affects MHC I recycling, required for loading of
cross-presented peptides (Belabed et al., 2020; Joffre et al.,
2012).MHC I recycling was higher in Faminp.284R compared to Fa-
minp.254I cDC1 (Figure 5J). AKB markedly reduced MHC I recy-
cling in Faminp.284R, and barely in Faminp.254I cDC1 (Figures 5K
and S5J), revealing that cytoplasmic NADH/NAD+ affects the
pace of MHC I recycling, too. This provided strong evidence
that enhanced antigen uptake and presentation in FAMIN-
impaired DCs is caused by increased reduction of cytoplasmic
NAD+ to NADH by IMPDH due to an imbalance in adenine-gua-
nine nucleotide interconversion cycles.
The FAMIN product inosine dampens T cell activation
during priming
Fixing splenic DCs with glutaraldehyde after pulsing with
OVA257-264 retained the ability of Faminp.284R compared to(D) IFNg secretion from OT-I T cells activated by anti-CD3/CD28 for 72 h in the p
splenic DCs cultured for 3 h in RPMI-1640/10% FBS (n = 3).
(E) IFNg secretion, proliferation indices, and CFSE overlays from OT-I T cells a
Faminp.254I, Faminp.254V, and Faminp.284R splenic DCs cultured for 3 h in OptiME
(F) IFNg from 72 h anti-CD3/CD28 stimulated naive OT-II T cells cultured in the pre
DCs (n = 6, 3 mice per genotype).
(G) Differential abundance of LC-MS features in supernatants from Famin/ ve
genotype). Data depicted as volcano plot showing p value and log2 fold change
(H) Representative extracted chromatograms, using normalized peak intensity, s
DCs and corresponding standard.
(I) Differential LC-MS features in supernatants from Faminp.254I and Faminp.284
Data depicted as volcano plot showing p value and log2 fold change for each de
(J and K) Relative inosine levels released in supernatants from Famin/ or Famin
3 h in OptiMEM (n = 8–9, 3 mice per genotype).
(L andM) IFNg secretion (L) and proliferation indices (M) of naiveOT-I T cells stimul
for 72 h (n = 3).
(N and O) Division index (N) and IFNg secretion (O) from Faminp.254I and Faminp
presence of indicated spiked-in concentrations of inosine into RPMI-1640 (conc
inosine release between Faminp.254I and Faminp.284R DC supernatants, such as i
(P and Q) IFNg secretion (P) and proliferation indices (Q) of OT-I T cells activated
supernatants from Faminp.254I, Faminp.254V, and Faminp.284R splenic DCs culture
Data represented as mean ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001 (one-wa
Figure S6.Faminp.254I DCs to prime for increased OT-I T cell proliferation
(Figure S6A), but the capacity to prime for enhanced IFNg secre-
tion by Faminp.284R DCs, however, was lost (Figure S6B). This
indicated that optimal priming requires mutual dynamic trans-
membrane signaling. It also raised the possibility that soluble
factors released from DCs might be involved, too. Famin/
and Famin+/+ DCs, and Faminp.254I, Faminp.254V, and Faminp.284R
cDC1s, were indistinguishable in their expression of co-stimula-
tory and co-inhibitorymolecules (Table S4; Figure S6C). Cell-free
supernatants of Famin/ DCs primed naive anti-CD3/CD28-
activated OT-I T cells to secrete 2-fold more IFNg compared to
supernatants of Famin+/+ DCs (Figure 6A). Transcriptomes of
OT-I T cells activated in the presence of Famin/ compared
to Famin+/+ DC supernatants were enriched for hallmark gene
sets (Leone et al., 2019) indicative of elevated effector function
(Figures 6B, 6C, S6D, and S6E; Table S6). IL-12p70 and IFNa
secretion by Famin/ and Famin+/+ DCs in co-culture with naive
OT-I T cells was indistinguishable (Figure S6F). Freeze-thaw cy-
cles did not affect Famin/ DC supernatants’ enhanced stimu-
latory capacity (data not shown), which was retained after pass-
ing through a 3 kDa filter (Figure 6D), pointing to a small
molecule. To enable its identification, we switched to serum-
free OptiMEM media to reduce complexity. OptiMEM superna-
tants of Famin/ and Famin+/+ DCs retained differences in
IFNg induction in anti-CD3/CD28-activated CD8+ T cells (Fig-
ure S6G). They were particularly stark between those elicited
by Faminp.254I compared to Faminp.254V DC supernatants (Fig-
ure 6E). CD4+ OT-II T cells were also primed for heightened
IFNg secretion by Faminp.254V and Faminp.284R compared to
Faminp.254I DC supernatants (Figure 6F). This suggested that
DCs secrete a small molecule in a FAMIN-dependent manner
that inhibits priming of naive CD4+ and CD8+ T cells.
Unbiased high-resolution LC-MS of supernatants of Famin+/+
and Famin/ splenic DCs resolved 
1,100 features, revealing
inosine as the top-ranking identifiable LC-MS feature of
differential abundance (Figures 6G and 6H). A second unbiased
LC-MS screen comparing Faminp.254I with Faminp.284R splenicresence of <3 kDa cut-off filtrates of supernatants from Famin+/+ or Famin/
ctivated with anti-CD3/CD28 for 72 h in the presence of supernatants from
M (n = 3).
sence of 3 h supernatant from Faminp.254I, Faminp.254V, and Faminp.284R splenic
rsus Famin+/+ splenic DCs cultured for 3 h in OptiMEM (n = 8–9, 3 mice per
for each detected LC-MS feature.
howing inosine detection in supernatants from Famin/ and Famin+/+ splenic
R splenic DCs cultured for 3 h in OptiMEM (n = 8–9, 3 mice per genotype).
tected LC-MS feature; FDR-adjusted p value is depicted.
+/+ (J) or Faminp.254I, Faminp.254V, and Faminp.284R (K) splenic DCs, cultured for
atedwith anti-CD3/CD28 in the presence of indicated concentrations of inosine
.284R splenic DCs pulsed with ovalbumin and co-cultured with OT-I T cells in
entration from 0.1 to 5 nM, with 5 nM reflecting the lower end of differences in
n Figure S6H) (n = 3–6, from 3 mice per genotype).
by anti-CD3/CD28 for 72 h in the presence of CGS21680 or SCH58261 and
d for 3 h in OptiMEM (n = 6, 3 mice per genotype).
y ANOVA or unpaired two-tailed Student’s t test where appropriate). See also
Cell Metabolism 34, 106–124, January 4, 2022 117
A0
1×105
2×105
3×105
4×105 *
13C5 
15N4 inosine
P
ea
k 
ar
ea
B Cells
0
2×105
4×105
6×105
***
***
13C5 
15N4 inosine
P
ea
k 
ar
ea
16
18
20
22
24 ***
***
13C5 
15N4 inosine
Fr
ac
tio
na
l i
nc
or
po
ra
tio
n 
(%
 o
f t
ot
al
)
Supernatants
Faminp.254V
Faminp.254I
Faminp.284R
Faminp.254V
Faminp.254I
Faminp.284R
5×105
1×106
1.5×106
2×106
2.5×106 **
*
unlabelled inosine
P
ea
k 
ar
ea
D Supernatants
Faminp.254V
Faminp.254I
Faminp.284R
E
3.0 3.2 3.4 3.6 3.8 4.0
0
20000
40000
60000
80000
100000
RPMI-1640 + 10% FBS
0.1 nM
5 nM
10 nM
P
ea
k 
in
te
ns
ity
Retention time (minutes)
F
0
5×105
1×106
In
os
in
e 
(p
ea
k 
ar
ea
)
Faminp.254V
Faminp.254I
Faminp.284R
IMPDH
MPA–
2×105
3×105
4×105
5×105
6×105
In
os
in
e 
(p
ea
k 
ar
ea
)
Faminp.254V
Faminp.254I
Faminp.284R
***
ADSS
Hadacidin–
G H
0
5×105
1×106
1.5×106
2×106
In
os
in
e 
(p
ea
k 
ar
ea
)
Faminp.254V
Faminp.254I
Faminp.284R
**
**
AMPD
Cpd3–
0
200
400
600
12 14 17 19 21 23 26
Days
Faminp.254V
Faminp.254I
Faminp.284R
Tu
m
ou
r v
ol
um
e 
(m
m
3 )
0
200
400
600
800 *
Tu
m
ou
r v
ol
um
e 
(m
m
3 )
Faminp.254V
Faminp.254I
Faminp.284R
Day 26
***
***
I J
Tu
m
ou
r d
ia
m
et
er
 (m
m
)
14 16 18 20 23 25 27 30
0
2
4
6
8
Days
Faminp.254V
Faminp.284R
0
5
10
15
20
*
Day 30
Faminp.254V
Faminp.284R
Tu
m
ou
r d
ia
m
et
er
 (m
m
)
***
O
O
OHHO
HO
N
N
NH
N
O
OHHO
HO
OPO൤
2–
+ P OHHO
O
O–
+
FAMIN
PNP
HypoxanthineRibose-1-phosphate Inosine Phosphate
Faminp.254V
Faminp.254I
Faminp.284R
***
***
*** **
C
O
N
HN
NH
N
Figure 7. FAMIN-dependent conversion of extracellular hypoxanthine into inosine
(A) FAMIN-catalyzed enzymatic conversion of [13C5
15N4] hypoxanthine to [
13C5
15N4] inosine.
(B) Cellular [13C5
15N4] inosine in Famin
p.254I, Faminp.254V, and Faminp.284R splenic DCs pre-equilibrated in OptiMEM for 3 h before a 3 h pulse with [13C5
15N4]
hypoxanthine in OptiMEM (n = 6, 3 mice per genotype).
(legend continued on next page)
ll
OPEN ACCESS Article
118 Cell Metabolism 34, 106–124, January 4, 2022
ll
OPEN ACCESSArticleDC supernatants identified two metabolites, inosine and pro-
pionyl-carnitine, overlapping with the first screen (Figure 6I). As
a catalytic product of FAMIN, inosine, whose levels were highest
with fully active FAMIN-254I (Figures 6J and 6K), was a plausible
candidate. Pure inosine dose-dependently reduced OT-I T cell
IFNg secretion and proliferation after anti-CD3/CD28 stimulation
(Figures 6L and 6M). Spiking co-cultures of OVA-presenting
splenic DCs and naive OT-I T cells with just 5 nM inosine, the
lower end of differentials in inosine levels between Faminp.254I
and Faminp.284R DC supernatants (Figure S6H), was sufficient
to decrease T cell proliferation and IFNg secretion (Figures 6N
and 6O). This suggested inosine as an inhibitory signal during
T cell priming. Among the four adenosine receptors, direct bind-
ing of, and activation by, inosine has been shown for A2AR
(Welihinda et al., 2018), the only adenosine receptor expressed
on T cells (Cekic et al., 2013). In OT-I T cells activated by anti-
CD3/CD28 in the presence of DC supernatants, the A2AR agonist
CGS21680 (Hutchison et al., 1989) inhibited proliferation and
IFNg secretion and abrogated Famin genotype-related differ-
ences (Figures 6P and 6Q). A2AR antagonism with SCH58261
(Zocchi et al., 1996) had the converse effect (Figures 6P and
6Q). This demonstrated that FAMIN-dependent release of
inosine from DCs inhibited activation of naive T cells.
FAMIN-dependent conversion of extracellular
hypoxanthine into inosine
Nucleobases and nucleosides equilibrate across the plasma
membrane via purine/pyrimidine transporters (Boswell-Casteel
and Hays, 2017), prompting us to consider whether external
nucleobases may supply the substrate for the synthesis of ino-
sine in DCs (Figure 7A). A 3 h pulse of splenic DCs with 25 mM
[13C5
15N4] hypoxanthine labeled over half of extracellular and
cellular hypoxanthine (Figures S7A–S7D). Labeled hypoxanthine
was converted to cellular [13C5
15N4] inosine, which was highest
in Faminp.254I and lowest in Faminp.284R DCs (Figures 7B and
S7E). This resulted in [13C5
15N4] inosine released into superna-
tants, whose levels were, in relative terms, 
35% and 
22%
higher in Faminp.254I and Faminp.254V, respectively, than in
Faminp.284R DCs (Figure 7C). Even fractional [13C5
15N4] inosine
labeling increased from Faminp.284R to Faminp.254V and
Faminp.254I supernatants, averaging at one-fifth of total (Fig-
ure 7C), despite unlabeled inosine levels increased alongside,
too (Figure 7D). Since OptiMEM is inosine-free and levels in
RPMI-1640/10% FBS negligible (<100 pM, below the lowest
detectable standard; Figure 7E), media change for labeling
studies prompt an immediate equilibrative efflux; hence, these(C) [13C5
15N4] inosine, and its fractional incorporation, in supernatants of Famin
p.
3 h before a 3 h pulse with [13C5
15N4] hypoxanthine in OptiMEM (n = 18, 3 mice
(D) Unlabeled (M+0) inosine in supernatants of Faminp.254I, Faminp.254V, and Fa
OptiMEM (n = 18, 3 mice per genotype).
