Article
Senescence-induced immune remodeling facilitates metastatic adrenal cancer in a sex-dimorphic manner
Received: 1 May 2022
Accepted: 12 April 2023
Published online: 25 May 2023
☒ Q Check for updates
Kate M. Warde 1, Lorenzo J. Smith 01, Lihua Liu1, Chris J. Stubben2, Brian K. Lohman2, Parker W. Willett1, Julia L. Ammer3,
Guadalupe Castaneda-Hernandez1, Sikiru O. Imodoye 01, Chenge Zhang1, Kara D. Jones1, Kimber Converso-Baran4, H. Atakan Ekiz5, Marc Barry6, Michael R. Clay7, Katja Kiseljak-Vassiliades8, Thomas J. Giordano3,9,10, Gary D. Hammer 3,10 & Kaitlin J. Basham ☒
Aging markedly increases cancer risk, yet our mechanistic understanding of how aging influences cancer initiation is limited. Here we demonstrate that the loss of ZNRF3, an inhibitor of Wnt signaling that is frequently mutated in adrenocortical carcinoma, leads to the induction of cellular senescence that remodels the tissue microenvironment and ultimately permits metastatic adrenal cancer in old animals. The effects are sexually dimorphic, with males exhibiting earlier senescence activation and a greater innate immune response, driven in part by androgens, resulting in high myeloid cell accumulation and lower incidence of malignancy. Conversely, females present a dampened immune response and increased susceptibility to metastatic cancer. Senescence-recruited myeloid cells become depleted as tumors progress, which is recapitulated in patients in whom a low myeloid signature is associated with worse outcomes. Our study uncovers a role for myeloid cells in restraining adrenal cancer with substantial prognostic value and provides a model for interrogating pleiotropic effects of cellular senescence in cancer.
Aging is the most important risk factor for cancer1. Incidence rates rise steeply with age, with -90% of newly diagnosed cancer occurring after the age of 50 years2,3. Age-driven changes in the tissue landscape cooperate with oncogenic mutations to determine overall cancer risk4. During aging, the tissue microenvironment evolves based on a multitude of physiological factors, including oxygen and nutrient
availability, extracellular matrix (ECM) composition and immune cell infiltration5. These tissue-level changes influence the functional impact of oncogenic mutations3. However, despite aging being a major carcinogen, we have limited understanding of how the aged microenvironment cooperates with genetic mutations to influence tumorigenesis.
1Department of Oncological Sciences, Huntsman Cancer Institute, University of Utah, Salt Lake City, UT, USA. 2Bioinformatics Shared Resource, Huntsman Cancer Institute, University of Utah, Salt Lake City, UT, USA. 3Department of Internal Medicine, Division of Metabolism, Endocrinology and Diabetes, University of Michigan, Ann Arbor, MI, USA. 4Frankel Cardiovascular Center Physiology and Phenotyping Core, University of Michigan, Ann Arbor, MI, USA. 5Department of Molecular Biology and Genetics, Izmir Institute of Technology, Urla Izmir, Turkey. 6Department of Pathology, University of Utah, Salt Lake City, UT, USA. 7Department of Pathology, University of Colorado School of Medicine at Colorado Anschutz Medical Campus, Aurora, CO, USA. 8Division of Endocrinology, Metabolism and Diabetes, Department of Medicine, University of Colorado School of Medicine at Colorado Anschutz Medical Campus, Aurora, CO, USA. 9Department of Pathology, University of Michigan, Ann Arbor, MI, USA. 1ºEndocrine Oncology Program, Rogel Cancer Center, University of Michigan, Ann Arbor, MI, USA. ☒ e-mail: kaitlin.basham@hci.utah.edu
a
4 weeks
6 weeks
12 weeks
♀
♀
♀
Control
1.84 mm3
3.32 mm3
3.55 mm
3
Znrf3 cKO
♀
♀
♀
22.91 mm3
39.78 mm3
23.03 mm3
3
c
Female mice
**** P < 1.00 × 10
-15
**** P = 3.43 x 10
-10
*P = 1.30 x 10
5
Adrenal weight (mg) per body weight (g)
6
** P = 9.89 x 10
3
Control
Znrf3 cKO
4
A
2
0
.
=
4
6
9
12
24
Fold change:
Age (weeks)
8.05
10.17
7.71
6.57
4.20
e
Female mice: Ki67 6 weeks
4 weeks
9 weeks
Control
Znrf3 cKO
P = 0.89
40
*P=9.60 x 10
-6
Ki67 proliferation index (%)
*** P= 4.05 ×10-4
**** P = 2.58 x10
5
P = 0.85
30
20
10
0
4 weeks
6 weeks
9 weeks
Control
Znrf3 cKO
b
Adrenal volume (mm3) per body weight (g)
2.5
** P = 9.32 × 10~
3
** P = 5.57 x 10
3
♀
TI
2.0
1.5
1.0
0.5
P = 0.31
P = 0.35
0
4
5
6
7
8
9
10
11
12
Age (weeks)
Control (n = 3)
Znrf3 cKO (n = 5)
d
Male mice **** P =1.90 ×10-11
**** P =1.48 x 10°
**** P =5.62 ×10
-4
Adrenal weight (mg) per body weight (g)
6
P = 0.46
Control
Znrf3 cKO
4
2
B
8
0
4
6
9
12
24
Fold change:
Age (weeks)
7.64
8.77
6.37
5.94
4.82
f
Male mice: Ki67 6 weeks
4 weeks
9 weeks
Control
Znrf3 cKO
**** P=5.24 x10 8
40
P = 0.84
Ki67 proliferation index (%)
**** P=2.90 ×10-5
P = 0.83
P = 0.99
30
20
10
0
4 weeks
6 weeks
9 weeks
Control
Znrf3 cKO
Adrenocortical carcinoma (ACC) is a highly pertinent model to study the interplay between the aging microenvironment and cell- intrinsic driver mutations. ACC is an aggressive cancer of the adre- nal cortex with no effective treatments6. Moreover, peak incidence occurs at the age of 45-55 years7, which corresponds to-12-18 months in mice8. One pathway frequently hyperactivated in ACC is Wnt-B- catenin, a highly conserved signaling cascade that mediates proper cell renewal and cell-to-cell interactions in many tissues9. Unlike most Wnt-driven tumors that are induced by activating mutations in CTNNB1 (B-catenin) or inactivation of the ß-catenin destruction complex, ACC tumors frequently harbor loss-of-function (LOF) alterations in ZNRF3 (refs. 10,11), encoding an E3 ubiquitin ligase that promotes endocytic turnover of Wnt receptors12,13. Consequently, ZNRF3 loss increases receptor availability and renders cells hypersensitive to Wnt ligands within the microenvironment.
