Articles
Investigating sex differences in T regulatory cells from
Lancet Rheumatol 2022;
4: e710–24
Published Online
August 31, 2022
https://doi.org/10.1016/
S2665-9913(22)00198-9
See Comment page e652
*Senior authors
Centre for Rheumatology
Research (G A Robinson PhD,
C Ciurtin PhD, Prof E C Jury PhD),
Centre for Adolescent
Rheumatology Versus Arthritis
(G A Robinson, J Peng MSc,
H Peckham MSc, C Ciurtin,
Prof E C Jury), and Centre for
Cardiometabolic and Vascular
Science
(Prof I Pineda-Torra PhD),
Division of Medicine,
University College London,
London, UK; Department of
Paediatric and Adolescent
Endocrinology, University
College London Hospital and
Great Ormond Street Institute
of Child Health, University
College London, London, UK
(Prof G Butler MD); Gender
Identity Development Service,
Tavistock and Portman NHS
Foundation Trust, London, UK
(Prof G Butler)
Correspondence to:
Prof Elizabeth C Jury, Centre for
Rheumatology Research,
Division of Medicine, University
College London, London,
W1CE 6JF, UK
e.jury@ucl.ac.uk
e710
cisgender and transgender healthy individuals and patients
with autoimmune inflammatory disease: a cross-sectional
study
George A Robinson, Junjie Peng, Hannah Peckham, Gary Butler, Ines Pineda-Torra*, Coziana Ciurtin*, Elizabeth C Jury*
Summary
Background Sexual dimorphisms, which vary depending on age group and pubertal status, have been described across
both the innate and adaptive immune system. We explored the influence of sex hormones on immune phenotype in
the context of adolescent health and autoimmunity.
Methods In this cross-sectional study, healthy, post-pubertal cisgender individuals (aged 16–25 years); healthy, pre-
pubertal cisgender individuals (aged 6–11 years); transgender individuals (aged 18–19 years) undergoing gender-
affirming treatment (testosterone in individuals assigned female sex at birth and oestradiol in individuals assigned
male sex at birth); and post-pubertal cisgender individuals (aged 14–25 years) with juvenile-onset systemic lupus
erythematosus (SLE) age-matched to cisgender individuals without juvenile-onset SLE were eligible for inclusion.
Frequencies of 28 immune-cell subsets (including different T cell, B cell, and monocyte subsets) from each participant
were measured in peripheral blood mononuclear cells by flow cytometry and analysed by balanced random forest
machine learning. RNA-sequencing was used to compare sex and gender differences in regulatory T (Treg) cell
phenotype between participants with juvenile-onset SLE, age-matched cis-gender participants without the disease,
and age matched transgender individuals on gender-affirming sex hormone treatment. Differentially expressed genes
were analysed by cluster and pathway analysis. Suppression assays assessed the anti-inflammatory function of Treg
cells in vitro.
Findings Between Sept 5, 2012, and Nov 6, 2019, peripheral blood was collected from 39 individuals in the post-
pubertal group (17 [44%] cisgender men, mean age 18·76 years [SD 2·66]; 22 [56%] cisgender women, mean age
18·59 years [2·81]), 14 children in the cisgender pre-pubertal group (seven [50%] cisgender boys, mean age 8·90
[1·66]; seven [50%] cisgender girls, mean age 8·40 [1·58]), ten people in the transgender group (five [50%]
transgender men, mean age 18·20 years [0·47]; five [50%] transgender women, mean age 18·70 years [0·55]), and
35 people in the juvenile-onset SLE group (12 [34%] cisgender men, mean age 18·58 years [2·35]; 23 [66%]
cisgender women, mean age 19·48 [3·08]). Statistically significantly elevated frequencies of Treg cells were one of
the top immune-cell features differentiating young post-pubertal cisgender men from similarly aged cisgender
women (p=0·0097). Treg cells from young cisgender men had a statistically significantly increased suppressive
capacity in vitro compared with those from cisgender women and a distinct transcriptomic signature significantly
enriched for genes in the PI3K–AKT signalling pathway. Gender-affirming sex hormones in transgender men and
transgender women induced multiple statistically significant changes in the Treg-cell transcriptome, many of
which enriched functional pathways that overlapped with those altered between cisgender men and cisgender
women, highlighting a hormonal influence on Treg-cell function by gender. Finally, sex differences in Treg-cell
frequency were absent and suppressive capacity was reversed in patients with juvenile-onset SLE, but sex
differences in Treg-cell transcriptional signatures were significantly more pronounced in patients with juvenile-
onset SLE compared with individuals without juvenile-onset SLE, suggesting that sex hormone signalling could be
dysregulated in autoimmunity.
Interpretation Sex-chromosomes and hormones might drive changes in Treg-cell frequency and function. Young
post-pubertal men have a more anti-inflammatory Treg-cell profile, which could explain inflammatory disease
susceptibilities, and inform sex-tailored therapeutic strategies.
Funding Versus Arthritis, UK National Institute for Health Research University College London Hospital Biomedical
Research Centre, Lupus UK, and The Rosetrees Trust.
Copyright © 2022 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0
license.
www.thelancet.com/rheumatology Vol 4 October 2022