(E) Representative extracted chromatograms, showing peak corresponding to in
inosine was below the quantification limit.
(F–H) Inosine levels released into supernatants from Faminp.254I, Faminp.254V, and F
18 h and cultured for 3 h in OptiMEM with inhibitors replenished in medium (n =
(I) Tumor volume over time, and on day 26, after subcutaneous inoculation of Fami
7, three mice were sacrificed due to fight wounds on day 23 during the blinded p
(J) Tumor diameter over time and on day 30 post inoculation of Faminp.254V and Fa
Data represented as mean ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001 (one-wa
Figure S7.marked differences likely underestimate the contribution of
FAMIN to inosine release during priming in situ. Plasma levels
of inosine were similar across Famin germline mutants and
DC-selective deletion (Figures S7F and S7G), consistent with a
model of localized release. Perturbation in adenine-guanine
nucleotide interconversion revealed increased inosine release
upon IMPDH inhibition and a slight decrease upon ADSS and
AMPD inhibition, with differential release across Famin geno-
types remaining intact (Figures 7F–7H). Altogether these studies
demonstrated that elevated inosine release by DCs with active
FAMIN amplifies an inhibitory signal during T cell priming, gener-
ated by phosphoribosylation of largely extracellularly derived
hypoxanthine.
Compromised FAMIN catalysis enhances tumor immune
surveillance
We finally turned to tumor immunosurveillance, a model not
confounded by increased viral replication associated with
IMPDH activity (To et al., 2016), to assess endogenous CTL func-
tion primed in the setting of polymorphic Famin variants.
Increased DC antigen presentation enhances tumor-specific im-
munity by inducing Th1 and CTL responses (Xia et al., 2018).
A2AR signaling prevents T cell anti-tumor immunity by
inhibiting CTL activation and maintaining naive T cells quiescent
(Cekic et al., 2013; Ohta et al., 2006). Faminp.254I and Faminp.254V
mice developed markedly larger tumors compared to
Faminp.284R mice (Figures 7I and 7J) when subcutaneously in-
jected with a syngeneic Lewis lung carcinoma cell line express-
ing ovalbumin (LL/2-OVA) (Kraman et al., 2010). Protection was
associated with nominally higher OVA257-264-specific CD8+
T cell numbers in peripheral blood in Faminp.284R compared to
Faminp.254I and Faminp.254V mice (Figure S7H). These results
were consistent with augmented priming that translated into
increased CTL activity and tumor immunosurveillance when
FAMIN activity is compromised.
DISCUSSION
Here we discovered that purine nucleotide and nucleoside turn-
over in DCs represses T cell immunity, with FAMIN acting as a
purely biochemical immune checkpoint. FAMIN achieves this
via two main routes. First, FAMIN restrains endocytosis, antigen
processing, and presentation via cytoplasmic NADH/NAD+
through balancing adenine-guanine nucleotide interconversion.
Second, it amplifies an inhibitory signal through the generation
of locally released inosine. The relative contribution of altered254I, Faminp.254V, and Faminp.284R splenic DCs pre-equilibrated in OptiMEM for
per genotype).
minp.284R splenic DCs following a 3 h pulse with [13C5
15N4] hypoxanthine in
osine for indicated standards and for RPMI-1640/10% FBS medium, in which
aminp.284R splenic DCs, pre-treated withMPA (F), hadacidin (G), or Cpd3 (H) for
9, 3 mice per genotype).
np.254I, Faminp.254V, and Faminp.284Rmice with 2.53 104 LL2-OVA cells (n = 6/7/
hase of the experiment; genotype was post hoc identified as Faminp.284R).
minp.284R mice subcutaneously injected with 23 104 LL2-OVA cells (n = 10/8).
y ANOVA or unpaired two-tailed Student’s t test where appropriate). See also
Cell Metabolism 34, 106–124, January 4, 2022 119
ll
OPEN ACCESS Articleantigen presentation versus altered inosine release to T cell
priming is impossible to disentangle, since both are directly cata-
lytically controlled by FAMIN. ADA, PNP, and MTAP do not
compensate for FAMIN’s absence, although they share three
of four of FAMIN’s catalytic activities. This suggests that FAMIN
is at the center of a dedicated purine metabolon that biochemi-
cally restrains DCs’ priming activity.
Pacing membrane trafficking via adenine-guanine nucleotide
interconversion cycles through an NADH/NAD+-sensitive
mechanism, consequent to hyperactive IMPDH that reduces
NAD+ to NADH, may represent a general principle. Vectorial
physical membrane displacements, which can be energized
by transmembrane e transport, occur upon plasma mem-
brane internalization and, in the opposite direction, during
membrane budding and vesicle formation (Morre´ and Morre´,
2011). NADH can activate vectorial membrane transfer,
elegantly demonstrated in cell-free systems for the transfer
from the transGolgi apparatus to the plasmamembrane (Rodri-
guez et al., 1992). Aside from IMPDH, NAD+ is reduced to
NADH by several cytoplasmic reactions, foremost by glyceral-
dehyde dehydrogenase of glycolysis, and re-oxidized by LDH
and the malate-aspartate shuttle via mitochondria. NADH/
NAD+ reductive stress due to perturbation in any of those reac-
tions might therefore also impact membrane trafficking and an-
tigen presentation. NADH/NAD+ reductive stress in the liver
emerges as the causal mechanism for features of the metabolic
syndrome associated with hypomorphicGCKR, such as hepat-
ic insulin resistance and increased triglyceride release
(Goodman et al., 2020). GCKR localizes at the probably most
pleiotropic genome-wide association study (GWAS) locus and
encodes liver-specific glucokinase regulatory protein, which
helps prevent a futile metabolic cycle with glycolysis during
gluconeogenesis (Goodman et al., 2020). Whether obesity,
which increases risk for autoimmunity (Versini et al., 2014),
causes NADH/NAD+ reductive stress in DCs is unknown.
Via equilibration of purine nucleobases and nucleosides
across the plasma membrane, DCs may survey their vicinity.
They respond to hypoxanthine by converting it to inosine, damp-
ening T cell activation. Since phosphorolysis is reversible, DCs
might bidirectionally respond to local hypoxanthine and inosine
availability during immunological synapse formation. T cells, in
particular naive CD4+ T cells, can release hypoxanthine (Fan
et al., 2019), whichmay enable a dynamic interactionwith FAMIN
catalysis in DCs affecting the priming threshold. Systemic hypo-
xanthine and inosine levels closely track each other, consistent
with their rapid interconversion via PNP (Sun et al., 2019). The in-
testinal microbiota may also affect inosine plasma levels (Mager
et al., 2020). Inosine activation of the A2AR on T cells is itself com-
plex: inosine either prevented Th1 differentiation and blunted
anti-tumor immunity in anti-CTLA4-treated mice or, conversely,
enhanced both, when co-supplied with IFNg in vitro and a
TLR9 agonist in vivo (Mager et al., 2020). The mechanism under-
lying this switch remained unclear. Inosine can also serve as an
alternative carbon source for CD8+ T cells when glucose is un-
available (Wang et al., 2020). The complete (254I versus 254V)
and almost-complete (254I versus 284R and Famin+/+ versus
Famin/) absence of transcriptomic changes excludes that
autocrine inosine-triggered receptor signaling in DCs dampens
their priming capacity.120 Cell Metabolism 34, 106–124, January 4, 2022The biochemical mechanism reported herein sheds new light
on predisposition for HLH/MAS. Hyperactivated CTLs and their
IFNg activate macrophages, triggering hemophagocytosis and
a cytokine storm (Brisse et al., 2016b). Persistence of antigen-
presenting DCs resulting in uncontrolled CTL priming occurs in
primary, genetic HLH, consequent to impaired perforin-medi-
ated antigen-selective removal of DCs (Lykens et al., 2011).
However, most patients with secondary, acquired HLH have un-
impaired cytotoxicity (Bryceson et al., 2012). AcquiredHLH/MAS
complicates diverse, mostly viral, infections, malignancies, auto-
immune and autoinflammatory disorders (Brisse et al., 2016a),
and treatment with chimeric antigen receptor (CAR) T cells (Nee-
lapu et al., 2018). Five of six candidate genes at the major
3p21.31 risk locus for severe COVID-19 (Ellinghaus et al.,
2020; Nakanishi et al., 2021; Pairo-Castineira et al., 2021) point
toward DC-T cell interactions (Kaser, 2020), remarkable as that
hyperinflammation shares features with HLH/MAS, including
hemophagocytosis (Lucas et al., 2020; Prieto-Pe´rez et al.,
2020). It remains unclear why excessive T cell activation by
FAMIN-impaired DCs results in enhanced immunosurveillance
of tumors but fails to control IAV infection.
A final point is that experiments in wild-type mice may grossly
over-estimate anti-viral and anti-tumor T cell immunity that can
be expected in the majority of humans, since mice naturally ex-
press hypomorphic FAMIN-254V. As we show here, the single
amino acid change to 254I (for which 
94% of humans are
homozygous or heterozygous) results in 4-fold lower numbers
of nucleoprotein-specific CTLs upon experimental IAV infection
and a profound reduction in T cell effector function. This poly-
morphism (rs3764147) has been linked to a possible founder
effect (Rivas et al., 2018), hinting that excessive priming may
have afforded evolutionary benefits.
Limitations of study
The IMP–S-AMP–AMP cycle operates at the center of energy
metabolism, directly and instantaneously affecting glycolysis,
electron transfer, FAO, Krebs cycle activity, glutamine oxidation,
and the urea cycle (Cader et al., 2020; Lowenstein, 1972, 1990).
The lack of technology with temporo-spatial resolution to resolve
metabolites across cellular compartments is a particularly acute
limitation, compounded by the unparalleled degree of intercon-
nectedness, fast substrate cycles, and redundancies within cen-
tral purine metabolism. This poses challenges, e.g., for directly
measuring flux through IMPDH, and for determining inosine
levels at the immunological synapse in situ.
STAR+METHODS
Detailed methods are provided in the online version of this paper
and include the following:
d KEY RESOURCES TABLE
d RESOURCE AVAILABILITYB Lead contact
B Materials availability
B Data and code availability
d EXPERIMENTAL MODEL AND SUBJECT DETAILS
B Mice
B Dendritic cell and T cell isolation
ll
OPEN ACCESSArticled METHOD DETAILS
B Influenza infection
B In vitro T cell priming and restimulation
B Cytotoxicity assays
B In vivo T cell priming
B Flow cytometry
B Cytokine measurement by ELISA
B RNA extraction and sequencing
B Immunoblot
B Metabolic tracing experiments
B Extraction of aqueous metabolites
B LC-MS sample preparation
B LC-MS analysis of aqueous metabolites
B Hydrazone derivatization of keto acids and hydroxy
carboxylic acids in cell culture supernatants and sub-
sequent LC-MS analysis
B LC-MS data processing
B Determination of extracellular acidification rate and ox-
ygen consumption rate
B Cytoplasmic pH assay
B Antigen uptake assay
B Endosome-to-cytosol uptake assay
B MHC I recycling assay
B siRNA transfection
B T cell activation assays in presence of DC-secreted
soluble factor
B Tumor xenograft model
d QUANTIFICATION AND STATISTICAL ANALYSIS
SUPPLEMENTAL INFORMATION
Supplemental information can be found online at https://doi.org/10.1016/j.
cmet.2021.12.009.
ACKNOWLEDGMENTS
We thank Paul J. Lehner for discussions and Caetano Reis e Sousa, Ziad Mal-
lat, Ken G.C. Smith, Suzanne Turner, and Douglas T. Fearon for reagents. This
work was supported by European Research Council grants 648889 and
260961 (A.K.); the Wellcome Trust via senior investigator awards 106260/Z/
14/Z and 222497/Z/21/Z (A.K.), intermediate clinical fellowships 105920/Z/
14/Z (J.C.L.) and 216630/Z/19/Z (N.C.K.), career re-entry fellowship 103077/
Z/13/Z (N.C.K.), clinical research career development fellowship 222056/Z/
20/Z (M.Z.C.), and PhD studentship 102163/B/13/Z (K.R.); the Academy of
Medical Sciences starter grants SGL018\1119 (M.Z.C.) and SGL022\1008
(G.W.S.); the National Council for Scientific and Technological Development,
Brazil, PhD scholarship 201664/2014-3 (R.P.d.A.R.); the Austrian Science
Fund FWF J4396 (L.W.U.); the National Institutes of Health RO1 grants
DK051362 and DK088199 (R.S.B.); the Ministry of Higher Education Scholar-
ship Programme, Malaysia grant MARA330408277134 (M.N.M.-I.); and the
German Research Foundation DFG fellowship HA 7731/1-1 (L.M.H.). We
acknowledge support by the NIHR Cambridge BRC. We thank the NIHR Cam-
bridge BRC Cell Phenotyping Hub for flow cytometry support, the MRCMeta-
bolic Diseases Unit [MC_UU_00014/5] for assistance with histology samples,
and the NIH Tetramer Core Facility for providing tetramers.