Wnt-driven ACC has been previously modeled in mice through CTNNB1 gain of function14 or APC LOF15. These approaches result in mild hyperplasia with late progression to benign adenomas or, in some rare cases, carcinomas. Combined CTNNB1 gain of function and p53 LOF shortens tumor latency but still requires 12 weeks or longer for tumors16. These studies hint at an important role for aging in the initiation and progression of adrenal tumors. However, these models are based on less frequent Wnt alterations in ACC, and the mechanism by which aging influences tumorigenesis remains unresolved.
We recently developed a model to specifically ablate Znrf3 in the adrenal cortex using steroidogenic factor 1 (SF-1)-Cre, which mediates recombination throughout the adrenal cortex beginning at embryonic day (E)9.5 (ref.17). We initially characterized Znrf3-conditional knock- out (cKO) mice between postnatal day (P)0 and adrenal maturation at 6 weeks18. We observed progressive hyperplasia that resulted in a more than eightfold increase in adrenal weight. Further characteriza- tion revealed a gradient of Wnt-B-catenin activity required for normal homeostasis with a specific role for ZNRF3 in limiting expansion of Wnt-low cells. Nevertheless, we found no evidence of neoplastic trans- formation during these early studies.
Here, we investigated the interplay between genetic Znrf3 inacti- vation and the aging microenvironment. We hypothesized that Znrf3- cKO mice would progress from adrenal hyperplasia to carcinoma with age. Unexpectedly, we found that Znrf3-cKO adrenals steadily regress over time. We demonstrate that this phenotypic switch from hyper- plasia to regression is driven by activation of cellular senescence and a subsequent senescence-associated secretory phenotype (SASP). We characterize diverse immune infiltrates that remodel the tissue microen- vironment following senescence activation, which exhibit a highly sex- dimorphic response that is enhanced in males by androgens. However, following prolonged inflammation, a large proportion of mice ultimately develop adrenal tumors. Males predominately form benign lesions, while females are significantly more prone to metastatic adrenal tumors, mirroring the higher incidence of ACC in women7. Myeloid-derived
immune cells are excluded from the tissue microenvironment as tumors progress, which is recapitulated in human patients with ACC in whom a low myeloid response is associated with worse outcome. In sum, our results suggest that Znrf3 loss is permissive for ACC tumor progression with advanced age (Graphical Abstract, Supplementary Fig. 1). More broadly, our work suggests a mechanism by which extrinsic factors, including sex and the immune microenvironment, cooperate with genetic lesions to promote age-induced malignant transformation.
Results
Ultrasound imaging measures adrenal size in real time
Given the significant adrenal hyperplasia that we previously observed at 6 weeks, we hypothesized that Znrf3-cKO animals would progress to form adrenal tumors with increased age. We employed ultrasound imaging as a non-invasive technique to quantify adrenal size in real time. We first validated this approach using a mixed cohort of mice with varying adrenal sizes. We measured adrenal area and volume by ultrasound and then euthanized mice 24 h after imaging to obtain adrenal weight (Extended Data Fig. 1a). We found that adrenal weight was significantly correlated with both ultrasound area (R2 = 0.90, ** P=1.08 ×10-3) (Extended Data Fig. 1b) and volume (R2 = 0.89, ** P=1.35 × 10-3) (Extended Data Fig. 1c). These results demonstrate that ultrasound imaging provides an accurate assessment of adrenal size in real time.
Following hyperplasia, ZNRF3-deficient adrenals regress
To monitor for potential tumors, we performed weekly adrenal ultra- sound imaging in Znrf3-cKO animals. We began imaging females at 4 weeks and observed an expected increase in adrenal size by 6 weeks (Fig. 1a,b). However, rather than continued growth, we unexpectedly observed a gradual decline in adrenal volume after 6 weeks (Fig. 1a,b).
To follow up on these observations, we collected cohorts of male and female mice at 4, 6, 9, 12 and 24 weeks of age. Consistent with ultrasound measurements, we initially observed increased adrenal weight in female Znrf3-cKO animals between 4 and 6 weeks of age, followed by a progressive decline in adrenal weight (Fig. 1c). In males, adrenal weight did not continue to increase after 4 weeks and, similar to females, gradually declined with increasing age (Fig. 1d). In sum, these results indicate that ZNRF3-deficient adrenals undergo a distinct phenotypic switch from hyperplasia to regression.