Research in context
Evidence before this study
Sex differences in the innate and adaptive immune response
result in differential responses to infection, vaccination, and
susceptibility to autoimmune diseases. Studies have reported
different sex specific observations by age group regarding
inflammatory profiles and disease presentation. This is
particularly relevant for juvenile-onset systemic lupus
erythematosus (SLE), which predominantly effects young
women and has a common onset at puberty. We searched
PubMed, Web of Science, and Google Scholar for research
articles published between Jan 1, 1990, and April 30, 2022,
using the search terms “sex”, “gender”, “transgender”, “sex
hormones”, “sex chromosomes”, “inflammation”, “immune
response”, “(juvenile-onset) systemic lupus erythematosus”,
and “autoimmunity”. We also searched for research articles
published in the same time window in the top rated
rheumatology, immunology, and endocrine-specific journals by
impact factor. Published abstracts were excluded from the
searches. The earliest referenced article was published in 2003.
We found that investigating the role of sex and gender in
immunological research is aiding the understanding of sexual
dimorphisms in the immune system in both healthy individuals
and in diseases with a sex bias. However, disparities across
published results have emerged, probably because of
differences in study age groups, pubertal status, gender
consideration, and use of different animal models.
Introduction
Men and women differ in their inflammatory response to
non-self and self antigens; consequently, they have
different responses to vaccination and risks of infection.1
This has been highlighted by the COVID-19 pandemic2
and historically by the risk developing autoimmune
diseases: men are generally more likely to have a more
severe disease course following SARS-CoV-2 infection,
whereas women represent about 80% of patients with
autoimmunity.1,3 Sex differences across the immune
system, including fundamental differences in the
frequency and activity of T-cell subsets, have been
described by gender across ethnicities.1,4 Generally,
previous observations show higher CD4+ T-cell and lower
CD8+ T-cell frequencies4,5 and higher proinflammatory,
cytotoxic, and antiviral T-cell responses in women
compared with men. Of note, some sex differences in
adaptive immunity are present throughout life whereas
others manifest after the onset of puberty or post-
reproductive senescence, implicating both genetic and
sex hormonal influences.7
Systemic lupus erythematosus (SLE) is a complex
autoimmune disorder characterised by loss of immune
cell regulation and the presence of autoantibodies against
nuclear and non-nuclear antigens, resulting in chronic
inflammation and organ damage. Sex hormones have
been hypothesised to be implicated in the cause of SLE
www.thelancet.com/rheumatology Vol 4 October 2022
Articles
Added value of this study
To our knowledge, this study is the first to report an in-depth
analysis of immune-cell phenotype comparing sex and gender
differences between young (aged 14–25 years) post-pubertal
cisgender men and cisgender women both with and without
juvenile-onset SLE, and in transgender individuals undergoing
gender-affirming sex hormone treatment who did not have
juvenile-onset SLE. During adolescence, rapid changes in sex
hormones drive changes in inflammatory responses and disease
susceptibility. Therefore, this age range is important in the
context of autoimmune research. We found unique changes in
the immune profile, specifically in regulatory T cells, between
cisgender men and cisgender women and sex hormone-
associated transcriptomic profiles that overlap by gender, which
could play a role in juvenile-onset SLE pathogenesis.
Implications of all the available evidence
The evidence from this study could be used to improve our
understanding of sexual dimorphisms in immune responses by
sex to improve sex-specific and gender-specific therapeutic
strategies in health and disease and to understand sex
differences in inflammatory responses, vaccine efficacy, and
autoimmune disease susceptibility. We also highlight specific
immune features and genes that could be targeted
therapeutically or used for sex-specific diagnosis and
investigation of juvenile-onset SLE pathogenic mechanisms.
due to significant sex bias (nine female patients to every
one male patient) and frequent disease onset during
childbearing age (15–44 years). Between 15% and 20% of
all patients with SLE have juvenile-onset disease (onset
before the patient reaches 18 years old). Juvenile-onset
SLE often starts at puberty and has a more severe disease
phenotype than adult onset disease.8 Together, this
highlights a key role for sex hormones in susceptibility to
juvenile-onset SLE.
Investigating the relationship between sex hormones
and inflammation is important for understanding the
cause of autoinflammatory diseases. Many clinical trials
do not consider sex or gender in recruitment and
outcome measures:9 historically less than 10% of
immunological studies analysed data by sex.10 The
absence of consideration of sex in research is particularly
important for the study of autoimmune diseases, which
are more prevalent in women.3 We aimed to investigate
the influence of sex chromosomes and hormones in
driving sexual dimorphisms in inflammatory profiles
across different sex, gender, age, and disease status
groups.
Methods
Study design and participants
In this cross-sectional study healthy, post-pubertal
cisgender individuals (aged 16–25 years); healthy,
e711
 Articles
pre-pubertal cisgender individuals (aged 6–11 years);
transgender individuals (aged 18–19 years) undergoing
gender-affirming treatment (testosterone in individuals
assigned female sex at birth and oestradiol in individuals
assigned male sex at birth); and post-pubertal cisgender
individuals (aged 14–25 years) with juvenile-onset SLE age-
matched to individuals without juvenile-onset were
eligible for inclusion. Transgender individuals were
contacted for participation via University College London
Hospital’s (London, UK) young people’s Gender Identity
Development Service liaison endocrine clinic, and
participants with juvenile-onset SLE were contacted for
inclusion via the Rheumatology Clinic at University
College London Hospital. Patients were eligible to be
included if they had juvenile-onset SLE according to the
American College of Rheumatology revised classification
criteria for lupus (1997) or the Systemic Lupus Inter­
national Collaborating Clinics (SLICC) criteria (2012), and
were diagnosed before the age of 18 years. For cisgender
inviduals without juvenile-onset SLE, volunteers who did
not have the disease or who had non-inflammatory, non-
infective conditions (eg, referred for assessment of non-
inflammatory musculoskeletal conditions) were recruited.
Additional inclusion criteria for post-pubertal cohorts was
a puberty Tanner stage of 4-5. Any patient who withheld
consent or whose carer withheld consent for inclusion
(as appropriate given patient’s competence) and any
patient who withdrew from the study were excluded from
the analyses.
Ethics approval for the study was obtained
from the London-Harrow Research Ethics Committee
(REC11/LO/0330). Informed written consent was
See Online for appendix 1 acquired from all participants (appendix 1 p 1).
Procedures
Peripheral blood mononuclear cell (PMBC) samples were
taken from cisgender individuals with and without
juvenile-onset SLE and transgender individuals without
juvenile-onset SLE. The samples were assessed for
28 immune-cell subsets (including different T-cell, B-cell,
and monocyte subsets) by multiparameter flow cytometry,
as described previously11 and in appendix 1 (pp 2–4).
To assess the suppressive capacity of regulatory T (Treg)
cells in post-pubertal cisgender individuals with and
without juvenile-onset SLE and transgender individuals
without juvenile-onset SLE, fluorescence-activated cell
sorting (FACS) was used to sort Treg cells, responder
T (Tresp) cells, and monocytes for in-vitro suppression
assays (appendix 1 p 5) using stored PBMCs from study
participants collected during the research project. Treg-
cell purity was confirmed by FOXP3 intracellular staining
(appendix 1 p 5).
RNA sequencing was used to assess differentially
expressed genes (DEGs) and gene ontology pathways
between different sexes and genders with and without
juvenile-onset SLE. Total RNA was isolated from FACS-
sorted Treg cells (appendix 1 pp 5–6). Quality control
e712
analysis and sequencing was done by UCL Genomics
(University College London, London, UK). Libraries were
prepped with the NEB Low input kit (New England
Biolabs, Ipswich, MA, USA) and were sequenced on the
NovaSeq SP flow-cell (Illumina, San Diego, CA, USA)
with a 100 base pair single read. Reads were demultiplexed
and converted to fastq files with Illumina’s bcl2fastq
conversion software (version 2.19) and analysis was done
by UCL Genomics. DEGs that were significantly different
between cohorts, some of which were expanded using
first-order protein–protein interaction networks, were
analysed by hierarchical clustering (Pearsons), sparce
partial least squares discriminant analysis, open target
disease association analysis, and gene ontology pathway
analysis to investigate differentially regulated immune
functions and associations in Treg cells (appendix 1 p 6).
Statistical analysis
For immunophenotyping, the analysis was exploratory
based on limited sample availability, as a result of
working with rare cohorts of young individuals and
patients. For RNA sequencing, a sample estimation was
carried out using a 90% power calculation (p<0·05) of
data from a Treg-cell suppression assay (Treg-cell
functional readout), using the proportion of Tresp cells
suppressed in cisgender men versus cisgender women
at a 1:1 Treg-cell to Tresp-cell ratio, yielding the required
sample size of five participants per group (appendix 1
p 5). This indicated the number of samples needed to
see functional transcriptomic changes in Treg cells by
sex. Statistical analyses were done on the basis of the
hypothesis that men and women have altered immune
cell profiles, including Treg-cell phenotype, resulting in
variation in inflammatory disease risk by sex and
gender. Statistical analyses were done using GraphPad
Prism (version 9). Demographic, clinical, immune
frequency, DEG normalised count, and Treg-cell
suppressive capacity measures data were tested for
normal distribution. Normally distributed data were
assessed with parametric tests; data that were not
normally distributed data were assessed with non-
parametric tests. For demographic, clinical, immune
phenotype, gene expression, and Treg-cell suppression
continuous data, unpaired t tests were used when
comparing two groups and one-way ANOVA (Turkey’s
post-hoc test) tests were used when comparing more
than two groups. For continuous data that were not
normally distributed, Mann-Whitney tests were used
when comparing two groups. For binary data Fisher’s
exact tests were used to compare two groups and χ² tests
were used when comparing more than two groups.
p values less than 0·050 were considered significant
unless mentioned otherwise. Multiple testing was
accounted for using the false discovery rate adjustment
for multiple comparisons Benjamini, Krieger, and
Yekutieli method to calculate p values for immune
phenotype data t tests. The balanced random forest
www.thelancet.com/rheumatology Vol 4 October 2022
 Articles

Cisgender and Cisgender and p value Cisgender and Cisgender and p value
transgender transgender transgender transgender
men women men women

Cisgender post-puberty (Continued from previous column)

Number of participants 17 22 ·· Tanner Stage at time of sample
Age (years) 18·76 (2·66) 18·59 (2·81) 0·85† Tanner stage 1 0 0 >1·00*
Ethnicity Tanner stage 2–3 0 0 >1·00*
White 8 (47%) 12 (55%) 0·75* Tanner stage 4–5 12 (100%) 23 (100%) >1·00*
Asian 6 (35%) 4 (18%) 0·14* Clinical feature
Black 1 (6%) 2 (9%) 1·00* Age at disease onset 13 (5·37) 12 (3·00) 0·32†
Other 2 (12%) 4 (18%) 0·68* (years)
Tanner Stage at time of sample SLEDAI 2·67 (2·87) 3·21 (3·18) 0·62†
Tanner stage 1 0 0 >1·00* Neurological 1 (8%) 1 (4%) 1·00*
involvement
Tanner stage 2–3 0 0 >1·00*
Serositis involvement 2 (17%) 6 (26%) 0·31*
Tanner stage 4–5 17 (100%) 22 (100%) >1·00*
Cutaneous involvement 10 (83%) 21 (91%) 0·69*
Taking contraception 0 2 (9%) 0·50*
(combined pill) Haematological 5 (42%) 9 (39%) 1·00*
involvement
Cisgender pre-puberty
Musculoskeletal 9 (75%) 21 (91%) 0·31*
Number of participants 7 7 ·· involvement
Age (years) 8·90 (1·66) 8·40 (1·58) 0·68† Renal involvement 4 (33%) 5 (22%) 0·69*
Ethnicity Erythrocyte 14·71 (14·22) 25·26 (32·93) 0·23†
White 4 (57%) 3 (43%) >1·00* sedimentation rate
Asian 2 (29%) 3 (43%) >1·00* dsDNA titre 53·5 9 (3–350) 0·46†
Black 1 (14%) 1 (14%) >1·00* (4·75–138·8)
Other 0 0 >1·00* C3 0·93 (0·29) 1·053 (0·35) 0·28†
Tanner Stage at time of sample Lymphocyte count 1·74 (0·8) 1·68 (0·6) 0·78†
Tanner stage 1 7 (100) 7 (100%) >1·00* Treatment
Tanner stage 2–3 0 0 >1·00* Hydroxychloroquine 11 (92%) 20 (87%) 1·00*
Tanner stage 4–5 0 0 >1·00* Mycophenolate mofetil 5 (42%) 13 (57%) 0·49*
Transgender on gender affirming hormones Prednisolone 5 (42%) 12 (52%) 0·72*
Number of participants 5 5 ·· Prednisolone dose (mg) 4·00 (6·34) 3·83 (4·57) 0·72*
Age (years) 18·20 (0·47) 18·70 (0·55) 0·16† Vitamin D 2 (17%) 6 (26%) 0·19*
Ethnicity Methotrexate 1 (8%) 2 (9%) 1·00*
White 5 (100%) 5 (100) >1·00* Azathioprine 4 (33%) 3 (13%) 0·20*
Asian 0 0 >1·00* Data are mean (SD), n (%), or median (IQR). For transgender individuals, the
Black 0 0 >1·00* Tanner stage was their most recent Tanner stage before puberty blocking therapy.
Tanner stage 1 is classified as pre-puberty, Tanner stages 2–3 were classified as in
Other 0 0 >1·00* the larche or gonadarche, and Tanner stages 4–5 were classified as post-puberty.
Tanner Stage at time of puberty block For patients with juvenile-onset SLE, common clinical measures of disease are
Tanner stage 1 0 0 >1·00* shown as well as treatments. Rituximab treatment was avoided in the cohort.
C3=complement component 3. dsDNA=anti-double-stranded-DNA antibodies.
Tanner stage 2–3 1 (20%) 2 (40%) >1·00* SLE=systemic lupus erythematosus. SLEDAI=systemic lupus erythematosus
Tanner stage 4–5 4 (80%) 3 (60%) >1·00* disease activity index. *p value obtained with Fisher’s exact test. †p value
Time on gender affirming 11·7 (11·08) 12 (7·16) 0·96† obtained with unpaired t test for normally distributed data and Mann-Whitney
hormone treatment test for not normally distributed data.
(months)
Table: Demographic and clinical characteristics
Cisgender post-puberty with juvenile-onset SLE
Number of participants 12 23 ··
Age (years) 18·58 (2·35) 19·48 (3·08) 0·39†
machine-learning approach was used to assess the
Ethnicity extent of the difference in the global immune phenotype
White 3 (25%) 10 (43%) 0·46* between healthy, post-pubertal cisgender men and
Asian 6 (50%) 7 (30%) 0·29* cisgender women and to highlight the key immune cell
Black 3 (25%) 4 (17%) 0·67* variables that were driving this difference. Receiver
Other 0 2 (9%) 0·54* operator characteristic (ROC) analysis and 10-fold cross-
(Table continues in next column) validation were used to assess the balanced random
forest model performance and classification accuracy,
which was optimised and adjusted for age and ethnicity
(appendix 1 p 6).