AUTHOR CONTRIBUTIONS
S.S., G.W.S., and K.R. together with M.Z.C., L.-M.H., R.P.d.A.R., L.W.U.,
A.B.I.-R., M.N.M.-I., J.O.J., and N.C.K. performed experiments; J.A.W. pro-
vided ultra-high-performance liquid chromatography and mass spectrometry;
S.C., L.-M.H., T.D.L., and G.D. contributed to in vivo experimentation; R.S.B.
contributed critical reagents; L.M.H. provided experimental support; J.C.L.,
C.B., G.W.S., and S.S. performed RNA-seq; A.B. generated CRISPR/Cas9mice lines; A.K. devised the study and, together with S.S., G.W.S., and K.R.
and input from all authors, coordinated the project, designed experiments, in-
terpreted data, and prepared the manuscript.
DECLARATION OF INTERESTS
The University of Cambridge has filed patent applications relating to this work.
The authors declare no other competing financial interests.
Received: June 18, 2021
Revised: September 28, 2021
Accepted: December 8, 2021
Published: January 4, 2022
REFERENCES
Admyre, T., Amrot-Fors, L., Andersson, M., Bauer, M., Bjursell, M., Drmota, T.,
Hallen, S., Hartleib-Geschwindner, J., Lindmark, B., Liu, J., et al. (2014).
Inhibition of AMP deaminase activity does not improve glucose control in ro-
dent models of insulin resistance or diabetes. Chem. Biol. 21, 1486–1496.
Al-Mayouf, S.M., Almutairi, A., Albrawi, S., Fathalla, B.M., Alzyoud, R., AlEnazi,
A., Abu-Shukair, M., Alwahadneh, A., Alsonbul, A., Zlenti, M., et al.; for
Pediatric Arab Rheumatology Group (PRAG) (2020). Pattern and diagnostic
evaluation of systemic autoinflammatory diseases other than familial
Mediterranean fever among Arab children: a multicenter study from the
Pediatric Rheumatology Arab Group (PRAG). Rheumatol. Int. 40, 49–56.
Alloatti, A., Kotsias, F., Magalhaes, J.G., and Amigorena, S. (2016). Dendritic
cell maturation and cross-presentation: timing matters! Immunol. Rev.
272, 97–108.
Atkinson, M.R., Morton, R.K., and Murray, A.W. (1964). Inhibition of adenylo-
succinate synthetase and adenylosuccinate lyase from Ehrlich ascites-tumour
cells by 6-thioinosine 50-phosphate. Biochem. J. 92, 398–404.
Barnden, M.J., Allison, J., Heath, W.R., and Carbone, F.R. (1998). Defective
TCR expression in transgenic mice constructed using cDNA-based alpha-
and beta-chain genes under the control of heterologous regulatory elements.
Immunol. Cell Biol. 76, 34–40.
Barrett, J.C., Hansoul, S., Nicolae, D.L., Cho, J.H., Duerr, R.H., Rioux, J.D.,
Brant, S.R., Silverberg, M.S., Taylor, K.D., Barmada, M.M., et al.; NIDDK IBD
Genetics Consortium; Belgian-French IBD Consortium; Wellcome Trust
Case Control Consortium (2008). Genome-wide association defines more
than 30 distinct susceptibility loci for Crohn’s disease. Nat. Genet. 40,
955–962.
Belabed, M., Mauvais, F.X., Maschalidi, S., Kurowska, M., Goudin, N., Huang,
J.D., Fischer, A., de Saint Basile, G., van Endert, P., Sepulveda, F.E., and
Me´nasche´, G. (2020). Kinesin-1 regulates antigen cross-presentation through
the scission of tubulations from early endosomes in dendritic cells. Nat.
Commun. 11, 1817.
Beutel, G., Wiesner, O., Eder, M., Hafer, C., Schneider, A.S., Kielstein, J.T.,
K€uhn, C., Heim, A., Ganzenm€uller, T., Kreipe, H.H., et al. (2011). Virus-associ-
ated hemophagocytic syndrome as a major contributor to death in patients
with 2009 influenza A (H1N1) infection. Crit. Care 15, R80.
Birsoy, K., Wang, T., Chen, W.W., Freinkman, E., Abu-Remaileh, M., and
Sabatini, D.M. (2015). An essential role of the mitochondrial electron transport
chain in cell proliferation is to enable aspartate synthesis. Cell 162, 540–551.
Blander, J.M. (2018). Regulation of the cell biology of antigen cross-presenta-
tion. Annu. Rev. Immunol. 36, 717–753.
Blum, J.S., Wearsch, P.A., and Cresswell, P. (2013). Pathways of antigen pro-
cessing. Annu. Rev. Immunol. 31, 443–473.
Borst, P. (2020). The malate-aspartate shuttle (Borst cycle): how it started and
developed into a major metabolic pathway. IUBMB Life 72, 2241–2259.
Boswell-Casteel, R.C., and Hays, F.A. (2017). Equilibrative nucleoside trans-
porters-A review. Nucleosides Nucleotides Nucleic Acids 36, 7–30.
Brisse, E., Matthys, P., and Wouters, C.H. (2016a). Understanding the spec-
trum of haemophagocytic lymphohistiocytosis: update on diagnostic chal-
lenges and therapeutic options. Br. J. Haematol. 174, 175–187.Cell Metabolism 34, 106–124, January 4, 2022 121
ll
OPEN ACCESS ArticleBrisse, E., Wouters, C.H., and Matthys, P. (2016b). Advances in the pathogen-
esis of primary and secondary haemophagocytic lymphohistiocytosis: differ-
ences and similarities. Br. J. Haematol. 174, 203–217.
Broen, J.C.A., and van Laar, J.M. (2020). Mycophenolate mofetil, azathioprine
and tacrolimus: mechanisms in rheumatology. Nat. Rev. Rheumatol. 16,
167–178.
Bryceson, Y.T., Pende, D., Maul-Pavicic, A., Gilmour, K.C., Ufheil, H., Vraetz,
T., Chiang, S.C., Marcenaro, S., Meazza, R., Bondzio, I., et al. (2012). A pro-
spective evaluation of degranulation assays in the rapid diagnosis of familial
hemophagocytic syndromes. Blood 119, 2754–2763.
Bzowska, A., Kulikowska, E., and Shugar, D. (2000). Purine nucleoside phos-
phorylases: properties, functions, and clinical aspects. Pharmacol. Ther. 88,
349–425.
Cader,M.Z., Boroviak, K., Zhang, Q., Assadi, G., Kempster, S.L., Sewell, G.W.,
Saveljeva, S., Ashcroft, J.W., Clare, S., Mukhopadhyay, S., et al. (2016).
C13orf31 (FAMIN) is a central regulator of immunometabolic function. Nat.
Immunol. 17, 1046–1056.
Cader, M.Z., de Almeida Rodrigues, R.P., West, J.A., Sewell, G.W., Md-
Ibrahim, M.N., Reikine, S., Sirago, G., Unger, L.W., Iglesias-Romero, A.B.,
Ramshorn, K., et al. (2020). FAMIN is a multifunctional purine enzyme enabling
the purine nucleotide cycle. Cell 180, 278–295.e23.
Cantrell, D. (2015). Signaling in lymphocyte activation. Cold Spring Harb.
Perspect. Biol. 7, a018788.
Cebrian, I., Visentin, G., Blanchard, N., Jouve, M., Bobard, A., Moita, C.,
Enninga, J., Moita, L.F., Amigorena, S., and Savina, A. (2011). Sec22b regu-
lates phagosomalmaturation and antigen crosspresentation by dendritic cells.
Cell 147, 1355–1368.
Cekic, C., Sag, D., Day, Y.J., and Linden, J. (2013). Extracellular adenosine
regulates naive T cell development and peripheral maintenance. J. Exp.
Med. 210, 2693–2706.
Dobin, A., and Gingeras, T.R. (2015). Mapping RNA-seq reads with STAR.
Curr. Protoc. Bioinformatics 51, 11.14.1–11.14.19.
Durai, V., and Murphy, K.M. (2016). Functions of murine dendritic cells.
Immunity 45, 719–736.
Ellinghaus, D., Degenhardt, F., Bujanda, L., Buti, M., Albillos, A., Invernizzi, P.,
Ferna´ndez, J., Prati, D., Baselli, G., Asselta, R., et al.; Severe Covid-19 GWAS
Group (2020). Genomewide association study of severe Covid-19with respira-
tory failure. N. Engl. J. Med. 383, 1522–1534.
Everitt, A.R., Clare, S., Pertel, T., John, S.P., Wash, R.S., Smith, S.E., Chin,
C.R., Feeley, E.M., Sims, J.S., Adams, D.J., et al.; GenISIS Investigators;
MOSAIC Investigators (2012). IFITM3 restricts the morbidity and mortality
associated with influenza. Nature 484, 519–523.
Fan, K.Q., Li, Y.Y., Wang, H.L., Mao, X.T., Guo, J.X., Wang, F., Huang, L.J., Li,
Y.N., Ma, X.Y., Gao, Z.J., et al. (2019). Stress-induced metabolic disorder in
peripheral CD4+ T cells leads to anxiety-like behavior. Cell 179, 864–879.e19.
Folch, J., Lees, M., and Sloane Stanley, G.H. (1957). A simple method for the
isolation and purification of total lipides from animal tissues. J. Biol. Chem.
226, 497–509.
Giblett, E.R., Anderson, J.E., Cohen, F., Pollara, B., and Meuwissen, H.J.
(1972). Adenosine-deaminase deficiency in two patients with severely
impaired cellular immunity. Lancet 2, 1067–1069.
Giblett, E.R., Ammann, A.J., Wara, D.W., Sandman, R., and Diamond, L.K.
(1975). Nucleoside-phosphorylase deficiency in a child with severely defective
T-cell immunity and normal B-cell immunity. Lancet 1, 1010–1013.
Goodman, R.P., Markhard, A.L., Shah, H., Sharma, R., Skinner, O.S., Clish,
C.B., Deik, A., Patgiri, A., Hsu, Y.H., Masia, R., et al. (2020). Hepatic NADH
reductive stress underlies common variation in metabolic traits. Nature 583,
122–126.
Grom, A.A., Horne, A., and De Benedetti, F. (2016). Macrophage activation
syndrome in the era of biologic therapy. Nat. Rev. Rheumatol. 12, 259–268.
Guicherit, O.M., Cooper, B.F., Rudolph, F.B., and Kellems, R.E. (1994).
Amplification of an adenylosuccinate synthetase gene in alanosine-resistant
murine T-lymphoma cells. Molecular cloning of a cDNA encoding the ‘‘non-
muscle’’ isozyme. J. Biol. Chem. 269, 4488–4496.122 Cell Metabolism 34, 106–124, January 4, 2022Han, J., Gagnon, S., Eckle, T., and Borchers, C.H. (2013). Metabolomic anal-
ysis of key central carbon metabolism carboxylic acids as their 3-nitrophenyl-
hydrazones by UPLC/ESI-MS. Electrophoresis 34, 2891–2900.
Hanauer, S.B., Sandborn, W.J., and Lichtenstein, G.R. (2019). Evolving con-
siderations for thiopurine therapy for inflammatory bowel diseases-a clinical
practice update: commentary. Gastroenterology 156, 36–42.
Hedstrom, L. (2012). The dynamic determinants of reaction specificity in the
IMPDH/GMPR family of (b/a)(8) barrel enzymes. Crit. Rev. Biochem. Mol.
Biol. 47, 250–263.
Heng, T.S., and Painter, M.W.; Immunological Genome Project Consortium
(2008). The Immunological Genome Project: networks of gene expression in
immune cells. Nat. Immunol. 9, 1091–1094.
Henter, J.I., Palmkvist-Kaijser, K., Holzgraefe, B., Bryceson, Y.T., and Palme´r,
K. (2010). Cytotoxic therapy for severe swine flu A/H1N1. Lancet 376, 2116.
Heuser, J. (1989). Changes in lysosome shape and distribution correlated with
changes in cytoplasmic pH. J. Cell Biol. 108, 855–864.
Hogquist, K.A., Jameson, S.C., Heath, W.R., Howard, J.L., Bevan, M.J., and
Carbone, F.R. (1994). T cell receptor antagonist peptides induce positive se-
lection. Cell 76, 17–27.
Hufford, M.M., Kim, T.S., Sun, J., and Braciale, T.J. (2011). Antiviral CD8+
T cell effector activities in situ are regulated by target cell type. J. Exp. Med.
208, 167–180.
Hutchison, A.J., Webb, R.L., Oei, H.H., Ghai, G.R., Zimmerman, M.B., and
Williams, M. (1989). CGS 21680C, an A2 selective adenosine receptor agonist
with preferential hypotensive activity. J. Pharmacol. Exp. Ther. 251, 47–55.
Joffre, O.P., Segura, E., Savina, A., and Amigorena, S. (2012). Cross-presen-
tation by dendritic cells. Nat. Rev. Immunol. 12, 557–569.