Adrenal regression begins with reduced proliferation
We hypothesized that declining adrenal size in Znrf3-cKO animals was caused by enhanced cell death. To test this, we measured apoptosis, which is the dominant form of cell death attributed to homeostatic turnover of the normal adrenal cortex19. Using immunohistochemistry (IHC) for cleaved caspase 3 (CC3), we found that CC3-positive cells primarily localized at the cortical-medullary boundary in controls, as expected (Extended Data Fig. 2a,b). In Znrf3-cKO animals, apoptotic
Fig. 2 | The switch from hyperplasia to regression is marked by activation of cellular senescence. a, Bulk RNA-seq reveals a significant change in gene expression signature during the switch from hyperplasia to regression. Heatmap of significant DEGs in adrenals from female Znrf3-cKO animals at 6, 9 and 12 weeks of age (Padj < 0.05). b, IPA identifies the most significantly altered canonical pathways in 9- versus 6-week-old female Znrf3-cKO adrenals; statistical analysis was performed using a right-tailed Fisher’s exact test (P < 0.05). AHR, aryl hydrocarbon receptor; CHK, Csk homologous kinase; GP6, Glycoprotein VI. c, DNA damage as measured by 53BP1 foci is significantly increased in female Znrf3-cKO adrenals compared to those of controls at 4 and 6 weeks of age. Quantification of 53BP1 foci was performed using QuPath digital image analysis based on the number of positive foci and normalized to total nuclei. Asterisks indicate 53BP1-positive foci. Scale bars, 10 um. d, IPA of the most significantly activated or inhibited transcriptional regulators in adrenals of 9- versus 6-week-old female Znrf3-cKO animals. Red asterisks highlight transcriptional
regulators associated with cellular senescence. p21 (e), p16INK4a (f) and SA-B- gal (g) are significantly increased in adrenals of 9-week-old female Znrf3-cKO animals compared to those of controls. Scale bars, 100 um. Quantification of p21 and p16INK4ª IHC was performed using QuPath digital image analysis based on the number of positive cells per high-powered field (HPF). Representative SA-ß-gal images obtained from analysis of three independent mice are shown. h, Co-staining of p21 and GFP in 9-week-old female Znrf3-cKO mice containing the R26RmTmG lineage reporter tool confirms that p21-positive senescent cells are GFP-positive adrenal cortex cells. Representative images obtained from analysis of five female and seven male animals are shown. Scale bar, 10 um. WT, wild type. b,d, IPA analysis includes four biological replicates per group. Statistical tests were performed using IPA (P < 0.05). c,e,f, Error bars represent mean ± s.e.m. Each dot represents an individual animal. Statistical analysis was performed using two-way ANOVA followed by Tukey’s multiple-comparison test.
a
Hyperplasia
Regression
b
Canonical pathway analysis: 9 9- vs 6-week Znrf3 cKO
9 Znrf3 cKO 6 weeks
1
9 weeks
12 weeks
Inhibited
Activated
3
Coronavirus pathogenesis pathway
·
2
Xenobiotic metabolism AHR signaling pathway
1
Role of CHK proteins in cell cycle checkpoint control
0
Senescence pathway
-1
Cell cycle: G2/M DNA damage checkpoint regulation
-2
GP6 signaling pathway
P value
-3
Glutathione-mediated detoxification
1 × 10-20
Estrogen-mediated S phase entry
1 × 10-15
1 × 10-10
Aryl hydrocarbon receptor signaling
· 1 × 10-5
Cyclins and cell cycle regulation
Base excision repair pathway
Kinetochore metaphase signaling pathway
Nucleotide excision repair, enhanced pathway
Cell cycle control of chromosomal replication
-4.5-4.0-3.5-3.0-2.5
1.5
1.8
2.1
2.4
Activation Z score
c
d Transcriptional regulator analysis: 99- vs 6-week Znrf3 cKO
♀
4 weeks
6 weeks
Inhibited
Activated
Percent 53BP1 foci (%)
6
P = 1.04 x 10
5
*** P=1.22 ×10
-4
TBX2
CEBPB
*
Control
CCND1
*
4
MITF
*
53BP1
53BP1
MYC
2
E2F1
P value
*
*
*
Znrf3 cKO
FOXM1
1 × 10-33
*
*
*
1 × 10-25
*
*
RBL1
*
*
1 x 10-17
*
0
*
*
*
*
4 weeks
6 weeks
TCF3
1 × 10-9
*
*
*
53BP1
53BP1
Control
Znrf3 cKO
*RB1
*
*
SMARCB1
* CDKN2A
* TP53
NUPR1
-6.0 -5.7 -5.4 -5.1
4
5
6
Activation Z score
e
6 weeks
f
♀
9 weeks
P = 0.78
♀ 6 weeks
9 weeks
*** P = 8.04 x10 **
4
200
P= 0.052
*** P = 7.23 × 10
4
60
P = 0.19
**** P = 2.15 x10 **
5
Control
p21+ cells per HPF
Control
p16+ cells per HPF
150
40
p21
p21
100
p16
p16
Znrf3 cKO
20
50
Znrf3 cKO
p21
p21
0
0
6 weeks
9 weeks
p16
p16
6 weeks
9 weeks
Control
Znrf3 cKO
Control
Znrf3 cKO
g
9 control (9 weeks)
9 Znrf3 cKO (9 weeks)
h
9 SF1Cre/+;Znrf3fl/fl;R26RmTmG/WT
Merge
p21
GFP
SA-ß-gal -
SA-ß-gal
p21 (senescence) GFP (adrenal cortex) Hoechst (nuclei)
cells were scattered throughout the cortex, but there was no increase in the proportion of CC3-positive cells compared to controls (Extended Data Fig. 2a,b). We concluded that the age-dependent decrease in adre- nal size in Znrf3-cKO animals was not caused by increased apoptotic cell death.
Next, we assessed proliferation in Znrf3-cKO adrenals. Previously, we observed hyperplasia at 6 weeks in females18. Consistent with these prior observations, the Ki67 index in female Znrf3-cKO animals was significantly higher than that in controls at 4 and 6 weeks (Fig. 1e). How- ever, proliferation was reduced to baseline control levels by 9 weeks. This striking drop in proliferation was coincident with the onset of adrenal regression that we observed by adrenal ultrasound and weight. Male Znrf3-cKO animals also displayed an initial phase of hyperplasia evident by the high rate of proliferation at 4 weeks. However, the Ki67 index returned to baseline by 6 weeks (Fig. 1f). These results indicate that adrenal cortex hyperplasia in Znrf3-cKO animals is not sustained. Furthermore, the onset of regression occurs earlier in males, suggest- ing an underlying sexual dimorphism.