www.thelancet.com/rheumatology Vol 4 October 2022 e713

 Articles

Role of the funding source mean age 19·48 [3·08]; late puberty Tanner stage 4–5).
The funding sources had no role in the study design, Participant characters are reported in the table.
data collection, analysis, interpretation, or writing of the Following optimisation and adjustment for age and
report. ethnicity, ROC analysis of the balanced random forest
model showed a strong classification accuracy (84·76%)
Results and good model efficiency in discriminating cisgender
Between Sept 5, 2012, and Nov 6, 2019, peripheral blood men from cisgender women by immune phenotype
was collected from 39 individuals in the cisgender post- (figure 1A, B). From this analysis, the predictive
pubertal group (17 [44%] cisgender men, mean age sensitivity was 82·61% and specificity was 82·25% for
18·76 years [SD 2·66]; 22 [56%] cisgender women, mean the BRF model, and the 10-times cross-validation
age 18·59 years [2·81]; late puberty Tanner stage 4–5; classification accuracy held a steady measure of 77·40%.
appendix 1 p 1), 14 children in the pre-pubertal cisgender The top contributing immune-cell features segregating
group (seven [50%] cisgender boys, mean age 8·90 years cisgender men from cisgender women were
[1·66]; seven [50%] cisgender girls, mean age 8·40 years CD4+CD25+CD127– Treg cells and CD4+CD25–CD127+
[1·58]; pre-puberty Tanner stage 1), ten individuals in the Tresp cells (figure 1C). This was confirmed by analysis of
transgender group (five [50%] transgender men, mean the frequency and absolute number of Treg cells—which
age 18·20 years [0·47]; five [50%] transgender women, were significantly increased (frequency p=0·0097;
mean age 18·70 years [0·55]; mid to late puberty Tanner absolute number p=0·041)—and Tresp cells—which
stage 2–5 at the time of puberty arrest), and 35 individuals were significantly reduced (frequency p<0·0001; absolute
in the juvenile-onset SLE group (12 [34%] cisgender men, number p=0·0021)—in cisgender men compared with
mean age 18·58 years [2·35]; 23 [66%] cisgender women, cisgender women (figure 1D–F; appendix 1 p 7). This

A B C D
1·0

CD4+ Tresp*
CD8+ EMRA
CD4+ EMRA
Cisgender women

CD8+ Naive
CD4+ Naive

CD4+ Treg*
0·4 Tresp

CD8+ CM
CD4+ CM

CD8+ EM
CD4+ EM
Cisgender men
Out of bag 0·8 Treg

CD8+
CD4+
Error rate CD4 CM
0·3 Cisgender men=0·2353 Bm1
0·6
Sensitivity

Cisgender women=0·1364 Bm2

Error

Late Bm5

*
*
0·4
0·2 Area under CD19 Unswitched memory
the curve=0·8476 Bm2 Translational
0·2 10-fold cross- CD19 Naive Cisgender men vs Cisgender women
0·1 validation=0·7740 CD14 (T cell subsets)
0
0 2500 5000 7500 10 000 1·0 0·8 0·6 0·4 0·2 0·0 0 0·5 1·0 1·5 2·0 2·5 3·0 3·5 p value
Trees Specificity Mean decrease Gini 0·0 0·05 0·3

E F G H
Cisgender men
g

g
g

g
in

in
in

in
di -

di -
vid

vid

Cisgender women
vid

vid
n

p<0·0001 p=0·0021
No

No
Di

Di

80 0·35 50 p=0·032
0·30 0:1
70
Proportion of Tresp suppressed

Unstim
0·25 40
Proportion of CD4+ T cells

60
0·20 1:1
Cell count (×109)/L

50
Treg:Tresp ratio

p=0·0097 0·15 p=0·041 30

40
1:2 p=0·088
8 20
0·025
6 1:4 p=0·033
0·020
4 0·015 10
2 0·010 0:1
0 0·005 0
Tresp Treg Tresp Treg –103 0 103 104 105 –103 0 103 104 105 1:1 1:2 1:4
Cell trace violet Cell trace violet Treg:Tresp ratio

Figure 1: Comparison of immune-cell subsets in young-post pubertal cisgender men and women
Comparison of 28 immune-cell subsets in 17 young, post-pubertal cisgender men versus 22 young, post-pubertal cisgender women using the balanced random forest model approach. (A) out of bag
error rate of the balanced random forest was 0·1795 (82·05% accuracy). (B) Receiver operator characteristic curve analysis used to validate the model, providing an area under the curve of 0·8476
(84·76% accuracy; 82·61% sensitivity; and 81·25% specificity), with 10-fold cross-validation classification accuracy of 77·40%. (C) The top ten variables contributing to the balanced random forest:
higher mean decrease in Gini score represents a higher importance of the variable. (D) Heatmap of p values comparing T-cell subset immunophenotyping in cisgender men versus cisgender women.
Cumulative cell frequency (E) and absolute counts (F) Tresp-cell (CD4+CD25–CD127+) and Treg-cell (CD4+CD25+CD127–) frequencies; data are mean (SE). (G) Treg-cell mediated suppression of activated
Tresp cells in four cisgender men and four cisgender women detected using cell trace violet and flow cytometry following 72 h activation using soluble anti-CD3 and anti-CD28 in the presence of
monocytes. Treg-cell to Tresp-cell ratios of 1:1 1:2, 1:4, and 0:1 were assessed (appendix 1 p 5); each leftward peak represents a round of proliferation. (H) Suppressive capacity of Treg cells at varying
Treg-cell to Tresp-cell ratios in cisgender men compared with cisgender women, calculated using the fold change of proportion of Tresp-cell proliferation with Treg cells (1:1, 1:2, and 1:4) compared
with Tresp cells without Treg cells (0:1); data are mean (SE). CM=central memory. EM=effector memory. EMRA=effector memory re-expressing CD45RA. Treg=regulatory T. Tresp=responder T.
*Significant p value following 10% false discovery rate adjustment for multiple comparisons.

e714 www.thelancet.com/rheumatology Vol 4 October 2022

 Articles

balance was reflected in the Treg-cell to Tresp-cell ratio, compared with Treg cells from young post-pubertal
which was significantly higher in cisgender men cisgender women (figure 1G, H; appendix 1 p 5). Thus,
compared with cisgender women (appendix 1 p 7). Of Treg cells from young cisgender men and women are
note, no significant differences in the frequency of other numerically and functionally distinct.
CD4+ and CD8+ T-cell subsets were seen (figure 1D). To investigate functional differences in Treg cells by
Treg cells from young cisgender men had a significantly sex, RNA sequencing, and DEGs from isolated Treg cells
higher capacity to suppress activated Tresp cells in vitro from young post-pubertal cisgender men and women

A B

LOC101928710
LOC389834

LINC00278
KCNQ1OT1
CCDC144A

MZF1-AS1
PPP1R2P9

SLC25A29

CATSPERG

FAM166A

FAM19A2
C9orf135
CLEC11A
MCOLN2

MAP7D2

SAMD12
FAM43A
ADGRA3

TMSB4Y
NLGN4Y

ZSWIM4
GPANK1
SH2D1B

CARD14
PNLDC1

TXLNGY
BCORP1
CXorf38

ZNF648
TOP1P1

TTTY10

TTTY15
TRIM46
SORCS3

RPS4Y1
CFAP58

KDM5D
KDM6A

SMAD5
HDHD1
ERCC6L

HECW2
EIF1AX

NR4A1
DDX3X

MPZL2
TXLNG

DDX3Y
EIF1AY
TIMD4
EIF2S3

TIAM2

PDGFB

USP9Y
BRCA2
SAR1B

ERAP2
FBLN5

AFAP1
CORIN
ZRSR2

KLRB1

WNK4
TCF19
SYAP1

MC1R
ESAM

DTX1
ASB9
NKRF
PRKX
CCR5

PRKY
GYPE

ESR1
TSIX
XIST

IRS1

UTY
ZFX
10–160

ZFY
JPX
DDX3Y
RPS4Y1 XIST
10–120 KDM5D
Cisgender men

10–80 UTY
TTTY15
TXLNGY
10–40 PRKY USP9Y
Split axis
Log10 p value

EIF1AY
10–30
ZFY
PPP1R2P9
10–20 BCORP1
LINC00278 TSIX
Cisgender women

10–8 NLGN4Y
Split axis
TMSB4Y JPX
10–6 FBLN5
TTTY10 ERCC6L
HECW2
10–4 SMAD5 NKRF
EIF1AX
ESAM CCR5
10–2 TCF19 NR4A1 SLC25A29 p=0·010
10–0
–15 –10 –5 0 5 10 15

Cisgender men Cisgender women 2 1 0 –1

Log2 fold change

C E
NR4A1 ESR1 IRS1 PDGFB
120 p=0·0015 80 p=0·0049 100 p=0·0068 200 p=0·0095
Gene ontology pathway

Histone lysine demethylation

Normalised tcounts

80 150
60
PI3K–AKT signalling pathway 80
60
40 100
Translational initiation 40
40
20 20 50
Renal system process
0 0 0 0
0·0 0·5 1·0 1·5 2·0 2·5 3·0 3·5 4·0 Cisgender Cisgender Cisgender Cisgender Cisgender Cisgender Cisgender Cisgender
men women men women men women men women
–Log10 p value

D F RPS4Y1 EIF2S3 EIF1AX G FOXP3

p<0·0001 p=0·00086 p<0·0001 p=0.33
40 Translation 5000 10000 1000 5000
initiation
Nuclear receptor transcription pathway 4000
8000 800 4000
30 ERa genomic pathway 3000
Enrichment

