A meta signature of murine Treg and Penk mRNA expression in lymphoid and non lymphoid organs.

A) Top 25 genes enriched in Treg compared to Tconv from lymphoid tissue from at least 10 of the 11 datasets analyzed ranked by fold change (LogFC) of mean expression relative to Tconv. The Penk gene is highlighted in red. B) Correlation of Penk and Foxp3 mRNA expression in all the cell types listed in the legend. The Pearson correlation coefficient is indicated. C) Expression of Penk mRNA in Treg (blue) and Tconv (red) isolated from thymus, visceral adipose tissue (VAT) and muscle. The source of the data is indicated below each graph.

Characteristics of the datasets used for the Treg meta-signature

(NA = not available; LN = Lymph Nodes)

Penk mRNA expression is regulated by TNFR signaling and the BATF transcription factor in Treg.

A) A network of the genes most correlated to Penk in Treg is shown. The Pearson correlation values were extracted from the Immuno-Navigator database (taking only Treg for the analysis) and integrated into Cytoscape v3.730. Each node is a gene linked by edges with width proportional to the Pearson correlation. (edge range: 0.538-0.758). B) Illustration of the correlation between expression of Penk and Batf in Treg. Each dot is a sample from the Immuno-Navigator database. The Pearson correlation coefficient is indicated on the figure. C) Penk and Batf mRNA expression after in vitro stimulation of purified Treg with the indicated TNFR agonists prior (0), and at 18 and 36hrs after stimulation. Each dot is a biological replicate from a single experiment. D) GEO2R analysis of the GSE89656 dataset for Penk mRNA variations between wild type control Treg (WT) and BATF-KO Treg.

Penk is prominently expressed by Treg at steady state

A-B) (right) UMAP representing all major cell types indicated in the figures determined by flow cytometry from lymph nodes (A) and colon (B) (left). Projection of Penk (blue) on the UMAP of lymph node and colon (right). Subsets were manually gated as depicted in Figure S1A. C) Scatter plot displaying the average population size and percentage of Penk+ cells for the listed cell populations and organs. Population size was calculated as the percentage out of total CD45+ single cells and represented on a log10 scale. (n = 3 mice for the VAT, spleen and bone marrow, n = 6 mice for the other groups; results cumulative of two independent experiments).

Heat hyperalgesia in mice deficient for Penk in Treg.

A) Withdrawal latency of WT and Lox mice before the administration of TMX (baseline). The mean of all 4 measures before administration of TMX for each mouse is shown. Statistical modeling was performed using a non parametric unpaired Mann-Whitney t-test and was deemed as non-significant (ns). B) Withdrawal latency of WT and Lox mice before administration of TMX (baseline) shown separately for female and male mice, as in A. Statistical modeling was performed using two-way ANOVA and the effect of the genotype on the results was deemed non-significant (ns). C) Withdrawal latency of WT and Lox mice post-TMX treatment. The mean of all 4 measures taken from D7 onward is represented for each mouse. Statistical modeling was performed using a non parametric unpaired Mann-Whitney t-test and was deemed significant (p=0,0038). D) Withdrawal latency of WT and Lox mice post-TMX treatment shown separately for female and male mice, as in C. Statistical modeling was performed using two-way ANOVA and the effect of the genotype on the results was deemed significant (p=0,0037). A multiple comparisons test using the Benjamini-Hochberg False Discovery Rate model was used. The adjusted p values were deemed significant for both female and male mice (q=0,043). Results shown in this figure are cumulative of two independent experiments with n = 44 mice (26 WT and 18 Lox).

Representative flow cytometry profiles in PenkCre-ERT2x ROSAtdtomato

A) Gating strategy used to define the population for the mapping of Penk expression by immune cells. B) Representative staining of TdTomato and mCherry

Penk expression in Tconv and Treg according to activation status in lymph node

Representative staining of TdTomato and mCherry in naive (CD62Lhigh CD44low), central memory (CD62Lhigh CD44high) and effector memory (CD62Llow CD44high) Tconv (top) and Treg (bottom).

A kinetics of latency periods in response to heat in WT and Lox mice.

Each dot is the mean value (± SD) of latency period in response to heat stimulation for all mice of the WT or Lox genotype. The results are represented as a function of days relative to TMX gavage (D0). Statistical modeling of the results was performed with linear regression curve fitting with a comparison of intercept and slopes. The low p value allow to reject the hypothesis that one curve fits all the datasets.