Cell cycle arrest triggers adrenal regression
To identify potential mechanisms that mediate the switch from hyper- plasia to regression, we performed bulk RNA-seq on control and Znrf3-cKO adrenals. We focused our analysis on females at 6 weeks (hyperplasia), 9 weeks (regression onset) and 12 weeks (regression). Control adrenals had a highly stable expression signature, with only 23 differentially expressed genes (DEGs) between 6 and 9 weeks (17 upregulated, 6 downregulated, adjusted P value (Padj) < 0.05) (Supplementary Table 1). By contrast, we observed a striking change in gene expression in Znrf3-cKO animals, with 720 significant DEGs (383 upregulated, 337 downregulated, Padj < 0.05) (Fig. 2a).
Using Ingenuity Pathway Analysis (IPA)20, we analyzed DEGs to identify canonical pathways associated with the switch from hyper- plasia to regression. Among the top inhibited pathways, we identified multiple pathways that regulate cell division (Fig. 2b). Conversely, among the top activated pathways, we detected several signatures associated with enhanced DNA damage and cell cycle arrest (Fig. 2b). We validated these results in situ by measuring 53BP1, which is a DNA damage-response factor recruited to DNA double-strand breaks21. Consistent with our RNA-seq data, we observed a significant increase in 53BP1 foci in Znrf3-cKO animals compared to controls (Fig. 2c and Extended Data Fig. 2c). In sum, these data suggest that ZNRF3-deficient adrenal cortex cells accumulate DNA damage and initiate cell cycle arrest near the onset of regression.
Adrenal regression is associated with senescence activation
We next performed upstream regulator analysis (URA) to identify candidate factors that could coordinate the observed change in gene expression. URA is a new IPA tool that analyzes the relationship between DEGs to identify potential upstream factors20. We focused our analysis on transcriptional regulators and identified 47 candidates (23 activated, 24 inhibited; P < 0.05). Among the most significantly activated regu- lators, we found TP53 (Z score = 6.20), CDKN2A (Z score = 5.68) and
RB1 (Z score =4.86) (Fig. 2d), which are central regulators of cellular senescence, one of the top activated canonical pathways (Fig. 2b).
Cellular senescence is an aging-associated stress response that causes proliferative arrest of damaged cells22. Enhanced cellular stress, which can be induced by a wide range of stimuli, activates the p53 and p16INK4ª tumor-suppressor proteins, leading to cell cycle inhi- bition through p21 and Rb. Based on our RNA-seq analysis, in com- bination with observed phenotypic changes, we hypothesized that cellular senescence was activated in Znrf3-cKO animals during the transition from hyperplasia to regression. To test this hypothesis, we measured multiple hallmarks of senescence at 6 and 9 weeks. We observed increased expression of p21 (Fig. 2e and Extended Data Fig. 2d) and p16INK4a (Fig. 2f and Extended Data Fig. 2e) as well as accumula- tion of senescence-associated ß-galactosidase (SA-ß-gal) (Fig. 2g and Extended Data Fig. 2f) in adrenals from 9-week-old Znrf3-cKO animals compared to those from controls. To confirm senescence activation in adrenal cortex cells, as opposed to the stromal or immune compart- ment, we performed co-staining for p21 and green fluorescent protein (GFP) in adrenals isolated from mice containing the R26RmTmG lineage reporter23. In these animals, the adrenal cortex is permanently labeled with GFP through SF-1-Cre-mediated recombination. We found that all p21-positive cells were positive for GFP at 9 weeks (Fig. 2h). Collectively, these data demonstrate that cellular senescence is activated in ZNRF3- deficient adrenal cortex cells during the switch from hyperplasia to regression.
Senescent adrenal cortex cells develop a secretory phenotype One defining feature of senescence that distinguishes it from other forms of growth arrest is the SASP24. The SASP comprises multiple families of secreted proteins including soluble signaling factors, secreted proteases and secreted insoluble proteins or ECM compo- nents. Although this phenotype is both temporally dynamic and het- erogeneous across tissue types or modes of senescence induction 25,26, the SASP collectively acts as a potent pro-inflammatory stimulus to facilitate tissue remodeling.
To begin to define the SASP in the context of Znrf3 loss in the adre- nal gland, we performed URA for activated cytokines and growth fac- tors in our RNA-seq dataset. During senescence activation at 9 weeks, we identified interleukin (IL)-1B (Z score =2.96), interferon (IFN)-y (Z score = 2.48) and tumor necrosis factor (TNF) (ligand) superfamily, member 13 (TNFSF13) (Z score = 2.00) as the most highly induced cytokines (Fig. 3a and Extended Data Fig. 3a). By 12 weeks, additional members of the IL-1 (IL-1a, Z score =2.38), IL-6 (IL-6, Z score = 2.15 and oncostatin M (OSM), Z score = 3.15) and TNF family (TNF, Z score =3.12 and TNFSF12, Z score =2.37) were significantly activated (Fig. 3b and Extended Data Fig. 3b). Gene set enrichment analysis (GSEA) at both 9 weeks (Fig. 3c) and 12 weeks (Fig. 3d) further confirmed significant posi- tive enrichment for genes associated with an inflammatory response. Growth factor analysis additionally revealed significant induction of SASP-associated factors, including transforming growth factor (TGF)-ß3 (ref. 27) and brain-derived neurotrophic factor (BDNF)28 (Extended Data Fig. 3c,d).
Fig. 3 | Senescent Znrf3-cKO adrenal glands develop a functional SASP. URA using bulk RNA-seq data predicts significantly activated and inhibited cytokines in adrenal tissue from Znrf3-cKO animals at 9 weeks (a) and 12 weeks (b) compared to 6 weeks. Data are representative of four biological replicates per group. Statistical analysis in IPA was performed using a right-tailed Fisher’s exact test (P < 0.05). c,d, GSEA for inflammatory signatures (IL-6-Janus kinase (JAK)-signal transducer and activator of transcription 3 (STAT3) signaling, chemokine signaling, the inflammatory response and the innate immune system) identifies positively enriched pathways in adrenal tissue from Znrf3- cKO animals at 9 weeks (c) and 12 weeks (d) compared to 6 weeks. FDR, false discovery rate; NES, normalized enrichment score. e, Heatmap representing
supervised hierarchical clustering of the 40 most significant DEGs that encode secreted proteins compared to controls; statistical analysis was performed using the Wald test (Padj < 0.05). f,g, Histological evaluation of adrenal tissue from female control and Znrf3-cKO animals based on H&E staining. Female Znrf3-cKO animals accumulate histiocytes (inset) between 12 and 24 weeks (wk) of age. h,i, Histiocytes (insets) accumulate earlier and occupy a larger proportion of the adrenal gland in male Znrf3-cKO animals than in female animals. Quantification was performed using QuPath digital analysis based on the proportion of histiocyte area normalized to the total adrenal cortex area. Scale bars, 100 um.