6000 600 3000

PI3K-Akt
2000
20
Normalised counts

signalling 10
pathway Oestrogen signalling 4000 400 2000
8
pathway
10 6
4 2000 200 1000
Cellular response
2
to hormone stimulus
0 0 0 0 0
25 50 75 100 Cisgender Cisgender Cisgender Cisgender Cisgender Cisgender Cisgender Cisgender
–Log10 p value men women men women men women men women

Figure 2: Comparison of Treg-cell transcriptomic profiles between young post-pubertal cisgender men and women
FACS-sorted Treg cells (CD4+CD25+CD127–) from five young, post-pubertal cisgender men and five young, post-pubertal cisgender women were analysed by RNA-sequencing and whole genome
expression was compared by sex. (A) Volcano plot displaying log2 fold changes and log10 p values of DEGs between cisgender men and women, where coloured points represent statistically significant
DEGs below p value threshold of 0·01). (B) Hierarchical clustering heatmap (Clustvis, Pearson’s) of normalised gene counts of statistically significantly altered DEGs between cisgender men and
women. (C) Pathway analysis plot displaying the –log10 p values of enriched DEG pathway ontology terms of Treg cells between cisgender men and women using the 82 statistically significant DEGs.
(D) Pathway analysis plot displaying the p value and enrichment ratios of enriched DEG pathway ontology terms of Treg cells between cisgender men and women using the extended gene list,
incorporating both original seed genes and the genes from respective protein–protein interaction networks; the size of the points are relative to the number of genes contributing to that pathway. Box
and whisker plots displaying Treg-cell gene expression by normalised counts of genes associated with the PI3K–AKT signalling (E), altered translation initiation pathways (F), and FOXP3 expression (G);
data are mean (SE). DEG=differentially expressed gene. ER=oestrogen receptor alpha. Treg=regulatory T. Tresp=responder T.

www.thelancet.com/rheumatology Vol 4 October 2022 e715

 Articles

A B C Cisgender Transgender D Cisgender Transgender

10–9 women men men women
10–7
p=0·033 TLR5
10–8
(proportion of CD4+ T cells)

p=0·017 10–6 MTRNR2L1

8 11 10–7
p=0·032 IRS2
10–5

Log10 p value

Log10 p value
10–6
Treg frequency

CXCR4 SOCS2
7 9 HCP5 IER5 CCR5
10–5 10–4 NR4A2 CCR3
SLC25A29 FOSL2 BEND3P3
7 10–4 ADARB2 CTSL TCF19 10–3 SLC25A29
6 CD161
GTF2H4 DRP2
10–3 TEKT4P2 NR4A1
CFAP58 TSPAN12 10–2
SYT6 NR4A1 p=0·010
5 5 10–2 p=0·010
10–1 10–1

4 3 1 1
sg n –10 –5 0 5 10 –10 –5 0 5 10

sg en

sg n

en
r

Tr womer

wo der
sg ys

rls
de

de

de
Ci me

de
an me
d

m
Log2 fold change Log2 fold change
Ci bo

gi
en

en

en

en

en

en
sg

sg

(vs cisgender women) (vs cisgender men)

Ci

Ci

an

E Tr F G
Cytokine-mediated signalling Transgender men Cisgender women Transgender men Cisgender women
Nitric oxide metabolic process
Gene ontology pathway

OAS2
Cytokine secretion IL17RB
HLA-B
Adaptive immune system LTA
NOD2
Cell recognition NOD2
α-amino acid biosynthetic process CD36
CD36
Response to fibroblast growth factor IRS1
Lymphocyte mediated immunity IL32
TLR5
Extracellular structure organisation SPATA2
HLA-DQA1
Gliogenesis NLRP3
IL17RB
0 1 2 3 4 5
–log10 p value -1 0 1 2

H I J

FOXO signaling pathway Transgender women Cisgender men Cisgender men Transgender women
Corticotropin-releasing hormone response TNFSF14
Gene ontology pathway

BCL6
CDKN1A ZFP36L2
Regulation of cell growth CDKN1B EMILIN1
Regulation of cell activation RGS2 IRS2
H3F3B NR4A3
Positive regulation of catabolic process CXCR4 LGALS9C
SGK1 TNFAIP3
Negative regulation of kinase activity ADNP2 CD83
Cellular senescence PTCH2 BCL6
FBP1 CDKN1A
Interleukin-7 signaling SUPV3L1 DUSP10
CISH ZBTB16
Epithelial cell proliferation FBLN5 SOX13
Regulation of macrophage migration SOCS2 CD9
NEDD4L SIRPB1
0 1 2 3 4 5
-log10 p value

K
TLR5 NLRP3 CD36 NOD2 LTA IL32
p<0·0001 p=0·0085 p=0·00087 p=0·0044
500 40 100 p=0·0013 40 40 000 p=0·0054
250
Normalised counts

400 30 80 200 30 30 000

300 60 150
20 20 20 000
200 40 100
100 10 20 10 10 000
50
0 0 0 0 0 0
Tr wo er

Tr wo er

Tr wo er

Tr wo er
Tr wo er

Tr wo er

sg en

sg en

sg en

sg en
sg en

sg en

r
r

en

en

en
en

en

en
de

de

de

de
de

de

d
d

an m

an m

an m

an m
an m

an m

m
m

m
en

en

en

en
en

en

en

en

en

en
en

en
sg

sg

sg

sg

sg
sg
Ci

Ci

Ci

Ci

Ci

Ci

L CD83 BCL6 NR4A3 CXCR4 ZFP36L2 SOCS2

p=0·0080 p=0·00099 p<0·0001 p<0·0001

800 p=0·0029 400 40 15 000 15 000 300
p=0·0070
Normalised counts

600 300 30
10 000 10 000 200
400 200 20
5000 5000 100
200 100 10

0 0 0 0 0 0
en

en

en

en

en

en
wo der

wo er

wo der

wo er

wo der

wo der
en

en

en

en

en

en
nd

nd
rm

rm

rm

rm

rm
m

m
en

en

en
ge

ge

ge
er
de

de

de

de

de
sg

sg

sg
d

ns

ns

ns
en

en

en

en

en

en
an

an

an
a

a
sg

sg

sg

sg

sg

sg
Tr

Tr

Tr

Tr

Tr

Tr
Ci

Ci

Ci

Ci

Ci

Ci

e716 www.thelancet.com/rheumatology Vol 4 October 2022


were analysed. 82 genes were differentially regulated in
Treg cells between cisgender men and women
(figure 2A, B; appendix 2); there were 50 upregulated
DEGs in cisgender men and 32 upregulated DEGs in
cisgender women. The most significantly altered genes
were located on the X or Y chromosomes: 16 (32%) DEGs
in cisgender men and 18 (56%) DEGs in cisgender
women (appendix 1 p 8). The other DEGs were expressed
on other chromosomes typically shared equally by sex
(appendix 1 p 8), suggesting both sex chromosomes and
hormones could affect Treg-cell transcriptomes.
Gene ontology pathway enrichment analysis of the
original 82 DEGs was done and validated using an
extended gene list based on known protein–protein
interactions (figure 2C, D; appendix 1 p 8). This
highlighted and validated many significantly enriched
functional pathways in Treg cells that differ by sex,
including altered PI3K–AKT signalling, nuclear receptor
transcription, translation initiation, oestrogen receptor
signalling, and cellular response to hormone stimulus.
Of note, several key node genes from the protein–
protein interaction network (appendix 1 p 8) were
exclusive to the PI3K–AKT signalling pathway and were
increased in cisgender men compared with cisgender
women, including NR4A1, ESR1, IRS1, and PDGFB
(figure 2E; appendix 1 pp 9–10), supporting increased
Treg-cell functionality in cisgender men. Other node
genes from the extended network were associated with
altered translation initiation pathways, including
EIF2S3, EIF1AX (X-linked), and RPS4Y1 (Y-linked);
these were increased in a respective sex chromosome-
dependent way (figure 2F; appendix 1 pp 9–10). No
significant difference in Treg cell FOXP3 expression
Figure 3: Assessment of Treg-cell functional gene expression and frequency
in cisgender and transgender individuals
Violin plots displaying circulating Treg-cell (CD4+CD25+CD127–) frequencies
between seven pre-pubertal cisgender boys and seven cisgender girls, assessed
by t test (A), and 17 post-pubertal cisgender men, 22 cisgender women,
five transgender men, and five transgender women, assessed by one-way
ANOVA (B); measured by flow cytometry; data are mean (SE). FACS-sorted Treg
cells (CD4+CD25+CD127–) from five young, post-pubertal cisgender men and
five cisgender women and five transgender men and five transgender women
were analysed by RNA-sequencing and whole genome expression was compared
by gender. Volcano plots displaying log2 fold changes and log10 p values of DEGs
between transgender men and cisgender women (C) and transgender women
and cisgender men (D): the red and blue points represent statistically
significantly DEGs below p value threshold (p<0·01). Cluster significance –log10
p values of enriched DEG pathway ontology terms of Treg cells between
transgender men and cisgender women (E). Hierarchical clustering heatmap
(Clustvis, Pearson’s) of normalised gene counts of DEGs found in the functional
genetic pathway ontology terms in for transgender men versus cisgender
women (F, G). Cluster significance –log10 p values of enriched DEG pathway
ontology terms of Treg cells between transgender women versus cisgender men
(H). Hierarchical clustering heatmap (Clustvis, Pearson’s) of normalised gene
counts of DEGs found in the functional genetic pathway ontology terms in for
transgender women versus cisgender men (I, J). Box and whisker plots of Treg-
cell gene expression by normalised counts of genes associated with the (K)
cytokine signalling, and (L) cell growth and activation pathways; data are mean
(SE). ANOVA=analysis of variance. DEGs=differentially expressed genes.
FACS=fluorescence-activated cell sorting. Treg=regulatory T. Tresp=responder T.
www.thelancet.com/rheumatology Vol 4 October 2022
Articles
was identified between cisgender men and cisgender
women, despite its gene location on the X chromosome
(figure 2G). See Online for appendix 2
To explore the specific influence of sex hormones on
the sexually dimorphic Treg-cell profiles, we used unique
cohorts of cisgender, pre-pubertal children and young
transgender individuals, who had their physiological
puberty blocked followed by supplementation of gender-
affirming sex hormones specific to their target gender
(table; appendix 1 p 11).
Pre-pubertal cisgender boys had increased Treg-cell
frequency compared with pre-pubertal cisgender girls
(figure 3A; p=0·017). Furthermore, gender-affirming sex
hormone therapy had an influence on Treg-cell
frequencies: Treg cells were significantly increased in
both cisgender men and transgender men compared
with cisgender women, with a similar trend in cisgender
men and transgender men compared with transgender
women (figure 3B), suggesting that dimorphism in
circulating Treg cell numbers might be driven by both
sex hormones and underlying sex chromosomes.
In addition, large transcriptomic changes in Treg-cell
gene expression were observed in both transgender men
(treated with testosterone) compared with cisgender
women (91 genes; figure 3C) and transgender women
(treated with oestradiol) compared with cisgender men
(235 genes; figure 3D; appendix 2), highlighting sex
hormone-driven effects over chromosome effects.
Pathway enrichment analysis highlighted key functional
pathways in Treg cells that were altered by sex hormones
in transgender individuals (figure 3E–J), including
upregulated cytokine mediated signalling and cytokine
secretion pathways in transgender men (figure 3E–G)
and downregulated cell growth and activation pathways
in transgender women (figure 3H–J). Many of these
genes have individually been associated with Treg-cell
function (figure 3K, L; appendix 1 pp 9–10).
The influence of sex hormones on Treg-cell gene
expression was also compared between cisgender men
and women and transgender men and women (figure 4A).
Two genes were altered in all gender-unique comparisons:
NR4A1 and SLC25A29 (figure 4B; appendix 1 pp 9–10),
and 24 genes overlapped between at least two gender-
unique comparisons (figure 4A; appendix 1 p 11). We were
able to cluster gender-unique groups using hierarchical
clustering (appendix 1 p 11) and sparse partial least
squares discriminant analysis (figure 4C). The top ranked
genes driving clustering in component 1 (highest
contributing component) were TIAM2 and MZF1-AS1
(figure 4D; appendix 1 p 11). Of the 62 DEGs unique to the
original cisgender analysis that did not overlap with
transgender comparisons, 32 (52%) were X chromosome
or Y chromosome linked (figure 4A). Of the remaining
DEGs that did overlap with transgender comparisons,
five genes (BCORP1, MAP7D2, SLC25A29, NR4A1, and
KLRB1) were validated as being significantly altered
between transgender men and transgender women,
e717
 Articles