Kang, J.-W., Yan, J., Ranjan, K., Zhang, X., Turner, J.R., and Abraham, C.
(2020). Myeloid cell expression of LACC1 is required for bacterial clearance
and control of intestinal inflammation. Gastroenterology 159, 1051–1067.
Kaser, A. (2020). Genetic risk of severe Covid-19. N. Engl. J. Med. 383,
1590–1591.
Kim, D., Langmead, B., and Salzberg, S.L. (2015). HISAT: a fast spliced aligner
with low memory requirements. Nat. Methods 12, 357–360.
Kirschner, M., and Mitchison, T. (1986). Beyond self-assembly: from microtu-
bules to morphogenesis. Cell 45, 329–342.
Korolchuk, V.I., Saiki, S., Lichtenberg,M., Siddiqi, F.H., Roberts, E.A., Imarisio,
S., Jahreiss, L., Sarkar, S., Futter, M., Menzies, F.M., et al. (2011). Lysosomal
positioning coordinates cellular nutrient responses. Nat. Cell Biol. 13,
453–460.
Kraman, M., Bambrough, P.J., Arnold, J.N., Roberts, E.W., Magiera, L., Jones,
J.O., Gopinathan, A., Tuveson, D.A., and Fearon, D.T. (2010). Suppression of
antitumor immunity by stromal cells expressing fibroblast activation protein-
alpha. Science 330, 827–830.
Krebs, H.A. (1967). The redox state of nicotinamide adenine dinucleotide in the
cytoplasm and mitochondria of rat liver. Adv. Enzyme Regul. 5, 409–434.
Lahiri, A., Hedl, M., Yan, J., and Abraham, C. (2017). Human LACC1 increases
innate receptor-induced responses and a LACC1 disease-risk variant modu-
lates these outcomes. Nat. Commun. 8, 15614.
Leone, R.D., Zhao, L., Englert, J.M., Sun, I.M., Oh, M.H., Sun, I.H., Arwood,
M.L., Bettencourt, I.A., Patel, C.H., Wen, J., et al. (2019). Glutamine blockade
induces divergent metabolic programs to overcome tumor immune evasion.
Science 366, 1013–1021.
Lowenstein, J.M. (1972). Ammonia production in muscle and other tissues: the
purine nucleotide cycle. Physiol. Rev. 52, 382–414.
Lowenstein, J.M. (1990). The purine nucleotide cycle revisited [corrected]. Int.
J. Sports Med. 11 (Suppl 2 ), S37–S46.
Lowenstein, J., and Tornheim, K. (1971). Ammonia production in muscle: the
purine nucleotide cycle. Science 171, 397–400.
Lucas, C., Wong, P., Klein, J., Castro, T.B.R., Silva, J., Sundaram, M.,
Ellingson, M.K., Mao, T., Oh, J.E., Israelow, B., et al.; Yale IMPACT Team
(2020). Longitudinal analyses reveal immunological misfiring in severe
COVID-19. Nature 584, 463–469.
ll
OPEN ACCESSArticleLykens, J.E., Terrell, C.E., Zoller, E.E., Risma, K., and Jordan, M.B. (2011).
Perforin is a critical physiologic regulator of T-cell activation. Blood 118,
618–626.
Mager, L.F., Burkhard, R., Pett, N., Cooke, N.C.A., Brown, K., Ramay, H., Paik,
S., Stagg, J., Groves, R.A., Gallo, M., et al. (2020). Microbiome-derived inosine
modulates response to checkpoint inhibitor immunotherapy. Science 369,
1481–1489.
Manjunath, N., Shankar, P., Wan, J., Weninger, W., Crowley, M.A., Hieshima,
K., Springer, T.A., Fan, X., Shen, H., Lieberman, J., and von Andrian, U.H.
(2001). Effector differentiation is not prerequisite for generation of memory
cytotoxic T lymphocytes. J. Clin. Invest. 108, 871–878.
Morre´, D.J., and Morre´, D.M. (2011). Non-mitochondrial coenzyme Q.
Biofactors 37, 355–360.
Naik, S.H., Proietto, A.I., Wilson, N.S., Dakic, A., Schnorrer, P., Fuchsberger,
M., Lahoud, M.H., O’Keeffe, M., Shao, Q.X., Chen, W.F., et al. (2005).
Cutting edge: generation of splenic CD8+ and CD8- dendritic cell equivalents
in Fms-like tyrosine kinase 3 ligand bone marrow cultures. J. Immunol. 174,
6592–6597.
Nakanishi, T., Pigazzini, S., Degenhardt, F., Cordioli, M., Butler-Laporte, G.,
Maya-Miles, D., Bujanda, L., Bouysran, Y., Niemi, M.E.K., Palom, A., et al.
(2021). Age-dependent impact of the major common genetic risk factor for
COVID-19 on severity and mortality. J. Clin. Invest. 131, e152386.
Neelapu, S.S., Tummala, S., Kebriaei, P., Wierda, W., Gutierrez, C., Locke,
F.L., Komanduri, K.V., Lin, Y., Jain, N., Daver, N., et al. (2018). Chimeric antigen
receptor T-cell therapy - assessment and management of toxicities. Nat. Rev.
Clin. Oncol. 15, 47–62.
Noto, A., Ngauv, P., and Trautmann, L. (2013). Cell-based flow cytometry
assay to measure cytotoxic activity. J. Vis. Exp. 82, e51105.
Ohta, A., Gorelik, E., Prasad, S.J., Ronchese, F., Lukashev, D., Wong, M.K.,
Huang, X., Caldwell, S., Liu, K., Smith, P., et al. (2006). A2A adenosine receptor
protects tumors from antitumor T cells. Proc. Natl. Acad. Sci. USA 103,
13132–13137.
Pairo-Castineira, E., Clohisey, S., Klaric, L., Bretherick, A.D., Rawlik, K.,
Pasko, D., Walker, S., Parkinson, N., Fourman, M.H., Russell, C.D., et al.;
GenOMICC Investigators; ISARIC4C Investigators; COVID-19 Human
Genetics Initiative; 23andMe Investigators; BRACOVID Investigators; Gen-
COVID Investigators (2021). Genetic mechanisms of critical illness in COVID-
19. Nature 591, 92–98.
Patel, N., El Mouzan, M.I., Al-Mayouf, S.M., Adly, N., Mohamed, J.Y., Al
Mofarreh, M.A., Ibrahim, N., Xiong, Y., Zhao, Q., Al-Saleem, K.A., and
Alkuraya, F.S. (2014). Study of Mendelian forms of Crohn’s disease in Saudi
Arabia reveals novel risk loci and alleles. Gut 63, 1831–1832.
Porgador, A., Yewdell, J.W., Deng, Y., Bennink, J.R., and Germain, R.N.
(1997). Localization, quantitation, and in situ detection of specific peptide-
MHC class I complexes using a monoclonal antibody. Immunity 6, 715–726.
Prabhu, N., Ho, A.W., Wong, K.H., Hutchinson, P.E., Chua, Y.L., Kandasamy,
M., Lee, D.C., Sivasankar, B., and Kemeny, D.M. (2013). Gamma interferon
regulates contraction of the influenza virus-specific CD8 T cell response and
limits the size of the memory population. J. Virol. 87, 12510–12522.
Prieto-Pe´rez, L., Fortes, J., Soto, C., Vidal-Gonza´lez, A´., Alonso-Rian˜o, M.,
Lafarga, M., Cortti, M.J., Lazaro-Garcia, A., Pe´rez-Tanoira, R., Trascasa, A´.,
et al. (2020). Histiocytic hyperplasia with hemophagocytosis and acute alve-
olar damage in COVID-19 infection. Mod. Pathol. 33, 2139–2146.
Rabionet, R., Remesal, A., Mensa-Vilaro´, A., Murı´as, S., Alcobendas, R.,
Gonza´lez-Roca, E., Ruiz-Ortiz, E., Anto´n, J., Iglesias, E., Modesto, C., et al.
(2019). Biallelic loss-of-function LACC1/FAMIN mutations presenting as rheu-
matoid factor-negative polyarticular juvenile idiopathic arthritis. Sci. Rep.
9, 4579.
Ralser, M. (2018). An appeal to magic? The discovery of a non-enzymatic
metabolism and its role in the origins of life. Biochem. J. 475, 2577–2592.
Rivas, M.A., Avila, B.E., Koskela, J., Huang, H., Stevens, C., Pirinen, M.,
Haritunians, T., Neale, B.M., Kurki, M., Ganna, A., et al.; International IBD
Genetics Consortium; NIDDK IBD Genetics Consortium; T2D-GENESConsortium (2018). Insights into the genetic epidemiology of Crohn’s and
rare diseases in the Ashkenazi Jewish population. PLoS Genet. 14, e1007329.
Rodriguez, M., Moreau, P., Paulik, M., Lawrence, J., Morre´, D.J., andMorre´, D.
(1992). NADH-activated cell-free transfer between Golgi apparatus and
plasma membranes of rat liver. Biochim. Biophys. Acta 1107, 131–138.
Sancho, D., Joffre, O.P., Keller, A.M., Rogers, N.C., Martı´nez, D., Hernanz-
Falco´n, P., Rosewell, I., and Reis e Sousa, C. (2009). Identification of a den-
dritic cell receptor that couples sensing of necrosis to immunity. Nature 458,
899–903.
Skon-Hegg, C., Zhang, J., Wu, X., Sagolla, M., Ota, N., Wuster, A., Tom, J.,
Doran, E., Ramamoorthi, N., Caplazi, P., et al. (2019). LACC1 regulates TNF
and IL-17 in mouse models of arthritis and inflammation. J. Immunol. 202,
183–193.
Stenmark, H. (2009). Rab GTPases as coordinators of vesicle traffic. Nat. Rev.
Mol. Cell Biol. 10, 513–525.
Sullivan, L.B., Gui, D.Y., Hosios, A.M., Bush, L.N., Freinkman, E., and Vander
Heiden, M.G. (2015). Supporting aspartate biosynthesis is an essential func-
tion of respiration in proliferating cells. Cell 162, 552–563.
Sun, J., Madan, R., Karp, C.L., and Braciale, T.J. (2009). Effector T cells control
lung inflammation during acute influenza virus infection by producing IL-10.
Nat. Med. 15, 277–284.
Sun, R., Huang, J., Yang, N., He, J., Yu, X., Feng, S., Xie, Y., Wang, G., Ye, H.,
and Aa, J. (2019). Purine catabolism shows a dampened circadian rhythmicity
in a high-fat diet-induced mouse model of obesity. Molecules 24, 4524.
Tesmer, J.J., Klem, T.J., Deras, M.L., Davisson, V.J., and Smith, J.L. (1996).
The crystal structure of GMP synthetase reveals a novel catalytic triad and is
a structural paradigm for two enzyme families. Nat. Struct. Biol. 3, 74–86.
Tiede, I., Fritz, G., Strand, S., Poppe, D., Dvorsky, R., Strand, D., Lehr, H.A.,
Wirtz, S., Becker, C., Atreya, R., et al. (2003). CD28-dependent Rac1 activation
is the molecular target of azathioprine in primary human CD4+ T lymphocytes.
J. Clin. Invest. 111, 1133–1145.
To, K.K.W., Mok, K.Y., Chan, A.S.F., Cheung, N.N., Wang, P., Lui, Y.M., Chan,
J.F.W., Chen, H., Chan, K.H., Kao, R.Y.T., and Yuen, K.Y. (2016).
Mycophenolic acid, an immunomodulator, has potent and broad-spectrum
in vitro antiviral activity against pandemic, seasonal and avian influenza viruses
affecting humans. J. Gen. Virol. 97, 1807–1817.
Udaka, S., and Moyed, H.S. (1963). Inhibition of parental and mutant xantho-
sine 50-phosphate aminases by psicofuranine. J. Biol. Chem. 238, 2797–2803.
Versini, M., Jeandel, P.Y., Rosenthal, E., and Shoenfeld, Y. (2014). Obesity in
autoimmune diseases: not a passive bystander. Autoimmun. Rev. 13,
981–1000.
Wakil, S.M., Monies, D.M., Abouelhoda, M., Al-Tassan, N., Al-Dusery, H.,
Naim, E.A., Al-Younes, B., Shinwari, J., Al-Mohanna, F.A., Meyer, B.F., and
Al-Mayouf, S. (2015). Association of a mutation in LACC1 with a monogenic
form of systemic juvenile idiopathic arthritis. Arthritis Rheumatol. 67, 288–295.
Walton, Z.E., Patel, C.H., Brooks, R.C., Yu, Y., Ibrahim-Hashim, A., Riddle, M.,
Porcu, A., Jiang, T., Ecker, B.L., Tameire, F., et al. (2018). Acid suspends the
circadian clock in hypoxia through inhibition of mTOR. Cell 174, 72–87.e32.
Wang, T., Gnanaprakasam, J.N.R., Chen, X., Kang, S., Xu, X., Sun, H., Liu, L.,
Rodgers, H., Miller, E., Cassel, T.A., et al. (2020). Inosine is an alternative car-
bon source for CD8+-T-cell function under glucose restriction. Nat. Metab. 2,
635–647.