Given significant activation of multiple SASP-associated cytokines and growth factors, we next sought to more broadly characterize this phenotype. We analyzed expression of genes encoding secreted
proteins29 that were significantly differentially expressed in Znrf3-cKO adrenals compared to controls at 6, 9 and 12 weeks. We then prior- itized top candidate SASP factors based on upregulated genes that
Cytokine analysis: º 12- vs 6-week Znrf3 cKO
Inhibited
Activated
CSF2
IL-10
IL-1B
IL-6
TNFSF12
P value
1 × 10-18
IL-1a
· 1 × 10-13
IL-4
· 1 × 10-8
IFN-Y
· 1 x 10-3
TNF
OSM
,A
3
2
1.8
2.1
2.4
2.7
3.0
Activation Z score
e
Znrf3 Q
2
Control
Mmp12
Lgals3
CKO
Trem2
1
Il7r
Age
Il1rn
Smpdl3a
6 weeks
Tfpi
0
9 weeks
C4b
Btd
-1
12 weeks
Ggh
Kars
Nsun2
Pdgfd
Clu
-2
Mmp19
Rnpep
C2
Dpp7
C1rl
Fstl3
Nxph1
Scube1
Tnfsf13b
Tnfrsf1b
Pros1
Tgfa
Ghr
Col4a6
Col4a5
Clqtnf6
Col5a1
Scube3
Txndc16
Crtap
Clqbp
Mif
Hmgb1
Cd63
Mfge8
Lgals1
g
Female mice
4 wk
6 wk
9 wk
12 wk 24 wk
100
Histiocyte
area
None
<2%
60
2-10%
40
10-20%
20
20-50%
>50%
0
Control
Znrf3 cKO
Control
Znrf3 cKO
Control
Znrf3 cKO
Control
Znrf3 cKO
Control
Znrf3 cKO
f
4 weeks
6 weeks
9 weeks
12 weeks
24 weeks
Control
Female
Znrf3 cKO
h
4 weeks
6 weeks
9 weeks
12 weeks
24 weeks
Control
Male
Znrf3 cKO
b
Cytokine analysis: 9 9- vs 6-week Znrf3 cKO
Inhibited
Activated
CSF2
IL-10
IL-6
P value
IFN-ß1
. 1 x 10-13 -13
1 x 10 -9
TNFSF13
· 1 x 10-5
IFN-Y
IL-13
1
6
5
x
1.5
2.0
2.5
3.0
Activation Z score
c
GSEA: 9- vs 6-week Znrf3 cKO
Enrichment score
0.5
IL-6-JAK-STAT3 signaling
Chemokine signaling pathway
NES = 1.5
Enrichment score
NES = 1.8
0.4
FDR = 0.1
0.4 -
0.3
P = 0.016
FDR = 0.1
P= 0.001
0.2
0.2 -
0.1
0
0
0
5,000
10,000
15,000
Rank
0
5,000 10,000 15,000
1
1
1
Rank
Inflammatory response
Innate immune system
Enrichment score
0.5
0.4
NES = 1.8
Enrichment score
FDR = 0.1
0.3
NES = 1.4
0.3 -
P = 0.0001
FDR = 0.1
0.2
P= 0.005
0.2
0.1
0.1 -
0
0
0
1
5,000
10,000
15,000
I
0
5,000
10,000
15,000
Rank
Rank
GSEA: 12- vs 6-week Znrf3 cKO
IL-6-JAK-STAT3 signaling
Chemokine signaling pathway
NEŠ = 1.8
FDR = 0.1
FDR = 0.1
P = 0.001
0.4 -
0.2 -
0
0
0
5,000
10,000
15,000
Rank
Rank
Inflammatory response
Enrichment score
NES = 1.6
FDR = 0.1
0.4 -
P = 0.0001
0.2 -
0
0
5,000
10,000
15,000
Rank
Enrichment score
1
1
0
5,000 10,000 15,000
Innate immune system
Enrichment score
NES =1.45
0.4 -
FDR = 0.1
0.3 -
P= 0.007
0.2 -
0.1 -
0
=
0
1
1
5,000
10,000
15,000
Rank
d
Enrichment score
NES = 1.7
0.4 -
P= 0.003
0.2
Proportion of mice (%)
80
i
Male mice
Proportion of mice (%)
4 wk
6 wk
9 k 12 wk 24 wk
100
Histiocyte
80
area
None
60
<2%
2-10%
40
10-20%
20
20-50%
0
>50%
Control
Znrf3 cKO
Control
Znrf3 cKO
Control
Znrf3 cKO
Control
Znrf3 cKO
Control
Znrf3 cKO
a
1
1
1
significantly changed during the phenotypic transition from hyper- plasia to senescence activation or regression. This analysis revealed a set of early secreted factors with high induction at 6 weeks and a set of later secreted factors with progressively increasing expression (Fig. 3e). We cross-referenced these early and late genes with the SASP Atlas25 and SeneQuest22 and found that >75% of the factors that we identified were established SASP components. These included the early-response gene Hmgb1, which encodes a damage-associated molecular pattern molecule actively secreted upon cellular stress25 that initiates an IL-6-dependent immune response30,31. Among our top late-response genes, we identified several central SASP components, including Mmp12 (refs. 25,32), Lgals3 (refs. 33,34) and Il1rn32,35, which play key roles in ECM remodeling and immune activation.