A B
MZF1-AS1, CARD14, CCR5 NR4A1 SLC25A29
Cisgender women BCORP1, FAM166A, TIAM2 p=0·0034
vs cisgender men MAP7D2, LOC101928710 p=0·0010
IRS1, CFAP58 FBLN5, HECW2, ESAM p=0·0032
62 120 160 p<0·0001
CD161, DTX1 p=0·0068
TRIM46, TCF19 Transgender p=0·0015
LOC389834 women 120 p<0·0001

Normalised counts
7 11 vs cisgender 80
2 men p=0·0048
Transgender 218 80
4
men 78
40
vs cisgender 40
women

0 0
MTUS1, GNLY
NR4A1,

wo der

wo der
en

sg n

en

sg n

sg n

en
Tr om r

Tr om r
an en

an en
de

de

de

de
SIRPG-AS1, GZMB

w de

w de
an me

an me
m

m
en

en

en

en

en

en
SLC25A29

en

en
sg

sg

sg

sg
sg
Ci

Ci
Ci

Ci
Tr

Tr
C D TIAM2 MZF1-AS1
8
Cisgender women p=0·0024 p=0·0064
6 p=0·0011 p=0·0021
80 200
Component 2 (17·1 %)

4
Normalised counts

60 150
2 Cisgender men
40 100
0
Transgender men
–2 20 50
Transgender women
–4 0 0
–6 –4 –2 0 2 4
r

wo der

wo der
sg n

sg n

en

sg n

sg n

en
Tr om r

Tr om r
an en

an en
de

de

de

de
w de

w de
e

an me

an me
m

m
Component 1 (23·2 %)
en

en

en

en

en

en
en

en
sg

sg

sg

sg
Ci

Ci
Ci

Ci
Tr

Tr
E F
Cisgender man Transgender man Cisgender woman Transgender woman p=0·042

Dividing Non- Dividing Non- Dividing Non- Dividing Non- p=0·0099

dividing dividing dividing dividing Proportion of Tresp suppressed
50 p=0·0074
0:1
Unstim 40
Treg:Tresp ratio

30
1:1
20

0:1 10

0
Cell trace violet Cell trace violet Cell trace violet Cell trace violet
Ci end en

n
de men

en
Tr end me
sg r m

om
sg r wo
er
e

rw
Tr end

e
sg

en
Ci

an

sg
an

Figure 4: Sex and gender specific transcriptomic and functional changes in Treg cells
Cohort details and gender-specific terminology descriptions are reported in appendix 1 (p 1). FACS-sorted Treg cells (CD4+CD25+CD127–) from five young, post pubertal
cisgender men, five cisgender women, five transgender men, and five transgender women were analysed by RNA-sequencing and whole genome expression was
compared by gender. (A) Overlap of DEGs (p<0·01) from Treg cells between different group comparisons: cisgender men versus cisgender women, transgender men
versus cisgender women, and transgender women versus cisgender men. (B) Treg-cell gene expression by normalised counts of genes that overlapped in all gender
comparisons; data were assessed by one-way ANOVA and are mean (SE). (C) Sparse partial least squares discriminant analysis plot clustering each gender group using
normalised gene counts of the overlapping genes (n=24); individual distribution points and confidence ellipses (ovals) are plotted for each gender group. (D) Treg-cell
gene expression by normalised counts of the top ranked loaded genes for component 1 (appendix 1 p 11); data were assessed by one-way ANOVA and are mean (SE).
Treg-cell mediated suppression of activated Tresp cells. Treg cells, Tresp cells, and monocytes were isolated from four young, post-pubertal cisgender men, four
cisgender women, three transgender men, and four transgender women. (E) Proliferation of Tresp cells at varying Treg-cell to Tresp-cell ratios in individuals of different
genders. (F) Suppressive capacity of Treg cells at a Treg-cell to Tresp-cell ratio of 1:1 between different gender groups, calculated using the fold change of the percentage
of Tresp-cell proliferation with Treg cells (1:1) compared with Tresp-cell proliferation without Treg cells (0:1); data were assessed with one-way ANOVA and are mean
(SE; appendix 1 p 5). DEGs=differentially expressed genes. FACS=f luorescence-activated cell sorting. ANOVA=analysis of variance. Treg=regulatory T. Tresp=responder
T. Unstim=Unstimulated.