Welihinda, A.A., Kaur, M., Raveendran, K.S., and Amento, E.P. (2018).
Enhancement of inosine-mediated A2AR signaling through positive allosteric
modulation. Cell. Signal. 42, 227–235.
Williams, A.P., Peh, C.A., Purcell, A.W., McCluskey, J., and Elliott, T. (2002).
Optimization of the MHC class I peptide cargo is dependent on tapasin.
Immunity 16, 509–520.
Williamson, D.H., Lund, P., and Krebs, H.A. (1967). The redox state of free
nicotinamide-adenine dinucleotide in the cytoplasm and mitochondria of rat
liver. Biochem. J. 103, 514–527.
Xia, Y., Xie, Y., Yu, Z., Xiao, H., Jiang, G., Zhou, X., Yang, Y., Li, X., Zhao, M., Li,
L., et al. (2018). The mevalonate pathway is a druggable target for vaccine
adjuvant discovery. Cell 175, 1059–1073.e21.Cell Metabolism 34, 106–124, January 4, 2022 123
ll
OPEN ACCESS ArticleYasin, S., and Schulert, G.S. (2018). Systemic juvenile idiopathic arthritis and
macrophage activation syndrome: update on pathogenesis and treatment.
Curr. Opin. Rheumatol. 30, 514–520.
Zelenay, S., Keller, A.M., Whitney, P.G., Schraml, B.U., Deddouche, S.,
Rogers, N.C., Schulz, O., Sancho, D., and Reis e Sousa, C. (2012). The den-
dritic cell receptor DNGR-1 controls endocytic handling of necrotic cell anti-
gens to favor cross-priming of CTLs in virus-infected mice. J. Clin. Invest.
122, 1615–1627.124 Cell Metabolism 34, 106–124, January 4, 2022Zhang, F.R., Huang, W., Chen, S.M., Sun, L.D., Liu, H., Li, Y., Cui, Y., Yan, X.X.,
Yang, H.T., Yang, R.D., et al. (2009). Genomewide association study of
leprosy. N. Engl. J. Med. 361, 2609–2618.
Zocchi, C., Ongini, E., Conti, A., Monopoli, A., Negretti, A., Baraldi, P.G., and
Dionisotti, S. (1996). The non-xanthine heterocyclic compound SCH 58261
is a new potent and selective A2a adenosine receptor antagonist.
J. Pharmacol. Exp. Ther. 276, 398–404.
ll
OPEN ACCESSArticleSTAR+METHODSKEY RESOURCES TABLEREAGENT or RESOURCE SOURCE IDENTIFIER
Antibodies
B220-PE Biolegend RRID: AB_312992
CD11b-Pe/Cy7 Biolegend RRID: AB_312799
CD11c-APC Biolegend RRID: AB_313778
CD11c-PE Biolegend RRID: AB_313777
CD16/CD32 Biolegend RRID: AB_312801
CD172a (SIRPa)-PE/Dazzle 594 Biolegend RRID: AB_2565279
CD24-AF647 Biolegend RRID: AB_493485
CD252-Pe/Cy7 Biolegend RRID: AB_2565744
CD274-BV785 Biolegend RRID: AB_2629659
CD275-PE Biolegend RRID: AB_2248797
CD28 (37.51) Thermo Fisher Scientific RRID: AB_468921
CD3-AF647 Biolegend RRID: AB_389323
CD3-PE/Cy5 Biolegend RRID: AB_312674
CD3-PECy7 Biolegend RRID: AB_312675
CD3ε (145-2C11) Thermo Fisher Scientific RRID: AB_468848
CD4-BV605 Biolegend RRID: AB_2564591
CD4-FITC Biolegend RRID: AB_312691
CD40-PE/Cy7 Biolegend RRID: AB_10933422
CD44-FITC Biolegend RRID: AB_493684
CD45-APC/Cy7 Biolegend RRID: AB_312980
CD45-Pe/Cy7 Biolegend RRID: AB_312978
CD45.1-APC Biolegend RRID: AB_313503
CD45R/B220-FITC Biolegend RRID: AB_312990
CD64-BV605 Biolegend RRID: AB_2629778
CD8a-BV605 Biolegend RRID: AB_2561352
CD8a-BV650 Biolegend RRID: AB_11124344
CD80-AF647 Biolegend RRID: AB_492824
CD86-FITC Biolegend RRID: AB_313149
Foxp3-PerCP/Cy5.5 Thermo Fisher Scientific RRID: AB_914349
H-2Kb-PerCP/Cy5.5 Biolegend RRID: AB_1967107
H-2Kb-SIINFEKL (25-D1.16)-APC Biolegend RRID: AB_11219402
H-2Kb/H-2Db-FITC Biolegend RRID: AB_313597
I-A/I-E-APC/Cy7 Biolegend RRID: AB_1659252
I-A/I-E-APC/Fire 750 Biolegend RRID: AB_2616728
I-A/I-E-BV510 Biolegend RRID: AB_2561397
IFNg-PE/Cy7 Biolegend RRID: AB_2295770
IL-17a-PerCP/Cy5.5 BD Horizon RRID: AB_2738642
IL-4-FITC Thermo Fisher Scientific RRID: AB_465387
Influenza A Virus Nucleoprotein-FITC Abcam RRID: AB_445914
SIRPa-PE Biolegend RRID: AB_2563549
TCR Va2-Pe/Cy7 Thermo Fisher Scientific RRID: AB_2573472
TCRVb5.1 5.2-PE Biolegend RRID: AB_10612761
Anti-mouse IgG, HRP-linked (polyclonal) Cell Signaling Technology RRID: AB_330924
Anti-rabbit IgG, HRP-linked (polyclonal) Cell Signaling Technology RRID: AB_2099233
(Continued on next page)
Cell Metabolism 34, 106–124.e1–e10, January 4, 2022 e1
Continued
REAGENT or RESOURCE SOURCE IDENTIFIER
Beta-actin (13-E4) (monoclonal) Cell Signaling Technology RRID: AB_2223172
ADA (polyclonal) Abcam Cat# ab217846
PNP (H-7) (monoclonal) Santa Cruz Biotechnology RRID: AB_10845931
MTAP (42-T) (monoclonal) Santa Cruz Biotechnology RRID: AB_2147095
Bacterial and virus strains
H3N2 IAV strain (A/X-31) (Everitt et al., 2012) N/A
Chemicals, peptides, and recombinant proteins
20X LumiGLO Reagent and 20X Peroxide Cell Signaling Technology Cat# 7003
3-(hydroxynitrosoamino)-L-alanine (L-
alanosine)
Cayman Chemicals Cat# 19545, CAS: 5854-93-3
3-Nitrophenylhydrazine hydrochloride Sigma-Aldrich Cat# N21804-5G
4X Laemmli sample buffer Bio-Rad Cat# 1610747
5(6)-CFDA, SE. Thermo Fisher Scientific Cat# C1157, CAS: 150347-59-4
6-Mercaptopurine monohydrate Sigma-Aldrich Cat# 852678-1G-A, CAS: 6112-76-1
Acetic acid, glacial Thermo Fisher Scientific Cat# 695092, CAS: 64-09-7
AMP Deaminase inhibitor, Cpd3 Sigma-Aldrich Cat# 533642
Antimycin A Sigma-Aldrich Cat# A867
BSA-Palmitate saturated fatty acid
complex (5 mM)
Cayman Chemicals Cat# 29558
CGS-21680 hydrochloride hydrate Sigma-Aldrich Cat# C141, CAS: 124182-57-6
Collagenase D Sigma-Aldrich Cat# 11088858001
cOmplete Protease Inhibitor Cocktail Roche Cat# 11836170001
D-Glucose-13C6 Sigma-Aldrich Cat# 389374, CAS: 110187-42-3
ExtrAvidin-R-Phycoerythrin Sigma-Aldrich Cat# E4011
FCCP Sigma-Aldrich Cat# C2920
Fixable viability dye eFluor450 Thermo Fisher Scientific Cat# 65-0863-14
Glutaraldehyde solution Sigma-Aldrich Cat# G7651, CAS 111-30-8
Guanine Sigma-Aldrich Cat# 51030, CAS: 635-39-2
H-2D(b) Influenza A NP366-374
ASNENMETM
NIH Tetramer Facility N/A
H-2D(b) Influenza A PA224-233
SSLENFRAYV
NIH Tetramer Facility N/A
H-2K(b) chicken OVA257-264 SIINFEKL NIH Tetramer Facility N/A
Hadacidin SantaCruz Biotechnology Cat# sc-490177, CAS: 689-13-4
Hypoxanthine-13C5,
15N4 Cambridge Isotope Laboratories Cat# CNLM-7894-0
Inosine Sigma-Aldrich Cat# I4625, CAS: 58-63-9
Inosine, 13C5 Omicron Biochemicals Cat# NUC-072
Inosine, 15N4 Cambridge Isotope Laboratories Cat# NLM-4264-PK
L-aspartic acid-13C4,
15N1 Sigma-Aldrich Cat# 607835, CAS 202468-27-7
L-Glutamine-13C5,
15N2 Sigma-Aldrich Cat# 607983
L-malic acid-13C4 Cambridge Isotope Laboratories Cat# CLM-8065-0
LPS (O111:B4) Sigma-Aldrich Cat# LPS25
mFlt3L Miltenyi Biotec Cat# 130-097-372
mIL-15 Miltenyi Biotec Cat# 130-094-072
mIL-2 Miltenyi Biotec Cat# 130-094-054
Mycophenolic acid Sigma-Aldrich Cat# M5255-50MG, CAS: 24280-93-1
N-(3-Dimethylaminopropyl)-N0-
ethylcarbodiimide hydrochloride
Sigma-Aldrich Cat# E6383-5G
Oligomycin Sigma-Aldrich Cat# 75371, CAS: 579-13-5
OVA257-264 SIINFEKL Invivogen Cat# vac-sin
(Continued on next page)
ll
OPEN ACCESS Article
e2 Cell Metabolism 34, 106–124.e1–e10, January 4, 2022
Continued
REAGENT or RESOURCE SOURCE IDENTIFIER
OVA323-339 ISQAVHAAHAEINEAGR Invivogen Cat# vac-isq
Ovalbumin Sigma-Aldrich Cat# A5503
Ovalbumin, Alexa Fluor 488 Conjugate Thermo Fisher Scientific Cat# O34781
Ovalbumin, Alexa Fluor 647 Conjugate Thermo Fisher Scientific Cat# O34784
Palmitic acid Sigma-Aldrich Cat# P0500, CAS 57-10-3
Palmitic acid-13C16 Sigma-Aldrich Cat# 605573, CAS: 56599-85-0
Paraformaldehyde Thermo Fisher Scientific Cat# 28908, CAS: 3025-89-4
Pierce BCA Protein Assay Kit Thermo Fisher Scientific Cat#23225
Poly-L-Lysine Sigma-Aldrich Cat# P8920-100ML, CAS: 25988-63-0
Psicofuranine Cayman Chemicals Cat# 19574, CAS: 1874-54-0
Pyridine Sigma-Aldrich Cat# 270970-100ML
Rotenone Sigma-Aldrich Cat# R8875, CAS: 83-79-4
SCH 58261 Sigma-Aldrich Cat# S4568, CAS: 160098-96-4
Sodium a-ketobutyrate Sigma-Aldrich Cat# K0875, CAS 2013-26-5
Sodium a-ketobutyrate-13C4 Cambridge Isotope Laboratories Cat# CLM-6164-0.5, CAS: 2483736-24-7
Sodium chloride Sigma-Aldrich Cat# S9888, CAS 7647-14-5
Sodium pyruvate Sigma-Aldrich Cat# P5280, CAS 113-24-6
Sodium pyruvate-13C3 Sigma-Aldrich Cat# 490717, CAS 142014-11-7
SYTOX Blue Dead Cell Stain Thermo Fisher Scientific Cat# S34857
SYTOX Green Nucleic Acid Stain Thermo Fisher Scientific Cat# S7020
b-lactamase Sigma-Aldrich Cat# P0389, CAS: 9073-60-3
Critical commercial assays
Anti-PE MicroBeads Miltenyi Biotec Cat# 130-048-80
CD11c MicroBeads UltraPure Miltenyi Biotec Cat# 130-108-338
CD4+ T Cell Isolation Kit, mouse Miltenyi Biotec Cat# 130-104-454
CD8a+ T Cell Isolation Kit, mouse Miltenyi Biotec Cat# 130-104-075
Foxp3 / Transcription Factor Staining
Buffer Set
Thermo Fisher Scientific Cat# 00-5523-00
IFNg gamma Mouse ELISA kit Thermo Fisher Scientific Cat# 88-7314-22; RRID: AB_2575066
In situ Cell Death detection Kit, POD Sigma-Aldrich Cat# 11684817910
LEGENDplex Mouse Inflammation Panel Biolegend Cat# 740150
LiveBLAzer FRET-B/G Loading Kit with
CCF4-AM
Thermo Fisher Scientific Cat# K1095
Mouse Dendritic Cell Nucleofection Kit Lonza Cat# VVPA-1011
Mouse Granzyme B DuoSet ELISA R&D Systems Cat# DY1865
Mouse IFN-a ELISA kit (TCM) PBL Assay Science Cat# 42120-1
Mouse IL-2 ELISA kit Thermo Fisher Scientific Cat# 15530997
pHrodo Red AM Intracellular pH Indicator Thermo Fisher Scientific Cat# P35372
Red Blood Cell Lysis Solution (10 3 ) Miltenyi Biotec Cat# 130-094-183
RNeasy Mini Kit QIAGEN Cat# 74104