The SASP recruits immune cells in a sex-dimorphic manner
The SASP primarily functions to induce local inflammation and attract immune cells to clear out damaged cells36,37. Given significant SASP activity in Znrf3-cKO animals, we hypothesized that immune cells would be recruited to the adrenal gland. We first tested this hypothesis by assessing broad histological changes using hematoxylin and eosin (H&E) staining. We performed H&E at each phenotypic stage from hyperplasia (4 and 6 weeks) to senescence activation (9 weeks) and regression (12 and 24 weeks). Strikingly, we observed accumulation of large multi-nucleated cells in Znrf3-cKO adrenals over time. These highly conspicuous cells first amassed in the innermost gland and became increasingly prevalent between 12 and 24 weeks in females (Fig. 3f,g). Upon pathologic review, we identified these as potential histiocytes, which are monocyte-derived immune cells known for their phagocytic role during tissue repair38,39. Interestingly, we observed an earlier and overall greater accumulation of these cells in males than in females (Fig. 3h,i), consistent with the earlier onset of senescence and regression.
Single-cell RNA-seq reveals a diverse senescence-driven immune response
Histiocytes are complex structures derived from antigen-presenting dendritic cells (DCs) and phagocytic macrophages38,39. Moreover, histio- cytes function in part to clear out neutrophils40. As a result, accumula- tion of histiocytes in Znrf3-cKO animals suggested that SASP-mediated tissue remodeling was likely driven by a multifaceted immune response. We performed single-cell RNA sequencing (scRNA-seq) to provide an unbiased profile of the immune microenvironment following senes- cence. We specifically focused on the initial switch from hyperplasia (6 weeks) to regression (9 weeks) in which we expected to identify the most proximal response. In total, we sequenced 18,294 cells from dissociated whole adrenal glands obtained from female Znrf3-cKO animals. We then used unsupervised clustering methods to define groups of cells with similar transcriptional profiles and determined the identity of each cluster based on established cell type-specific markers in conjunction with Cluster Identity Predictor41.
scRNA-seq revealed a total of 19 cell clusters (Fig. 4a,b). Expression of cell type-specific markers for each cluster is provided in Supplemen- tary Fig. 2. Within the major adrenal clusters, we observed a massive decrease in zona fasciculata cells between 6 and 9 weeks (Fig. 4c,d and Supplementary Table 2), consistent with adrenal cortex regression. In the immune compartment, we observed increases in both myeloid and lymphoid lineages. The most striking myeloid changes included increased macrophages, conventional type 1 DCs and suppressive neutrophils. Additionally, we observed a distinct cluster of neutro- phils uniquely present at 9 weeks. These cells expressed high levels of transcripts characteristic of immature neutrophils42,43, including Chil3, Ltf and Camp (Supplementary Fig. 2n). Within the lymphoid lineage, natural killer and B cells were the main increased populations. Overall, scRNA-seq uncovered a complex immune response following senescence activation with a prominent role for myeloid cells.
Given the large proportion of immune cells detected by scRNA- seq, we next performed IHC to measure in situ changes. We chose the broad myeloid marker CD68, which was highly expressed across our DC, monocyte, neutrophil and macrophage cell clusters (Supplementary Fig. 2f). While we observed some heterogeneity between animals, CD68- positive myeloid cells progressively increased over time in female Znrf3- cKO animals compared to in controls (Fig. 4e,f). In males, we observed a significant increase in CD68-positive cells at earlier stages than in females. Moreover, CD68-positive cells occupied nearly the entire male gland by 24 weeks (Fig. 4g,h). These results suggest that cellular senescence triggers a robust immune response, which is more advanced in males, to mediate clearance of damaged adrenal cortex cells.
To further investigate the origin of this profound myeloid response, we analyzed the top DEGs within the dominant myeloid cell clusters using Enrichr44. This analysis identified multiple monocyte- derived cells among the most highly enriched cell types (Fig. 4i), sug- gesting a central role for bone marrow-derived immune cells. To follow up on these observations, we performed multiplex IHC for CD11b and CD11c as well as F4/80. While these markers can be expressed by multiple cell types, CD11b and CD11c are classically expressed by monocyte-derived macrophages and DCs45-48 as compared with F4/80-positive tissue-resident macrophages49. Multiplex IHC revealed a heterogeneous immune response, with contributions from both CD11c- positive as well as CD11b-positive cell populations, and involvement of resident F4/80-positive macrophages (Fig. 4j). Large multi-nucleated histiocytes were comprised of diverse cell types, including single-, double- and even some triple-positive cells. These results highlight the diversity of immune cells activated during the senescence-driven immune response and suggest a prominent role for migratory mac- rophages and DCs.
Males exhibit higher SASP activation
While both males and females developed an inflammatory response following senescence activation, immune cells amassed earlier and to a
Fig. 4 | scRNA-seq reveals activation of innate and adaptive immune systems in response to cellular senescence and the SASP. Uniform manifold approximation and projection (UMAP) plots with cluster identification of cell types in adrenal glands of female Znrf3-cKO animals at 6 weeks (9,800 cells) (a) compared to 9 weeks (8,494 cells) (b). Adrenal zG, adrenal zona glomerulosa; adrenal zF, adrenal zona fasciculata; cDC1, conventional type 1DC; cDC2, conventional type 2 DC; MDSC, myeloid-derived suppressor cell; MoDC, monocyte-derived dendritic cell; NK, natural killer; pDC, plasmacytoid dendritic cell. c,d, Proportions of major adrenal, myeloid, lymphoid and other populations in 6-week and 9-week Znrf3-cKO samples are visualized according to the percentage of total cells. In situ validation of myeloid cell accumulation based on IHC for CD68 in control and Znrf3-cKO adrenal tissue from female (e,f) and male (g,h) cohorts. Quantification was performed using QuPath digital analysis based on the number of positive cells normalized to total nuclei. Each dot represents an individual animal. Box-and- whisker plots indicate the median (line) within the upper (75%) and lower (25%)
quartiles, and whiskers represent the range. Statistical analysis was performed on log2 transformed data using two-way ANOVA followed by Tukey’s multiple- comparison test. Scale bars, 100 um. i, Enrichment analysis using Enrichr for myeloid cell clusters (macrophages, DCs, monocytes and myeloid-derived suppressor cells) based on significant DEGs between 9- and 6-week Znrf3-cKO mice. Combined score indicates the log (Pvalue) x z score. Padj values are indicated for the respective color groups and were calculated using Fisher’s exact test. j, Multiplex immunofluorescence images of multi-nucleated myeloid clusters based on staining for CD11c (green), F4/80 (purple), CD11b (yellow) and 4,6-diamidino-2-phenylindole (DAPI) (blue). Images shown are from a 24-week-old female Znrf3-cKO animal and are representative of clusters observed in staining of five independent 24-week-old female Znrf3-cKO animals and four 12-week-old male Znrf3-cKO mice. Quantification was performed using QuPath digital analysis on manually annotated fused cell clusters based on the percentage of total nuclei in each cluster and are visualized as parts of a whole. Scale bars, 50 um.