e718 www.thelancet.com/rheumatology Vol 4 October 2022


supporting a role of sex hormones in driving their
expression (appendix 1 p 11; appendix 2).
Analysis of DEG lists using first-order protein–protein
interaction extended network gene lists from cisgender
and transgender comparisons revealed significant
overlap of functional pathways associated with cell
signalling by both secondary messengers and
interleukins (appendix 1 p 12). Despite these functional
transcriptomic changes, Treg cells from cisgender men
were more suppressive in vitro than were Treg cells from
all other individuals (figure 4E, F), suggesting that both
sex chromosomes and hormones might play a role in the
suppressive functions of Treg cells (appendix 1 p 13).
SLE is an autoimmune disease with a 90% female sex
bias that is characterised by defective Treg function.12 Of
note, a significant association was identified by open
target analysis between the original 82 Treg-cell genes
differentially expressed between cisgender men and
women (figure 2A) and genes previously associated with
SLE in public databases (p=0·020). This association was
driven by ten genes (ESR1, CCR5, TIMD4, DTX1, BRCA2,
KDM6A, PDGFB, MC1R, NR4A1, and IRS1;
appendix 1 pp 9–10), and supports a role for these genes
in sex specific SLE susceptibility. Of note, KDM6A is
X chromosome linked, and DTX1, IRS1, CCR5, and
NR4A1 were regulated by testosterone, oestradiol, or both
in transgender individuals on gender-affirming sex
hormone treatment (figure 4A; appendix 1 p 11). Of the
gender associated and SLE associated genes, CCR5 and
ESR1 (increased in cisgender women), and PDGFB and
MC1R (increased in cisgender men) are already
established drug targets for HIV (CCR5); breast cancer,
infertility, and obesity (ESR1); macular degeneration
(PDGFB); and erythropoietic protoporphyria and kidney
injury (MC1R; appendix 1 p 11).
Sex differences in Treg-cell frequencies were absent in
juvenile-onset SLE, probably due to the significant increase
in Treg-cell frequency in cisgender women with juvenile-
onset SLE (figure 5A). Of note, there was no significant
difference in clinical measures or treatments between
cisgender men or cisgender women with juvenile-onset
SLE (table 1), nor was there a significant effect of these
features on Treg-cell frequencies in the combined juvenile-
onset SLE cohort (appendix 1 p 13). This was observed
despite multiple patients being treated with glucocorticoids
(17 [49%] of 35 patients), hydroxychloroquine (31 [89%]
patients), or methotrexate (three [9%]; appendix 1 p 13).
Additionally, Treg cells from patients with juvenile-onset
SLE were more suppressive in cisgender women compared
with cisgender men, the opposite of that seen in individuals
who do not have juvenile-onset SLE, probably due to the
significant loss of Treg-cell suppressive function in
cisgender men with juvenile-onset SLE (figure 5B).
When Treg-cell gene expression was compared between
cisgender men with juvenile-onset SLE and cisgender
women with juvenile-onset SLE, 411 DEGs were
identified (figure 5C; appendix 2) compared with
www.thelancet.com/rheumatology Vol 4 October 2022
Articles
82 sex specific genes in cisgender individuals who did
not have juvenile-onset SLE (figure 2A; appendix 2). Of
these genes, 223 were increased in cisgender women
with juvenile-onset SLE compared with 188 in cisgender
men with juvenile-onset SLE. Only 27 genes overlapped
between the cisgender men and cisgender women in the
juvenile-onset SLE and healthy post-pubertal groups
(appendix 1 p 14); all 27 overlapping genes were X or Y
chromosome linked. Gene ontology pathway analysis of
the 384 genes altered by sex that were unique to the
juvenile-onset SLE group showed significant enrichment
of major histocompatibility complex class II antigen
presentation, membrane trafficking, protein-DNA
complex assembly, histone 3 lys 27 methylation, and ras
protein signal transduction pathways (appendix 1 p 14).
Furthermore, sex differences in Treg-cell functional
pathways were more pronounced in patients with
juvenile-onset SLE compared with those who did not
have the disease (appendix 1 p 14).
Of note, sex differences in the expression of CCR5,
DTX1, and NR4A1 (genes that were significantly altered
between cisgender men and cisgender women who did
not have juvenile-onset SLE, were affected by gender
affirming treatment hormones in transgender individuals
and were associated with SLE by open target analysis),
were absent in patients with juvenile-onset SLE (figure 5D).
By contrast, sex differences in sex-chromosome linked
genes (eg, KDM6A), that were associated with SLE by open
target analysis, were maintained in patients with juvenile-
onset SLE (figure 5E). In addition, comparing patients
with juvenile-onset SLE with individuals who did not have
the disease by sex showed very little overlap of DEGs
between cisgender men and cisgender women, suggesting
that sex hormones might play a differential role in juvenile-
onset SLE pathogenesis and disease presentation by sex
(figure 5F; appendix 1 p 13). This was validated by pathway
enrichment analysis, in which genes unique to juvenile-
onset SLE pathogenesis in cisgender men represented
mRNA splicing and cytokine signalling, whereas those in
cisgender women represented innate immune responses,
cellular scenenscence, and toll-like receptor signalling
(figure 5G). As expected, shared pathways included a
strong upregulation of interferon signalling.
Discussion
Our study identified key sex differences in Treg-cell
phenotype and function between healthy individuals,
which might explain differences in autoimmune disease
susceptibilities and response to infection. Specifically,
we identified that the global immune profile was altered
by sex, with circulating Treg cells more numerous and
suppressive in young post-pubertal cisgender men
compared, with cisgender women; Treg cells had an
altered transcriptomic profile between young cisgender
men and cisgender women (beyond sex chromosomes)
associated with increased secondary messenger
signalling in cisgender men; sex hormones altered the
e719
 Articles

A Cisgender men
B Cisgender men with Cisgender women with p=0·0285 p=0·031
Cisgender women juvenile-onset SLE juvenile-onset SLE
p=0·0069 p=0·0242
p=0·0089 Dividing Undivided Dividing Undivided
Figure 5: Sex differences in 0:1 p=0·032 p=0·088
8 p=0·0097 50
(proportion of CD4+ T cells)

Treg cells in patients with Unstim

40

Tresp suppressed
juvenile-onset SLE 1:1

Proportion of
6
Treg frequency

(A) Cumulative cell frequency 30

flow cytometry data of Treg 4 1:2
20
cells (CD4+CD25+CD127–) 15
1:4
comparing 17 young post- 2 10
pubertal healthy cisgender 0:1 5
men, 12 cisgender men with 0 0
Healthy Participants –103 0 103 104 105 –103 0 103 104 105

rti cip y

ju tici cipa y
juvenile-onset SLE, 22 young

se h

se th
ni nt ts

ni nt ts
LE

E
Pa arti alth

h
on it

SL
Pa part ealt
ju cip an

ve pa n
on wi
participants with

tS
Cell trace violet Cell trace violet

le- s w
p He

t
le- s
post-pubertal healthy

H
juvenile-onset

i
ve a
cisgender women, and SLE

r
23 cisgender women with 1:1 1:2
juvenile-onset SLE; data were Treg:Tresp ratio
assessed by t test and are
mean (SE). (B) Treg cell- C D E
mediated suppression of 188 223 CCR5 DTX1 NR4A1 KDM6A
activated Tresp cells in five upregulated upregulated (oestradiol) (testosterone) (oestradiol and (X chromosome)
cisgender men and five 10–160 testosterone) p=0·0022
cisgender women with XIST p=0·0008
juvenile-onset SLE, and DDX3Y 800 200 150 2500 p=0·014
p=0·0019 p=0·0059 p=0·0005
four cisgender men and 10–120
Normalised counts

KDM5D
four cisgender women 600 150 p=0·001 p=0·0015 2000
RPS4Y1 p=0·0008
without juvenile-onset SLE; 100
10 –80
TXLNGY
data were assessed with t test TTTY15 400 100 1500
and are mean (SE). (C) Log2 USP9Y
UTY
fold changes and log10 p values 10–40 Split axis 50
of DEGs between 200 50 500
EIF1AY
five cisgender men and PRKY
five cisgender women with 10–30 0 0 0 0
Log10 p value

juvenile-onset SLE, coloured

Ci r wo n

en er n
r w en
en

Ci r wo en

en er n
r w en
en

Ci r wo n

en er n
r w en
en

Ci r wo n

en er n
r w en
en
e
Ci en me

Ci sgen me

e
Ci sgen me

e
Ci sgen me
de m

de m
om

de m

de m
om

de m

de m
om

de m

de m
om
ZFY
points represent statistically FMN1
en er

en er

en er

en er
10–20
sg d

sg d

sg d

sg d

sg d

sg d

sg d

sg d
Ci sgen

Ci sgen

Ci sgen

Ci sgen
significantly upregulated DEGs BCORP1 PPP1R2P9
s g

NLGN4Y
Ci

Ci

Ci

Ci
below p value threshold LINC00278 TSIX
(p<0·01). Treg-cell gene 10–10 Split axis
expression by normalised GPSM3
PLA2G4C
F G
counts of hormone (gender) 10–8 DDX3X Pathways enriched in cisgender women with juvenile-onset SLE
TMSB4Y
(D) or sex chromosome KDM5C Regulation of innate immune response
10–6 LTA HBB Cellular senescence
specific genes (E) that were RNF5 GTF2H4 Regulation of toll-like receptor signalling pathway
significantly associated with 10 –4
MIR6723 Toll-like receptor cascades
HBA2 Regulation of viral life cycle
SLE by open target analysis 10–2 Regulation of nervous system development
(five per group); data were p=0·010 Regulation of cell adhesion
10
0
15 Deubiquitination
assessed with one-way ANOVA Regulation of activin receptor signalling pathway
–15 –10 –5 0 5 10
and are mean (SE); open bars 192
Negative regulation of cell population proliferation
Log2 fold change
are participants without 0 1 2 3 4 5 6 7
juvenile-onset SLE and closed Cisgender women Pathways enriched in cisgender men and cisgender women with juvenile-onset SLE
bars are participants with Cisgender Cisgender Interferon α and β signalling
men women Neutrophil extracellular trap formation
juvenile-onset SLE. (F) Overlap Type II interferon signalling
of DEGs (p<0·01) from Treg Response to interferon-α
Network map of SARS-CoV-2 signalling pathway
cells comparing patients with 44 Extrafollicular B-cell activation
juvenile-onset SLE with Response to interferon β
Regulation of pattern recognition receptor signalling
healthy participants, stratified Non-genomic actions of 1,25 dihydroxyvitamin D3
by sex (cisgender men and Response to bacterium
Cisgender men
cisgender women). (G) 0 5 10 15 20 25
Pathway analysis bar charts of 214
Pathways enriched in cisgender men with juvenile-onset SLE
enriched pathway ontology mRNA splicing, via spliceosome
terms from sex-unique or Cytokine signalling in immune system
Epstein-Barr virus infection
overlapping DEGs. Systemic lupus erythematosus
ANOVA=analysis of variance. Extrafollicular B cell activation
Glial cell activation
SLE=systemic lupus Cognition
erythematosus. Cell activation
Embryonic cranial skeleton morphogenesis
DEGs=differentially expressed Positive regulation of interferon-α production
genes. Treg=regulatory T. 0 2 4 6 8 10 12
Tresp=responder T. –Log10 p value
Unstim=unstimulated.