RNeasy Plus Micro Kit QIAGEN Cat# 74034
Thermo Fisher Scientific IL-12 p70 Mouse
Uncoated ELISA Kit
Thermo Fisher Scientific Cat# 12384003
TruSeq stranded mRNA library prep kit Illumina Cat# 20020594
Deposited data
RNA-Seq (Dendritic cell dataset) This paper GEO: GSE126473
RNA-Seq (T cell dataset) This paper GEO: GSE147370
Experimental models: Cell lines
LL2-ovalbumin (Kraman et al., 2010) N/A
MEFs, bm1 T OVA (Sancho et al., 2009) N/A
(Continued on next page)
ll
OPEN ACCESSArticle
Cell Metabolism 34, 106–124.e1–e10, January 4, 2022 e3
Continued
REAGENT or RESOURCE SOURCE IDENTIFIER
Experimental models: Organisms/strains
Mouse: Famin/ (Cader et al., 2016) N/A
Mouse: Famin+/+ (Cader et al., 2016) N/A
Mouse: Faminp.254I (Cader et al., 2016) N/A
Mouse: Faminp.254V (Cader et al., 2016) N/A
Mouse: Faminp.284R (Cader et al., 2016) N/A
Mouse: FaminWT (Faminfl/fl) N/A N/A
Mouse: FaminDDC (Faminfl/fl;Cd11c-Cre) N/A N/A
Mouse: OT-I;Rag1/ (Barnden et al., 1998) N/A
Mouse: OT-II;Rag2/ (Hogquist et al., 1994) N/A
Oligonucleotides
ON-TARGETplus Mouse Adsl siRNA Horizon Cat# L-064380-01
ON-TARGETplus Mouse Adss siRNA Horizon Cat# L-060265-01
ON-TARGETplus Mouse Ampd2 siRNA Horizon Cat# L-063716-01
ON-TARGETplus Mouse Ampd3 siRNA Horizon Cat# L-042904-01
ON-TARGETplus Mouse Gmpr siRNA Horizon Cat# L-046519-01
ON-TARGETplus Mouse Gmps siRNA Horizon Cat# L-049796-01
ON-TARGETplus Mouse Impdh1 siRNA Horizon Cat# L-042235-01
ON-TARGETplus Mouse Impdh2 siRNA Horizon Cat# L-062809-01
ON-TARGETplus Mouse Cpt1 siRNA Horizon Cat# L-042456-01
ON-TARGETplus Non-targeting
Control Pool
Horizon Cat# D-001810-10
Murine Adss F 50-
CTGGCCACACAGTTGTCGTA-30; R 50-
AAGCCTTTTCTCCCAGCCAT-30
Thermo Fisher Scientific N/A
Murine Adsl F 50-
GGATCACCAGAAGGTGGAGC-30; R 50-
TGTGCACCGATGCTCCTAAG-30
Thermo Fisher Scientific N/A
Murine Ampd2 F 50-
CTCCTTGCATTTGCCATCCAT-30; R 50-
CCTCTCCGCTACAGTCTGC-30
Thermo Fisher Scientific N/A
Murine Ampd3 F 50-
CTGCGACCGGATCATCTTGAA 30; R 50-
GTTGGCGGAGAAGGTGTTTG 30
Thermo Fisher Scientific N/A
Murine Cpt1a F 50-
TGGCATCATCACTGGTGTGTT-30; R 50-
GTCTAGGGTCCGATTGATCTTTG-30
Thermo Fisher Scientific N/A
Murine Gmpr2 F 50-
CAGCATCCATCAGTGGCAAGAG-30; R 50-
CCGTTAGCCACATCCAGGCATA-30
Thermo Fisher Scientific N/A
Murine Gmps F 50-
CCTTGTTGCCAGTGGTAAAGCC30; R 50-
TCTTCTGGCAGGTCAAGCTCTC-30
Thermo Fisher Scientific N/A
Murine Impdh1 F 50-
GGCTACGTTCCCGAGGATG 30; R 50-
GGCTGATGTCAGGTCCACTT-30
Thermo Fisher Scientific N/A
Murine Impdh2 F 50-
CTTGCTGGTGTGGATGTAGTGG 30; R
50-GCCTCCAATGACCTGTAGACTG-30
Thermo Fisher Scientific N/A
Murine Famin F 50-
TGGGGTTGCTCACTCCGGCTG-30; R 50-
GGAGACTGCTGATTCTTTGGGAAGA-30
Thermo Fisher Scientific N/A
(Continued on next page)
ll
OPEN ACCESS Article
e4 Cell Metabolism 34, 106–124.e1–e10, January 4, 2022
Continued
REAGENT or RESOURCE SOURCE IDENTIFIER
Murine ActinB F 50-
GATGCTCCCCGGGCTGTATT-30; R 50-
GGGGTACTTCAGGGTCAGGA-30
Thermo Fisher Scientific N/A
Influenza A virus H3N2 A/X-31 M protein F
50- GGACTGCAGCGTTAGACGCTT-30; R
50-CATCCTGTTGTATATGAGGCCCAT-30
Thermo Fisher Scientific N/A
Software and algorithms
Adobe Illustrator CC 2019 (23.0.3) Adobe https://www.adobe.com/products/
illustrator.html; RRID: SCR_010279
Compound Discoverer 2.1 and Compound
Discoverer 3.1
Thermo Fisher Scientific Cat# OPTON-30834
FlowJo FlowJo LLC version v10.6.2
GraphPad Prism 8.0 GraphPad Software LLC version 8.3.0
R 3.6.1 The Comprehensive R Archive Network Version 3.6.1
Thermo Xcalibur 4.1 Thermo Fisher Scientific Cat# OPTON-30382; RRID: SCR_014593
Other
Algal lyophilized cells 13C
(Synechococcus sp.)
Sigma-Aldrich Cat# 487945-1G
Fetal Bovine Serum Sigma-Aldrich Cat# F7524-500ML
HBSS Thermo Fisher Scientific Cat# 14170112
HEPES Thermo Fisher Scientific Cat# 15630049
OptiMEM Thermo Fisher Scientific Cat# 31985062
Penicillin-streptomycin Sigma-Aldrich Cat# P0781-50ML
RPMI 1640 Thermo Fisher Scientific Cat# 11875093
RPMI 1640, no glutamine Thermo Fisher Scientific Cat# 21870084
RPMI 1640, no glucose Thermo Fisher Scientific Cat# 11879020
Seahorse XF base medium Thermo Fisher Scientific Cat# 102353-100
ll
OPEN ACCESSArticleRESOURCE AVAILABILITY
Lead contact
Further information and requests for materials should be directed to and will be fulfilled by the Lead Contact, Arthur Kaser (ak729@
cam.ac.uk).
Materials availability
Unique resources generated in this study are available on reasonable request, although may require completion of a Materials
Transfer Agreement.
Data and code availability
d RNA Sequencing datasets generated in this study have been deposited at the Gene Expression Omnibus (GSE126473 and
GSE147370) and are publicly available as of the date of publication.
d This paper does not report original code.
d Any additional information required to reanalyse the data reported in this paper is available from the lead contact upon reason-
able request.EXPERIMENTAL MODEL AND SUBJECT DETAILS
Mice
Age-, gender-, and, whenever possible, littermate-matched 6- to 11-week-oldmice were used for all experiments. Micewere housed
in a 19-21C environment on a 12 h light/ dark cycle. Health status screening was performed every three months using sentinel mice.
Mice were fed on a universal maintenance chow diet purchased from Safe (Safe 105). Famin+/+, Famin/, Faminp.254I, Faminp.254VCell Metabolism 34, 106–124.e1–e10, January 4, 2022 e5
ll
OPEN ACCESS Articleand Faminp.284R mice, generated on a C57BL/6NTac background, have previously been described (Cader et al., 2016). The loxP-
flanked conditional allele (Faminfl/fl) allele was generated by converting mice with homologous recombination of the tm1a(KOMP)
Wtsi construct (Cader et al., 2016) with a FlpO recombinase-transgenic mouse. FaminDDC mice were then generated by crossing
Faminfl/fl mice with Itgax-Cre (CD11c-Cre) mice (B6.Cg-Tg(Itgax-cre)1-1Reiz/J, on a C57BL/6 background). Itgax-Cre;Faminfl/fl
mice and their controls were born at Mendelian ratio and developed normally, consistent with what we previously reported for the
other Famin mutant mice used in this study (Cader et al., 2016). None of these mice develop spontaneous autoimmunity or autoin-
flammation under specific pathogen-free conditions. OT-I;Rag/ (‘OT-I’) and OT-II;Ptprca;Rag2/ (‘OT-II’) mice on a C57BL/6
background have previously been described (Barnden et al., 1998; Hogquist et al., 1994). Maintenance and breeding under specific
pathogen-free conditions was performed at the Central Biomedical Services facility or Phenomics Laboratory, University of
Cambridge. UK Home Office and local ethics approval has been granted for all experimental procedures.
Dendritic cell and T cell isolation
DCs were isolated from spleens digested in 1 mg/mL collagenase D (Sigma, 11088858001) in complete media in the presence of
DNase (Sigma, D4263). Positive selection for CD11c+ DCswas performed usingCD11cUltrapure beads following themanufacturer’s
protocol (Miltenyi, 130-100-875). BM-derived DCs were isolated from murine tibias and femurs by flushing in complete RPMI-1640
medium, filtering through a 70 mm cell strainer, lysing red blood cells using red blood cell lysis solution (Miltenyi, 130-094-183)
in accordance with manufacturer’s instructions, and subsequent resuspension in complete RPMI-1640 medium (containing
100 U/mL of penicillin-streptomycin, 10 mM HEPES buffer and 10% FBS) followed by culture for 9 days in mFlt3L (100 ng/mL,
Miltenyi, 130-097-372), replenished on days 3 and 6. After 9 days in culture, isolation of BM-derived DCs into cDC1 and cDC2 sub-
sets was performed using a protocol adapted from (Zelenay et al., 2012). Briefly, 9-day Flt3L-expanded BM-derived cDC2s were
isolated by positive selection using SIRPa-PE antibody (Biolegend, 144011) and anti-PEmicrobeads (Miltenyi, 130-048-80), followed
by negative selection of cDC1 through depletion of plasmacytoid DCs with anti-B220-PE antibody (Biolegend, 103207) and anti-PE
microbeads (Miltenyi, 130-048-80). For maturation prior to all subsequent analysis, splenic DCs were rested overnight, while BM-
derived cDC1s or cDC2s were treated with 1 mg/mL LPS (O111:B4, Sigma, LPS25) for 18 h. All treatments of splenic DCs or BM-
derived cDC1swere performed during this maturation step unless specified otherwise. Inhibitors were used at the following final con-
centrations: 25 mM (Figures 4A and S4A) or 100 mM (Figure 5C) L-alanosine (CaymanChemicals,19545), 50 mMhadacidin (SantaCruz,
sc-490177), 5 mM 6-mercaptopurine (Sigma, 852678-1G-A), 5 mM Cpd3 (Sigma, 533642), 0.8 (Figures 4H and 5B) or 1 mM (Figures
5D, 5E, and 7F) mycophenolic acid (Sigma, M5255-50MG), 100 mMpsicofuranine (Cayman Chemicals, 19574). Guanine was used at
100 mM (Sigma, 51030); pyruvate (Sigma, P5280) and a-ketobutyrate (Sigma, K0875) were used at 1 mM.
T cells isolated from spleens and lymph nodes of OT-I;Rag/ and OT-II;Ptprca;Rag2/ mice were purified using negative
selection for CD8+ or CD4+ T cells (Miltenyi, 130-104-075 and 130-104-454). Purity (80%–90%) was confirmed by flow cytometry.
Either male or female mice were taken for DC/ T cell isolation – each individual experiment was gender-matched. All cells were
cultured at 37C, 5% CO2.