greater extent in males. To determine whether this differential response was due to the SASP composition and/or magnitude, we performed bulk RNA-seq on male adrenals at 4 (hyperplasia), 6 (regression onset) and
9 (regression) weeks. During the phenotypic switch from hyperpla- sia to regression, we found 3,161 significant DEGs (1,558 upregulated, 1,603 downregulated; Padj < 0.05) (Fig. 5a). Although we observed many
a
9 6 week Znrf3 cKO
b
9 9 week Znrf3 cKO
c
9 6 week Znrf3 cKO
Adrenal (53,7%)
Adrenal zF (38.2%)
Endothelial
Endothelial
Adrenal zG (15.4%)
Other adrenal (0.1%)
-Neutrophils (2.5%)
1
:
NK cells (4.1%) B cells (1.7%)
Dendritic cells (3.9%)
Other lymphoid (1.1%)
-MDSCs (0.2%)
Proliferating immune
Unknown (1%)
Monocytes (1.6%)
Adrenal capsule
NK cells
CDC1
NK cells
-
Proliferating immune
CDC1
Proliferating immune (1.1%)
Macrophages (11%)
A
cDC2
Adrenal capsule
CDC2
Lymphoid cells (6.9%)
Endothelial (18.3%)
Myeloid cells (19.2%)
T cells
PDCs
B cells
B cells
Adrenal medulla
Monocytes
Adrenal medulla
T cells
PDCs
Monocytes
MoDCs
/MDSCs
-
MoDCs
Adrenal zF
Immature neutrophils
7 MDSCs
d
Adrenal zF
9 9 week Znrf3 cKO
Lymphoid cells (15.9%)
Adrenal (16.8%)
Adrenal zG
Adrenal zG
NK cells (9.6%)
Adrenal zG (9.2%)
B cells (3.3%)
Adrenal zF (7.4%)
Suppressive neutrophils
B cells
Other lymphoid (3%) Unknown (1%)
Other adrenal (0.2%)
B cells
Macrophages
Suppressive neutrophils
Macrophages
Neutrophils (7.8%)
UMAP2
Proliferating immune (4.1%)
Unknown
Unknown
Dendritic cells (12.4%)
9,800 cells
8,494 cells
UMAP1
Endothelial (16.5%)
Adrenal
Capsule
Myeloid
Macrophages
Suppressive neutrophils
Lymphoid NK cells T cells B cells
MDSCs (0.7%) Monocytes (2.9%)
Other
Adrenal zG
Monocytes
Immature neutrophils
Endothelial cells
Macrophages (22.1%)
Myeloid cells (45.9%)
Adrenal zF
MDSCs
Proliferating immune
Medulla
cDC1 cells
PDCs
Unknown
cDC2 cells
MoDCs
e
9 weeks
12 weeks
24 weeks
f
Female mice
Control
log2 (percent CD68+)
8
*P = 2.33 × 10
6
P = 0.87
** P= 3.50 × 10
-
-
-
Female
4
P = 0.20
2
Znrf3 cKO
0
-2
6 wk
9 wk
12 wk
24 wk
Control
Znrf3 cKO
-
-
-
g
9 weeks
12 weeks
24 weeks
h
Control
Male mice
*** P = 5.75 x10
log2(percent CD68+)
8
** P = 8.75 ×10
** P = 2.75 x10
-
-
-
6
= 0.88
T
Male
4
2
Znrf3 cKO
0
-2
6 wk
9 wk
12 wk
24 wk
Control
Znrf3 cKO
j
-
-
-
i scRNA-seq myeloid cluster enrichment analysis
֏ 24 weeks
CD11b+ (1.7%)
Myeloid cell clusters
CD11c+ (10.3%)
CD11c
F4/80
♀ Enrichr: Tabula Sapiens 9- vs 6-week Znrf3 cKO
Q Enrichr: CellMarker augmented 9- vs 6-week Znrf3 cKO
F4/80* (17.2%)
CD11b+CD11c+ (2.7%)
·F4/80+CD11b+ (6.8%)
-F4/80+CD11c+ (18.9%)
Skin, macrophage
P value
Megakaryocyte-erythroid
progenitor, germ
P value
1.25 × 10
-7
F4/80+CD11c+ CD11b+ (20.6%)
Lymph node, macrophage
8.84 x 10
2.56 ×10
Granulocyte-monocyte
progenitor, germ
4.88 × 10
Bone marrow, monocyte
Langerhans cell, skin
8.93 x 10
Negative (21.9%)
6.38 x 10
CD11b
Merge
Lymph node,
non-classical monocyte Tongue, Schwann cell
Mesenchymal stem cell
o 12 weeks
Circulating precursor cell,
CD11b+ (0.4%)
peripheral blood
CD11c+ (38.8%)
Liver, liver dendritic cell -
Periodontal ligament stem cell
Lung, capillary aerocyte
Dendritic cell, bone marrow
F4/80* (8.3%)
CD11b CD11c (1.8%)
Lung, basophil
Monocyte, fetal kidney
F4/80*CD11b+ (1%)
.F4/80+CD11c+ (28.4%)
0
100
200
300
F4/80+CD11c+CD11b+ (3.1%)
0
500
1,000
1,500
CD11c F4/80 CD11b DAPI
Negative (18.1%)
Enrichr combined score
Enrichr combined score
a
Hyperplasia