e720 www.thelancet.com/rheumatology Vol 4 October 2022


Treg-cell frequency and transcriptomic functional profile
between cisgender men and women, validated using
transgender individuals undergoing gender-affirming
sex hormone therapy; and differences in Treg frequency
and function were absent or altered in an age-matched
post-pubertal cisgender patients with juvenile-onset
SLE. The information reported here might help us to
understand immunopathological mechanisms of
sexually dimorphic autoimmune disease development
and contribute to basic understanding of immunology
by sex and gender.
We showed that use of broad-spectrum immune
phenotyping with machine learning can be used to
classify sex differences with confidence. To our
knowledge, no other studies have used machine learning
to address innate and adaptive immunological differences
between men and women, particularly surrounding
prediction accuracy as a method to describe how different
these sexual dimorphisms are. However, other studies,
including our own, have addressed this from an
autoimmunity standpoint, comparing individuals with
and without specific diseases (eg, juvenile-onset SLE
and juvenile idiopathic arthritis). These studies of similar
immune landscapes found that a balanced random
forest model had a strong classification accuracy:
90·9% (4·8% higher than this study) for classifying
patients with juvenile-onset SLE,11 and 89·6% (6·1% higher
than our study) for classifying patients with juvenile
idiopathic arthritis13 from healthy controls. For identifying
patients with juvenile-onset SLE, there was a diagnostic
sensitivity of 89·6% (7·0% higher than this study) and
specificity of 82·1% (0·2% lower than this study). This
highlights that the extent of immune dimorphisms
observed between healthy cisgender men and women
are not too dissimilar to those observed between people
who have and those who do not have juvenile-onset SLE,
supporting our hypothesis of altered autoimmune risk
by sex. Treg cells were one of the top contributing
immune-cell features responsible for segregating
cisgender men from cisgender women using machine
learning. This supports previous observations of human
sexual dimorphism in Treg-cell frequency across
different post-pubertal age groups;1,14 however, mouse
studies of Treg cells by sex are contradictory, probably
due use of disease models and examining organ-specific
rather than peripheral blood Treg cells.14 We found that
young post-pubertal cisgender men had increased Treg-
cell frequency and suppressive function compared with
cisgender women, suggesting a more anti-inflammatory
circulating immune profile. These characteristics could
contribute to the better response of women to vaccination
and infection that has been observed (highlighted in the
COVID-19 pandemic, in which men had an increased
risk of mortality)1,2 and the increased susceptibility of
women to autoimmune disease compared with men.3
Despite these observations, multiple studies have
reported a key role of oestradiol in supporting the
www.thelancet.com/rheumatology Vol 4 October 2022
Articles
expansion15,16 and suppressive capabilities17,18 of Treg cells.
Treg-cell numbers have been shown to increase during
the menstrual cycle when oestradiol concentrations are
highest before ovulation,19 suggesting a Treg-cell response
to sex hormones. By contrast, men who received
gonadotropin-releasing hormone agonists, a puberty
blocker which significantly reduces testosterone
concentrations, have lower circulating Treg cells than do
men who received placebo.20 As observed in cisgender
men, we found that transgender men had increased
Treg-cell frequencies following puberty blocking and
early-stage gender-affirming testosterone administration
(mean 12 months). Observed sex differences in Treg-cell
frequencies in pre-pubertal children matched those seen
in older participants post-puberty suggesting that sex
hormones and chromosomes both play dominant roles
in driving changes in Treg-cell frequency by sex. Women
with Turner syndrome, characterised by the presence of
one normal X chromosome and a missing or structurally
abnormal second one, have an increased frequency of
Treg cells compared with women with two intact X
chromosomes, suggesting a more suppressive role of the
X chromosome regarding Treg-cell proliferation.21
Despite these observations, no differences in Treg-cell
frequencies have been reported in infants (who have sex
chromosomes but very low concentrations of sex
hormones) from birth to 1 year of age;22 therefore, the
influence of the X chromosome on the number of
circulating Treg cells might occur after the initial years of
development in children.
Treg-cell FOXP3 expression was not altered by sex in
this study, despite its location on the X chromosome and
previous reports of increased mRNA expression in
cisgender men compared with women in an older cohort
with a wider age range.23 Our data suggests that differences
in Treg-cell function by sex might be maintained via
different mechanisms in younger individuals. Transcrip­
tomic analysis identified that the PI3K signalling pathway
was increased in young cisgender men compared with
young cisgender women in our study. PI3K signalling can
support Treg-cell metabolism and function; however,
overstimulation can reverse this in vitro.24 Inhibiting
PI3K signalling in vivo results in reduced Treg-cell
frequency and suppressive capacity,25 supporting a role for
PI3K signalling in the increased suppressive capacity of
Treg cells in young men compared with young women.
Of note, secondary messenger signalling pathways were
associated with hormone changes in transgender
individuals by extended pathway analysis. A gene of
interest from these pathways with increased expression in
cisgender men and transgender men was NR4A1 (and
NR4A3 in transgender men); these receptors are crucial
for Treg-cell function,26 highlighting a therapeutic target to
control autoimmune pathogenesis in women. Despite
this, Treg cells in young cisgender men were more
suppressive than those in transgender men and cisgender
and in transgender women, suggesting that both sex
e721
 Articles
hormones and sex chromosomes might play a role in
driving increased suppresive function of Treg cells in
cisgender men. Treg cells from women with Turner
syndrome have impaired suppressive functions compared
with women with two intact X chromosomes.21 A previous
Treg-cell transcript­omics study combined a list of genes
into a specific human Treg-cell signature, using previously
published datasets, to produce a subsequent microarray
test of 62 genes.27 Of these genes, we only identified
a single gene, CCR5, that was significantly altered between
cisgender men and cisgender women. PECAM1 was the
only altered gene in transgender men compared with
cisgender women in our study. Our findings suggest that
we might have identified multiple novel genes associated
with Treg-cell function by sex. Pfoertner and colleagues27
analysed the combined data from six women and five men
(aged 26–58 years) which might have substantially affected
the results, as highlighted by our observations of sexual
dimorphisms in Treg cells.
Although SLE is more common in females, sexual
dimorphisms in SLE are important to consider because
sex-specific clinical features have been observed in adults
with SLE. Men are affected by more severe renal
manifestations and higher end-stage renal disease risk,
requiring increased monitoring in clinical practice.28 Of
note, DEGs from Treg cells between young cisgender
men and cisgender women were significantly associated
with SLE, some of which were altered in transgender
individuals. Treg cells have been implicated in SLE
pathogenesis and disproportionate T helper 17-cell to Treg-
cell ratios, resulting in a proinflammatory phenotype.29
However, these studies did not address sex differences
and the populations included less than 7% men. It is also
reported that women with SLE have increased plasma
oestradiol and decreased testosterone concentration
compared with women without SLE, which could account
for differences in Treg-cell function; however, this study
did not take into consideration the individual’s stage of
menstrual cycle.23 We showed that Treg cells were more
numerous in patients with juvenile-onset SLE compared
with individuals without juvenile-onset SLE in cisgender
women, but not in men; therefore, Treg-cell frequency
was not significantly different by sex in patients with
juvenile-onset SLE. This increase in Treg-cell frequency in
women with juvenile-onset SLE could represent an anti-
inflammatory pathway aimed to counter the over-riding
inflammation. We also showed that Treg cells from young
cisgender women with juvenile-onset SLE had a signif­
icantly higher suppressive capacity compared with those
from cisgender men with juvenile-onset SLE, the opposite
of that seen in healthy individuals. However, this
suppressive capacity of Treg cells from cisgender women
with juvenile-onset SLE was still much lower than that
observed in cisgender men and similar to that observed in
cisgender women who did not have the disease. These
findings suggest that the underlying function of Treg cells
was not improved in cisgender women with juvenile-
e722
onset SLE, despite the increased number of cells. This
could explain why anti-inflammatory feedback mech­
anisms are often not substantial enough to control
disease; thus, therapies that promote Treg-cell function
and number could still hold promise in juvenile-onset
SLE for both sexes. We also identified strong sex
differences in Treg-cell transcriptomic profiles in juvenile-
onset SLE and highlighted both sex specific (mRNA
splicing in cisgender men and immune regulation in
cisgender women) and common (interferon signalling)
pathogenic processes, which together might relate to sex
differences in clinical outcomes for patients. Of note, we
did not find an association between oestrogen-enriched
and interferon-enriched pathways in the cohort of patients
with juvenile-onset SLE in our study, which has been
described in older patients (mean age 38·5 [SD 15·3]) with
SLE.33 Finally, histone lysine demethylation was a key
pathway altered between young cisgender men and
cisgender women, supporting previous evidence
surrounding the role of female-biased epigenetic alteration
of T-cell function associated with SLE.31 Together, this
supports a key role for Treg-cell targeted therapies in SLE
and juvenile-onset SLE.12
Our study has several limitations. Difficulty in recruiting
and taking larger volumes of blood from rare cohorts of
young individuals, especially transgender individuals and
children, limited our ability to assess the phenotype and
function of Treg cells. Validation in larger cohorts of
young individuals would be beneficial, but this is beyond
the scope of this study. As highlighted by our study, it is
also possible that Tresp cells could be important for the
inflammatory balance between cisgender men and
women, which would require additional phenotype and
functional analyses. The increased suppressive capacity of
Treg cells in cisgender men could also be due to increased
proliferation of Tresp cells in cisgender women, or these
cells not being as amenable to suppression compared with
Tresp cells from cisgender men. We attempted to account
for this by plotting the suppressive capacity of Treg cells as
the fold change between Tresp-cell proliferation in the
presence and absence of Treg cells, thus normalising the
data for basal Tresp-cell proliferative capacity upon
stimulation; however, this remains a study limitation. It
was beyond the scope of the study to measure sex
hormone concentrations in the serum of study
participants; however, clinical monitoring aims to keep
concentrations within physiological ranges. There was no
follow-up of transgender individuals, which prevented us
from observing the long-term effects of hormone therapy
on Treg cells. Despite using a diverse healthy cohort for
the immunophenotyping and accounting for ethnicity in
analysis, the RNA-sequencing was done in White
transgender individuals; thus, our findings will require
validation in larger, more diverse cohorts. Although we
have speculated on the functional implications on Treg
cells of specific DEGs and pathways between sexes, the
physiological relevance of these differences would need to
www.thelancet.com/rheumatology Vol 4 October 2022