METHOD DETAILS
Influenza infection
Isoflurane anesthetized 6-8 week-old female mice were infected by intranasal inoculation with 104 plaque forming units (PFU) Influ-
enza A virus H3N2 A/X-31 strain in 50 mL of sterile PBS. Disease activity was scored in a blinded to genotype and group allocation,
using the following scoring system: 5 ‒ healthy mouse; 4 ‒mouse is huddled, looks hypothermic; 3 ‒ ruffled fur, squinting eyes; 2 ‒
mouse is hunched; 1 ‒mouse looks lethargic. LEGENDplex Mouse Inflammation Panel (Biolegend, 740150) was used to determine
cytokine levels in plasma according to manufacturer’s protocol and fluorescent signals assessed by flow cytometry. Absolute
quantification was performed using standard curves generated in the assay. For analysis of influenza-specific T cells bronchoalveolar
lavage with cold PBS was performed. Influenza viral load was determined in total RNA extracted from lungs (RNeasy kit, QIAGEN,
74104), via qRT-PCR using primers for influenza M protein transcript, 50-GGACTGCAGCGTTAGACGCTT-30; and 50-CATCCTGTT
GTATATGAGGCCCAT-30 (Prabhu et al., 2013). Influenza M protein mRNA levels were normalized to mActb (50-GATG
CTCCCCGGGCTGTATT-30 and 50-GGGGTACTTCAGGGTCAGGA-30). Terminal deoxynucleotidyl transferase (TdT) dUTP nick-end
labeling (TUNEL) was performed on formalin-fixed paraffin embedded sections of lungs using the In Situ Cell Death Detection Kit,
POD (Roche, 11684817910) following manufacturer’s instructions. An Olympus CX31microscope with Nikon DS-Fi2 camera attach-
ment was used to analyze the sections.
In vitro T cell priming and restimulation
Splenic DCs were rested overnight in normal media, and BM-derived cDC1 or cDC2s were activated with 1 mg/mL LPS (O111:B4,
Sigma, LPS25) for 18 h prior to pulsing with antigen. Mature DCs were pulsed for 30 min with 1 mg/mL (splenic CD11c+ DCs) or
0.5 mg/mL (cDC1 and cDC2 BM-derived DCs) OVA257-264 peptide (SIINFEKL) (Invivogen, vac-sin), 1 mg/mL OVA323-339 peptide
(ISQAVHAAHAEINEAGR) (Invivogen, vac-isq), 1 mg/mL of ovalbumin (Sigma, A5503), or UV-irradiated bm1T-OVA MEFs (Sancho
et al., 2009) as indicated. OT-I and OT-II cells were labeled with carboxyfluorescein succinimidyl ester (CFSE; Molecular Probes,
C1157) according to manufacturer’s protocol prior to co-culture with antigen-pulsed DCs. OT-I cells and DCs were co-cultured
for 72 h at a ratio of 5:1 for cDC1 and splenic DCs; OT-II cells and DCs were co-cultured at 2:1 for 96 h. For differentiation into TEe6 Cell Metabolism 34, 106–124.e1–e10, January 4, 2022
ll
OPEN ACCESSArticleand TEM cells 72 h after priming, OT-I cells were passaged for 5 days with mIL-2 (5,000 IU/mL, Miltenyi, 130-094-055) or mIL-15
(100 IU/mL, Miltenyi, 130-094-072), respectively (Manjunath et al., 2001). Primed OT-I cells were re-seeded at the indicated time-
points at 106 cells/mL, restimulated for 5 h with plate-bound anti-CD3 (5 mg/mL, Thermo Fisher Scientific, 16-0037-81) and soluble
anti-CD28 (2 mg/mL, Thermo Fisher Scientific, 16-0289-81), and supernatants were then analyzed for IFNg release. For analysis of
intracellular cytokines, OT-II cells were rested for 3 days following 96 h co-culture with DCs prior to restimulation for 5 h with 1 mg/mL
OVA323-339 or anti-CD3/CD28. For co-culture experiments using fixed dendritic cells, mature splenic DCs were pre-loaded with
1 mg/mL of OVA257-264 for 30min and fixed for 10min in 0.01 or 0.05%of glutaraldehyde solution (Sigma, G7651). Following 2washes
in PBS, CFSE-labeled OT-I T cells were added at 5:1 ratio and co-cultured for 72 h.
Cytotoxicity assays
Cytotoxicity wasmeasured as previously described (Noto et al., 2013). Briefly, after 72 h of priming, OT-I T cells were passaged twice
at 48 h intervals inmedia containing 5,000 IU/mLmIL-2. Effector T cells were then combinedwith target cells at the indicated ratios for
5 h at 37C. Target cells were splenocytes fromwild-type mice pulsed with 1 mg/mL of OVA257-264 for 30min and labeled with 5 mMof
CFSE (CFSEhigh), which were combined at 1:1 ratio with control splenocytes labeled with 0.5 mMCFSE (CFSElow). Specific lysis was
equal to 100 – ((CFSEhigh/CFSElow) in the presence of cytotoxic CD8 T cells / (CFSEhigh/CFSElow) in the absence of cytotoxic CD8
T cells) 3 100. Supernatants were harvested for cytokine analysis by ELISA.
In vivo T cell priming
CD8+ T cells and CD4+ T cells were isolated from spleens and lymph nodes of OT-I;Rag/ andOT-II;Ptprca;Rag2/mice, respec-
tively. Following labeling with 5 mM of CFSE, 53 106 T cells/mouse were intravenously injected into FaminDDC mice and their con-
trols. 24 h later, mice were immunized with 25 mg ovalbumin intraperitoneally. Three days later, proliferation indices of splenic
T cells were calculated based on CFSE dilution and, in the case of OT-I cells, CTL activity of total splenocytes analyzed as
described above.
Flow cytometry
Cells were blocked with purified anti-CD16/CD32 antibody (Biolegend, 101302) for 30 min on ice and stained with corresponding
antibodies for cell surface molecules. For intracellular staining, samples were fixed with 2% paraformaldehyde (Thermo Fisher Sci-
entific, 28908), permeabilized using Wash/Perm buffer (BD Biosciences, 554723), and stained for intracellular cytokines using stan-
dard protocols; FOXP3 was stained with Mouse Foxp3 buffer set (Biolegend, 560409) per manufacturer’s protocol. Peptide-tetramer
staining was analyzed in whole blood obtained through tail bleeds, with antibodies and peptide-loaded tetramers added directly
to blood, and samples diluted with buffer prior to analysis by flow cytometry. Monomers for H-2Db Influenza A NP366-374
(ASNENMETM), H-2Db Influenza A PA224-233 (SSLENFRAYV), andH-2KbOVA257-264 (SIINFEKL) were obtained from the NIH Tetramer
Core Facility and tetramers prepared using Extravidin–PE (Sigma, E4011) as per standard protocol. Single cell fluorescence was
analyzed on BD LSRFortessa or Attune NxT flow cytometers. Data analysis was performed using FlowJo software v9/v10; gating
strategies for experiments are listed in Table S7.
Cytokine measurement by ELISA
Supernatants from experiments were analyzed using ELISA according to the manufacturer’s instructions (IFNg, Thermo Fisher Sci-
entific, 15501107; IFN-a, PBL Assay Science, 42120-2; IL-2, Thermo Fisher Scientific, 15133787; IL-12p70, Thermo Fisher Scientific,
12384003; granzyme B, R&D Systems, DY1865).
RNA extraction and sequencing
RNeasyMini Kit (QIAGEN, 74104) was used to extract RNA and samples were quantified with a NanoDrop ND-1000 spectrophotom-
eter (Thermo Fisher Scientific). RNA quality was assessed using the Agilent 2200 or 2100 TapeStation system (Agilent Technologies).
Libraries were prepared using TruSeq stranded mRNA library prep kit (Illumina, 20020594) in accordance with the manufacturer’s
instructions. Sequencing of libraries was performed using an Illumina NextSeq 500 platform with NextSeq 500-Mid Output kit gener-
ating 1x75 bp end reads (T cell dataset GSE147370) or 2x150bp end reads (dendritic cell dataset GSE126473). FastQ files were qual-
ity-checked (FastQC; http://www.bioinformatics.babraham.ac.uk/projects/fastqc/) and any residual adaptor sequences were
removed (TrimGalore; http://www.bioinformatics.babraham.ac.uk/projects/trim_galore/). Reads were subsequently aligned to the
appropriate reference genome (mm10, UCSC for cDCs, Ensembl Mus_musculus.GRCm38 for T cells) using HISAT2 (Kim et al.,
2015) (for cDCs) or STAR (Dobin and Gingeras, 2015) (for T cells). Analysis of differentially expressed genes was conducted on
read count files using the limma package in R with the Voom transformation (for cDCs) or edgeR (for T cells). Gene Set Enrichment
Analysis (GSEA) was undertaken using log counts-per-million (CPM) data. RNA-Seq data generated as part of this study can be
accessed at the Gene Expression Omnibus (GEO: GSE126473, GSE147370).
Immunoblot
63 106 BM-derived cDC1s were washed once in ice-cold PBS and lysed in ice-cold RIPA buffer supplemented with protease inhib-
itors (cOmplete Protease Inhibitor Cocktail, Roche). After lysing for 15min on ice, cell debris was removed by centrifugation for 15min
at 4C. Protein levels were quantified using the Pierce BCA Protein Assay Kit (Thermo Scientific) and samples were normalizedCell Metabolism 34, 106–124.e1–e10, January 4, 2022 e7
ll
OPEN ACCESS Articleto protein content before addition of 4X laemmli buffer (Bio-Rad) and boiling at 95C for 5 min. Samples were run on a 10%
SDS-PAGE gel. Proteins were transferred to a nitrocellulose membrane using a Trans-Blot Turbo transfer system before blocking
for 1 h at room temperature in 5% milk in TBS-T. Membranes were incubated with primary antibodies overnight at 4C in 5% milk
in TBS-T. These were detected by incubation with HRP-conjugated secondary antibodies for 1 h at room temperature and visualized
using 20X LumiGlo reagent (Cell Signaling).
Metabolic tracing experiments
For metabolic tracing experiments using BM-derived cDC1s, cells were isolated and stimulated for 18 h with 1 mg/mL LPS as
described above. Following this, cells were incubated for 3 h with 300 mM [13C4] malic acid (Cambridge Isotope Laboratories,
CLM-8065-0), 100 mM [13C16] palmitic acid (Sigma-Aldrich, P0500), 1 mM [
13C3] sodium pyruvate (Sigma-Aldrich, 490717),
1 mM [13C4] a-ketobutyrate (Cambridge Isotope Laboratories, CLM-6164) or 150 mM L-aspartic acid-
13C4,
15N1 (Sigma-Aldrich,
607835) supplemented into complete RPMI-1640 medium as applicable. In glutamine labeling experiments, BM-derived cDC1s
were incubated for 3 h with 2 mM [15N2
13C5] glutamine (Sigma-Aldrich, 607983), added into RPMI glutamine free medium. For
glucose labeling experiments, BM-derived cDC1s were pulsed for 1 h with 2 g/L [13C6] glucose (Sigma-Aldrich, 389374) sup-
plemented into RPMI glucose free medium. For metabolic tracing experiments utilizing splenic CD11c+ DCs, cells were isolated
and rested overnight before pre-equilibration in OptiMEM for 3 h and subsequent addition of [13C5
15N4] hypoxanthine (Cam-
bridge Isotope Laboratories, CNLM-7894-0) to a final concentration of 25 mM for 3 h prior to harvesting of supernatants and
cell extracts.
Formeasurements of systemic inosine levels plasmawas collected by cardiac puncture followed by centrifugation in EDTA-coated
tubes for 15 min at 4C at 2000 g. 20 mL aliquots were taken and prepared for LC-MS by addition of 100 uL 4:1 methanol:water fol-
lowed by vortexing and centrifugation at 20 000 g. The supernatants were then dried using a centrifugal evaporator (Savant, Thermo-
Fisher). Samples were reconstituted in 100 mL ammonium acetate containing 2 mM [13C10,
15N5] adenosine monophosphate and
adenosine triphosphate, 10 mM [13C4] succinic acid, a 1 in 5000 diluted [U
13C, U15N]mixture of amino acids (all purchased fromSigma
Aldrich) and 50 nM [13C5] inosine (Cambridge Isotope Laboratories) as internal standards. Famin
p.254I, Faminp.254V and Faminp.284R
mice were fasted for 18 h prior to harvesting.
Extraction of aqueous metabolites
After washing with PBS or 162 mM ammonium acetate adjusted to pH 7.4 (as appropriate), cell pellets were then extracted using the
2:1 chloroform:methanol method described by Folch (Folch et al., 1957) with modifications to the method as previously detailed
(Cader et al., 2020). All solvents used were HPLC or LC-MS grade and obtained from Fisher Scientific. Aqueous extracts were stored
at 80C prior to analysis.
LC-MS sample preparation
Aqueous extracts of cells were dried using a centrifugal evaporator (Savant, ThermoFisher) and reconstituted in 10 mM ammonium
acetate containing 2 mM [13C10,
15N5] adenosine monophosphate and adenosine triphosphate, 10 mM [
13C4] succinic acid, and a 1 in
5000 diluted [U13C, U15N] mixture of amino acids (all purchased from Sigma Aldrich) as internal standards. Where appropriate, inter-
nal standardswere omitted during isotopic labeling experiments to prevent contamination with labeled substrates. The sampleswere
then vortexed and sonicated for 5 min, followed by brief pulsed centrifugation to recover maximum volume.