be confirmed by mechanistic studies and analysis of
proteins. The use of conventional flow cytometry limited
the number of markers that we could measure for
phenotyping. In future studies, it would be useful to
investigate additional surface markers for Treg cells and
Tresp cells to assess function and partition T cells into
sublineages. A wider array of markers were used to this
effect by Lambert and colleagues,32 who revealed important
features that vary with autoimmune disease states and by
age. Finally, the patients with juvenile-onset SLE were
clinically heterogeneous and were on different treatments,
presenting possible confounding effects on the immune
system, especially considering that dsDNA has previously
been shown to correlate negatively with Treg cells in
adults with SLE33 and that immuno­ suppressive drugs,
such as glucocorticoids, hydroxy­ chloroquine, and
methotrexate, could influence Treg-cell frequency and
function.34,35 Despite finding no difference in clinical
measures or treatments between cisgender men and
cisgender women with juvenile-onset SLE, or any
association of these juvenile-onset SLE features with Treg
frequency in our study, the potentially different treatment
strategies are important to consider because treatments
for the disease could explain the absence of differences in
Treg-cell frequency in patients with juvenile-onset SLE. It
is possible that this cohort—which includes patients
with predominantly well controlled and inactive juvenile-
onset SLE, with a high proportion of patients on hydroxy­
chloroquine and several other treatments—might mask
the potential effect of treatment and disease features on
Treg-cell frequency and function by sex. Because juvenile-
onset SLE is a severe disease frequently requiring timely
commencement of therapy, recruitment of treatment-
naive patients is not usually feasible.
To conclude, we have highlighted sexual dimorphisms
in Treg-cell profiles and function by sex in young healthy
individuals and patients with autoimmunity, showing
differential influences of sex chromosomes and
hormones using unique transgender cohorts. This
information helps us to understand the mechanistic
pathogenesis of autoimmune disease and the bias
towards cisgender women. Our study will help inform
the future consideration of sex as a biological variable in
inflammatory research and clinical trials.
Contributors
ECJ, CC, IP-T, and GAR designed the study. GAR, HP, and GB acquired
the data. CC, HP, and GB recruited the patients. All authors had access
to the data. GAR, ECJ, and JP accessed and verified the data. GAR and
JP analysied the data. GAR and ECJ wrote the manuscript. CC, IP-T,
and GB reviewed the manuscript. All authors approved the final
version.
Declaration of interests
We declare no competing interests.
Data sharing
Immune phenotype data can be found at Mendeley Data. RNA
sequencing data can be found at ArrayExpress (Accession number:
E-MTAB-11919) repositories. This data will be available from manuscript
publication date.
www.thelancet.com/rheumatology Vol 4 October 2022
Articles
Acknowledgments
This work was supported by a Versus Arthritis Career Development
Fellowship (22856), grants from the National Institute for Health and
Care Research University College London Hospital Biomedical Research
Centre (grant numbers BRC772/III/EJ/101350 and BRC773/III/
CC/101350), and Lupus UK and The Rosetrees Trust (M409). The study
was done at the Centre for Adolescent Rheumatology Versus Arthritis at
University College London. University College London Hospital and
Great Ormand Street Hospital supported by grants from Versus
Arthritis (21593 and 20164), Great Ormond Street Hospital Charity, and
the National Institute for Health and Care Research-Biomedical
Research Centres at both Great Ormand Street Hospital and University
College London Hospital. The views expressed are those of the authors
and not necessarily those of the UK NHS, the National Institute for
Health and Care Research or the UK Department of Health.
References
1 Klein SL, Flanagan KL. Sex differences in immune responses.
Nat Rev Immunol 2016; 16: 626–38.
2 Peckham H, de Gruijter NM, Raine C, et al. Male sex identified by
global COVID-19 meta-analysis as a risk factor for death and ITU
admission. Nat Commun 2020; 11: 6317.
3 Billi AC, Kahlenberg JM, Gudjonsson JE. Sex bias in autoimmunity.
Curr Opin Rheumatol 2019; 31: 53–61.
4 Uppal SS, Verma S, Dhot PS. Normal values of CD4 and CD8
lymphocyte subsets in healthy indian adults and the effects of sex,
age, ethnicity, and smoking. Cytometry B Clin Cytom 2003; 52: 32–36.
5 Abdullah M, Chai PS, Chong MY, et al. Gender effect on in vitro
lymphocyte subset levels of healthy individuals. Cell Immunol 2012;
272: 214–19.
6 Hewagama A, Patel D, Yarlagadda S, Strickland FM,
Richardson BC. Stronger inflammatory/cytotoxic T-cell response in
women identified by microarray analysis. Genes Immun 2009;
10: 509–16.
7 Hughes GC, Choubey D. Modulation of autoimmune rheumatic
diseases by oestrogen and progesterone. Nat Rev Rheumatol 2014;
10: 740–51.
8 Ambrose N, Morgan TA, Galloway J, Ionnoau Y, Beresford MW,
Isenberg DA. Differences in disease phenotype and severity in SLE
across age groups. Lupus 2016; 25: 1542–50.
9 Holdcroft A. Gender bias in research: how does it affect evidence
based medicine? J R Soc Med 2007; 100: 2–3.
10 Beery AK, Zucker I. Sex bias in neuroscience and biomedical
research. Neurosci Biobehav Rev 2011; 35: 565–72.
11 Robinson GA, Peng J, Dönnes P, et al. Disease-associated and
patient-specific immune cell signatures in juvenile-onset systemic
lupus erythematosus: patient stratification using a machine-
learning approach. Lancet Rheumatol 2020; 2: e485–96.
12 Mizui M, Tsokos GC. Targeting regulatory T cells to treat patients
with systemic lupus erythematosus. Front Immunol 2018; 9: 786.
13 Van Nieuwenhove E, Lagou V, Van Eyck L, et al. Machine learning
identifies an immunological pattern associated with multiple
juvenile idiopathic arthritis subtypes. Ann Rheum Dis 2019;
78: 617–28.
14 Afshan G, Afzal N, Qureshi S. CD4+CD25(hi) regulatory T cells in
healthy males and females mediate gender difference in the
prevalence of autoimmune diseases. Clin Lab 2012; 58: 567–71.
15 Tai P, Wang J, Jin H, et al. Induction of regulatory T cells by
physiological level estrogen. J Cell Physiol 2008; 214: 456–64.
16 Polanczyk MJ, Carson BD, Subramanian S, et al. Cutting edge:
estrogen drives expansion of the CD4+CD25+ regulatory T cell
compartment. J Immunol 2004; 173: 2227–30.
17 Goodman WA, Bedoyan SM, Havran HL, Richardson B,
Cameron MJ, Pizarro TT. Impaired estrogen signaling underlies
regulatory T cell loss-of-function in the chronically inflamed
intestine. Proc Natl Acad Sci USA 2020; 117: 17166–76.
18 Polanczyk MJ, Hopke C, Huan J, Vandenbark AA, Offner H.
Enhanced FoxP3 expression and Treg cell function in pregnant and For more on the immune
estrogen-treated mice. J Neuroimmunol 2005; 170: 85–92. phenotype see http://dx.doi.
19 Arruvito L, Sanz M, Banham AH, Fainboim L. Expansion of org/10.17632/mcsypbk6nz.2
CD4+CD25+and FOXP3+ regulatory T cells during the follicular
phase of the menstrual cycle: implications for human reproduction. For more on the RNA
J Immunol 2007; 178: 2572–78. sequencing data see www.ebi.
ac.uk/arrayexpress
e723
 Articles
20 Page ST, Plymate SR, Bremner WJ, et al. Effect of medical
castration on CD4+ CD25+ T cells, CD8+ T cell IFN-gamma
expression, and NK cells: a physiological role for testosterone and/
or its metabolites. Am J Physiol Endocrinol Metab 2006;
290: E856–63.
21 Lee YA, Kim HR, Lee JS, et al. CD4+ FOXP3+ regulatory T cells
exhibit impaired ability to suppress effector T cell proliferation in
patients with turner syndrome. PLoS One 2015; 10: e0144549.
22 Collier FM, Tang ML, Martino D, et al. The ontogeny of naïve and
regulatory CD4(+) T-cell subsets during the first postnatal year:
a cohort study. Clin Transl Immunology 2015; 4: e34.
23 Singh RP, Bischoff DS. Sex hormones and gender influence the
expression of markers of regulatory T cells in SLE patients.
Front Immunol 2021; 12: 619268.
24 Pompura SL, Dominguez-Villar M. The PI3K/AKT signaling
pathway in regulatory T-cell development, stability, and function.
J Leukoc Biol 2018; 103: 1065–76.
25 Soond DR, Slack ECM, Garden OA, Patton DT, Okkenhaug K.
Does the PI3K pathway promote or antagonize regulatory T cell
development and function? Front Immunol 2012; 3: 244.
26 Hiwa R, Nielsen HV, Mueller JL, Mandla R, Zikherman J. NR4A
family members regulate T cell tolerance to preserve immune
homeostasis and suppress autoimmunity. JCI Insight 2021;
6: 151005.
27 Pfoertner S, Jeron A, Probst-Kepper M, et al. Signatures of human
regulatory T cells: an encounter with old friends and new players.
Genome Biol 2006; 7: R54.
e724
28 Ramírez Sepúlveda JI, Bolin K, Mofors J, et al. Sex differences in
clinical presentation of systemic lupus erythematosus.
Biol Sex Differ 2019; 10: 60.
29 Yang J, Chu Y, Yang X, et al. Th17 and natural Treg cell population
dynamics in systemic lupus erythematosus. Arthritis Rheum 2009;
60: 1472–83.
30 Singh RP, Hahn BH, Bischoff DS. Interferon Genes Are Influenced
by 17β-Estradiol in SLE. Front Immunol 2021; 12: 725325.
31 Christou EAA, Banos A, Kosmara D, Bertsias GK, Boumpas DT.
Sexual dimorphism in SLE: above and beyond sex hormones. Lupus
2019; 28: 3–10.
32 Lambert K, Moo KG, Arnett A, et al. Deep immune phenotyping
reveals similarities between aging, Down syndrome, and
autoimmunity. Sci Transl Med 2022; 14: eabi4888.
33 Lee HY, Hong YK, Yun HJ, Kim YM, Kim JR, Yoo WH. Altered
frequency and migration capacity of CD4+CD25+ regulatory T cells
in systemic lupus erythematosus. Rheumatology (Oxford) 2008;
47: 789–94.
34 Mathian A, Jouenne R, Chader D, et al. Regulatory T cell responses
to high-dose methylprednisolone in active systemic lupus
erythematosus. PLoS One 2015; 10: e0143689.
35 Scottà C, Fanelli G, Hoong SJ, et al. Impact of immunosuppressive
drugs on the therapeutic efficacy of ex vivo expanded human
regulatory T cells. Haematologica 2016; 101: 91–100.
www.thelancet.com/rheumatology Vol 4 October 2022