Molecular formula determination using accurate mass and isotopic mass distribution, confirmed by authentic standard, were
used to validate identification of inosine as the top-ranking identifiable LC-MS feature of differential abundance between super-
natants of Famin+/+ and Famin/ splenic DCs. For analysis of cell culture supernatants in subsequent experiments, 20 ml of
supernatant was aliquoted directly onto a styrene 96 well plate (Corning) followed by dilution with 100 ml of 10 mM ammonium
acetate containing 50 nM [13C5] inosine (Omicron Biochemicals) or 50 nM [
15N4] inosine (Cambridge Isotope Laboratories) as
an internal standard. Where appropriate the internal standard was omitted. For absolute quantitation of both labeled and
unlabelled inosine, an inosine calibration line was prepared in the appropriate cell culture medium in the following concentra-
tions: 100 pM, 1 nM, 5 nM, 10 nM, 50 nM, 100 nM and 1 mM. These calibrants were then subjected to the same dilution and
preparation described above. All plates were sealed with a pre-slit silicone sealing mat prior to injection (Thermo Fisher
Scientific).
LC-MS analysis of aqueous metabolites
A Q Exactive Plus orbitrap coupled to a Vanquish Horizon ultra high performance liquid chromatography system was used for all the
analysis. LC-MS methodology used corresponds to the ACE C18-PFP and the Phenomenex Gemini-NX protocols described previ-
ously (Cader et al., 2020), utilizing identical chromatographic and MS parameters. The majority of analyses (for example detection of
nucleotides, nucleosides and organic acids etc.) was carried out using the ACEC18-PFP column and, where appropriate, nucleoside
phosphates were measured on a BEH amide HILIC column as detailed in Cader et al. (2020). For analysis of supernatants, where
sensitivity was critical, 10 mL was injected with the first minute of chromatography being switched to waste to prevent build-up of
matrix containing contaminants in the source of the mass spectrometer.
All solvents and additives used were LC-MS or Optima grade and obtained from Fisher Scientific or Merck.e8 Cell Metabolism 34, 106–124.e1–e10, January 4, 2022
ll
OPEN ACCESSArticleHydrazone derivatization of keto acids and hydroxy carboxylic acids in cell culture supernatants and subsequent
LC-MS analysis
An internal standard solution was prepared by extracting 100 mg of U13C lyophilized algae (Sigma) using the Folch extraction
described above. Supernatants were dried using a centrifugal evaporator (Savant, Thermo Fisher) and derivatised according to a
modified version of the protocol previously described (Han et al., 2013). Briefly, 50 ml of 75% aqueous methanol was added to the
dried culture medium followed by 10 ml of the internal standard mix. To this mixture, 30 ml of 250 mM 3-nitrophenylhydrazine
(in 50%aqueousmethanol), 30 ml of 150mM1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (in methanol) and 30 ml of 7.5%pyridine
(in 75%aqueousmethanol) were added sequentially. The resultingmixture was vortexed and samples allowed to derivatise for 1 h on
ice. Samples were subsequently quenched with 5 mg/mL butylated hydroxytoluene and 420 ml of water and centrifuged to pellet any
salts from the media.
The LC gradient employed for the separation of the hydrazone derivatives utilized a binary solvent mixture consisting of mobile
phase A, 0.1% formic acid in water and B, 0.1% formic acid in methanol and the column was an Acquity CSH C18 (100 3
2.1mm, 1.7 mm). The gradient program was as follows: 18% B was increased to 90% B in a linear gradient over 6.75 min, held at
90% for a further minute followed by re-equilibration for 1 min to give a total run time of 9 min. The flow rate was 400 ml/min and
the column oven temperature was 40C. The injection volume was 5 ml. To prevent derivatisation reagents from entering the ion
source, a switch was employed for the first 2 min of the gradient program. Samples were run in negative ion mode using MS
parameters previously described (Cader et al., 2020).
LC-MS data processing
All data were acquired using Xcalibur (Version 4.1, Thermo Fisher Scientific). Targeted processing was carried out using Xcalibur and
unbiased analysis using Compound Discoverer (Version 2.1 or Version 3.1, Thermo Fisher Scientific). Untargeted analysis utilized
data from both positive and negative ionization modes. Chromatogram peaks for each differential metabolite were manually verified
using XCalibur (Version 4.1, Thermo Fisher Scientific) and identities validated using the high-resolution m/z METLIN database
(Scripps Research Institute). To confirm identification of inosine and in cases of ambiguity, compound retention times were
validated against known external standard solutions. For all cellular and serummetabolite analysis, target peak areas corresponding
to metabolites were normalized to total ion content unless otherwise indicated. For absolute quantitation of inosine in supernatants,
normalization of target peaks was performed with reference to internal standards, and quantitation performed with reference to a
calibration line between 10 pM and 1 mM prepared in the appropriate sample matrix. Relative quantitation of metabolite levels in
supernatant tracing studies were not normalized.
All sample data were processed using Compound Discoverer (Version 2.1 and Version 3.1, Thermo Fisher Scientific) to accurately
calculate total ion content for use as a normalization factor. For labeling studies, incorporation into specific compounds was deter-
mined by accurate mass shift of +1.0034 and +0.9970 for 13C and 15N respectively. Endogenous levels of 13C and 15N compounds of
interest were determined by reference to control samples pulsedwith unlabelled compounds of investigation, and endogenous levels
subtracted from quantified isotopomers in the labeled samples as applicable.
Determination of extracellular acidification rate and oxygen consumption rate
Bone marrow-derived cDC1 stimulated overnight with LPS (O111:B4, Sigma-Aldrich, LPS25), or isolated splenic DCs matured
overnight, were seeded prior to analysis at 3 3 105 cells per well on Poly-L-lysine-coated plates (Sigma-Aldrich, P8920), as indi-
cated. Cells were then washed twice and incubated for 1 h in XF assay medium (unbuffered DMEM pH 7.4 with 10 mM glucose,
100 mM sodium palmitate and 2mM L-glutamine) in a non-CO2 incubator at 37
C as per manufacturer’s instructions (Agilent Tech-
nologies). Measurements of extracellular acidification rate (ECAR) and oxygen consumption rate (OCR) were determined using an
XF-96 Extracellular Flux Analyzer (Agilent Technologies). Serial measurements were obtained under basal conditions and
following addition of 1 mM oligomycin (Sigma-Aldrich, 75371), 1.5 mM FCCP (Sigma-Aldrich, C2920) and 100 nM rotenone with
1 mM antimycin A (Sigma-Aldrich, R8875 and A867). For determination of glycolysis, ECAR measurements were obtained under
basal conditions.
Cytoplasmic pH assay
Intracellular pH was compared using pHrodo Red AM (Thermo Fisher Scientific, P35372) fluorogenic probe for measurement of
cytoplasmic pH according tomanufacturer’s protocol. In brief, splenic DCswerematured overnight and incubated with 5 mMpHrodo
for 30min at 37C in a non-CO2 incubator in Hank’s Balanced Salt Solution (HBSS) and washed once before fluorescence signal was
measured using a microplate reader (Tecan infinite M1000, Tecan Group or CLARIOstar plus, BMG Labtech) with an excitation/
emission of 560/580nm.
Antigen uptake assay
LPS-treated cDC1s or splenic DCs were incubated with 0.05 mg/mL of OVA-AF647 (Thermo Fisher Scientific O34784) in HBSS
containing HEPES at 37C in a non-CO2 incubator for 30 min, while control cells were kept on ice to account for passive diffusion.
Following washes in ice cold HBSS, samples were prepared for analysis by flow cytometry as described above. For analysis of
splenic cDC1s and cDC2s cells were gated on CD11c+MHC II+CD8+CD11b and CD11c+MHC II+CD11b+CD64, respectively.Cell Metabolism 34, 106–124.e1–e10, January 4, 2022 e9
ll
OPEN ACCESS ArticleEndosome-to-cytosol uptake assay
Assay was performed as previously described (Cebrian et al., 2011), with some modifications using the LiveBLAzer FRET-B/G
Loading Kit with CCF4-AM (Thermo Fisher Scientific, K1095). LPS-treated BM-derived DCs were loaded with 1 mM of CCF4-AM
at room temperature, followed by incubation with 2 mg/mL b-lactamase (Sigma, P0389) for the indicated time. Cells were subse-
quently analyzed by flow cytometry, and blue-to-green (excitation 405 nm, emission 450 nm/525 nm). FRET ratio was used as an
indicator for efficiency of antigen export into the cytosol. Response ratios were calculated as per manufacturer’s instructions and
normalized to the signal intensity of control cells, which had been incubated on ice.
MHC I recycling assay
The rate of MHC I recycling was assessed as previously described (Belabed et al., 2020). LPS-treated cDC1s were incubated in the
presence of 1 mM sodium a-ketobutyrate (Sigma-Aldrich, K0875), or control for 18 h. Cells were blocked with anti-CD16/CD32 anti-
body (Biolegend, 101302) and subsequently incubated with anti-H-2Kb-FITC (Biolegend, 114605) for 30 min on ice. To enable inter-
nalisation, cells were incubated for 30min at 37C in complete RPMI-1640mediumwith 1mM sodium a-ketobutyrate replenished as
applicable. After washing in 1% BSA-PBS, cells were subsequently incubated in stripping buffer (0.5 M NaCl, 0.5% acetic acid, pH
3.0) for 10 min on ice. After washing with ice cold PBS, cells were fixed in 2% paraformaldehyde (to determine basal MHC I staining
after internalisation step) or re-incubated in complete pre-warmed RPMI-1640 medium (with 1 mM sodium a-ketobutyrate replen-
ished if applicable) for 15 min or 30 min to allow MHC I recycling. After re-incubation, cells were incubated in stripping buffer for
10 min on ice, washed in ice cold PBS and fixed in 2% paraformaldehyde followed by analysis by flow cytometry. The difference
in mean fluorescence intensity after re-incubation was determined to calculate the%MHC I recycled at each time point as described
previously.
siRNA transfection
Freshly isolated splenic CD11c+ DCs were transfected with 30 pmol/sample of siRNA (purchased from Dharmacon, Horizon
Discovery) using aMouseDendritic Cell Nucleofector Kit (Lonza, VVPA-1011) andNucleofector 2b. Cells were rested for 36-48 h prior
to assays (antigen uptake or co-culture). RNA for knockdown validation was extracted using RNeasy Plus Micro Kit (QIAGEN),
reverse transcribed using M-MLV Reverse Transcriptase (Thermo Fisher) and SYBR Green Q-PCR (Eurogentec) was performed
using QuantStudio 7 Flex (Thermo Fisher). For primer sequences see Key resources table.
T cell activation assays in presence of DC-secreted soluble factor
Splenic CD11c+ DCs were rested overnight in RPMI, and on the next day cultured in OptiMEM or RPMI for 3 h. Cell-free supernatants
were harvested and frozen immediately or centrifuged using 3 kDa cut-off spin columns. OT-I or OT-II T cells isolated from spleens
and LNs were seeded at 105 cells/mL and stimulated for 72 h with plate-bound anti-CD3 (5 mg/mL, 16-0037-81, Thermo Fisher
Scientific) and soluble anti-CD28 (2 mg/mL, 16-0289-81, Thermo Fisher Scientific) in presence of supernatants harvested from
DCs, soluble inosine with or without 100 nM SCH 58261 (Sigma Aldrich, S4568) or 0.5 mM CGS-21680 (Sigma Aldrich, C141). For
T cell RNA-seq experiments, naive OT-I T cells were cultured in the presence of 2 h splenic DC supernatant and stimulated for 24
h with anti-CD3/CD28 prior to RNA extraction.
Tumor xenograft model
2.53 104 LL2-OVA cells (for experiments comparing Faminp.254I, Faminp.254V, Faminp.284R mice) or 23 104 LL2-OVA cells (for exper-
iment comparing Faminp.254V and Faminp.284R mice) in PBS were subcutaneously injected into the left flank of gender- and age-
matched mice of 6-10 weeks of age. Tumor growth was assessed at least every other day in a fully blinded fashion, using a calliper
for both the long (L) and short (S) dimensions, and tumor volume calculated using the equation volume = (L 3 S2)/2.
QUANTIFICATION AND STATISTICAL ANALYSIS
Statistical analyses were performed using Graphpad Prism 6.0 /8 /9 or, and as described in LC-MS analysis methods, Compound
Discoverer 2.1 / Compound Discoverer 3.1 (Thermo Scientific). Unless otherwise stated, statistical significance was calculated as
appropriate using unpaired, two-tailed Student’s t test or ordinary one-way ANOVA and Tukey post hoc test as described in the figure
legends. Formal statistical determination of whether the data met assumptions of the approach was not undertaken. Grubbs’ test
was used to identify outliers within datasets. Where indicated, FDR-adjusted p values were calculated using the Benjamini-Hochberg
procedure. All in vivo experiments were performed in a blindedmanner. Data are represented asmean and standard error of themean
(SEM). A P value of < 0.05 was considered significant.e10 Cell Metabolism 34, 106–124.e1–e10, January 4, 2022