Regression
b
Znrf3 cKO
4 weeks
6 weeks
9 weeks
3
Age
Pathway analysis: 6- vs 4-week Znrf3 cKO
LE
2
4 weeks
Inhibited
Activated
6 weeks
Kinetochore metaphase signaling pathway
-
1
9 weeks
Mitotic roles of polo-like kinase
0
Cyclins and cell cycle regulation
Estrogen-mediated S phase entry
-1
Cell cycle regulation by BTG family proteins
P value
-2
Cell cycle control of chromosomal replication
1 × 10-16
-3
Pyridoxal 5’-phosphate salvage pathway
1 × 10-12
Natural killer cell signaling
1 × 10-8
·
1 × 10-4
Senescence pathway
TREM1 signaling
Production of nitric oxide and
reactive oxygen species in macrophages
IL-8 signaling
HIF-1a signaling
Phagosome formation
-3.00
-2.75
-2.50
-2.25
-2.00
2.0
2.5
3.0
3.5
4.0
Activation Z score
Transcriptional regulators: 6- vs 4-week Znrf3 cko d 36-
Cytokines: 5 6- vs 4-week Znrf3 cKO
e
GSEA: 3 6- vs 4-week Znrf3 cKO
Inhibited
Activated
Inhibited
Activated
Enrichment score
IL-6-JAK-STAT3 signaling
TBX2
P value
₡ CSF2
P value
0.5
NES = 1.8”
Enrichment score
Chemokine signaling pathway
FOXM1
0.4 -
CCND1
1 × 10
CX3CL1
NES = 1.7
26
1 × 10
-36
0.4
-
FDR = 0.1
FDR = 0.1
MYC
1 × 10
20
IL-3
-26
TNFSF10
1 × 10
~16
0.3
P = 0.001
0.3
-
P= 0.002
CEBPB
1 x 10
·
1× 10
0.2
0.2
-
E2F3
1 × 10
9
IL-13
1 × 10
6
E2F1
TNFSF12
0.1
-
0.1
-
E2F2
CD40LG
0
0
FOXO1
IL-18
I
1
1
I
CITED2
CSF3
0
5,000
10,000
15,000
1
1
I
1
0
5,000
10,000
15,000
MEF2D
CSF1
Rank
EGR1
Rank
TRPS1
IL-33
SMARCA4
₡‘IL-4
Enrichment score
0.6
Inflammatory response
Enrichment score
Innate immune system
SPI1
0.4
NES = 1.9
CEBPA
₡IL-6
NES =2.3
-
IL-1a
0.4
FDR = 0.1
FDR = 0.1
-
* CDKN2A
P = 0.0006
0.3
-
P = 0.004
KDM5B
₡ TNF
* TP53
₡IL-13
0.2 -
0.2
-
NUPR1
₡IFN-Y
0.1
-
0
-
-
-
-5
-4
-3
3
4
5
0
-
6
-2.80
-2.75
-2.70
-2.65
2.0
2.5
3.0
3.5
4.0
1
1
1
A
1
1
I
I
Activation Z score
Activation Z score
0
5,000
10,000
15,000
0
5,000
10,000
15,000
Rank
Rank
f
Top genes encoding senescence-associated secreted proteins
g
Tfpi
Top SASP genes: male vs female
Tgfbi
4 weeks
6 weeks
9 weeks
Btd
Adgre5
C1qbp
C1qtnf6
Anxa1
Ctrl
Apln
C2
♀
♂
Cd200r1
Creg1
Mmp12
C4b
Lgals3
2
Znrf36
Cd63
Ctss
Fabp5
Control
Col4a5
Clu
Il7r
Efna1
Dpp7
Emid1
Il1rn
Epor
Illa
1
CKO
Col4a6
Lgals3
Creg1
Fstl3
Col5a1
Mmp12
Esm1
Pla2g7
0
Age
Ghr
Cd200r1
Hmgb1
Crtap
Tnfrsf1b
Fabp5
Kars
Gc
IL7r
-1
Ggh
Tnfsf13b
Grn
Ctss
4 weeks
Lgals1
Il1rn
Igf1
Trem2
-2
6 weeks
Mfge8
Trem2
Mif
Il18bp
lgf1
Gr
9 weeks
Mmp19
Illa
Kazald1
Anxa1
S100a13 Adgre5
Nsun2
Nxph1 Pdgfd
Lnpep
Loxl2
Gc
Mmp11
Plau
Pros1
Pla2g7
IL18bp
Rnpep
Mmp11
Scube1
Plau
Efna1
Tnfrsf1b
Scube3
Pros1
Smpdl3a
S100a13
Selenop
Selenop
Tgfa
Txndc16
Smpdl3a
Apln
Esm1
Wnt10b
Epor
Wnt10b
Tgfbi
h
Kazald1
TD139 (15 mg/kg) i.p. every 3 d
Lnpep
No. phagocytosis beads/DAPI
Col5a1
Clu
Pros1
10
-6
Tnfsf13b
6 weeks
**** P =2.55 x10
Smpdl3a
12 weeks
log2 normalized
Crtap
Ggh
Col4a6
₫ SF1Cre/+; Znrf3fl/f\; R26RmTmG/WT
5
Loxl2
*
Emid1
*
*
*
*
*
*
*
*
*
GFP (adrenal cortex lineage)
0
*
*
*
*
tdTomato (non-cortex cells)
*
*
*
*
*
Internalized phagocytosis beads
*
DAPI (nuclei)
-5
*
*
*
Vehicle
TD139
*
*
*
*
*
*
Vehicle
TD139
sex-specific differences in gene expression (Extended Data Fig. 4), IPA identified several canonical pathways that regulate cell division as the most significantly inhibited (Fig. 5b), mirroring what we observed in females. Moreover, cellular senescence was among the top activated pathways (Fig. 5b), and the senescent mediators CDKN2A and TP53 were two of the top activated transcriptional regulators (Fig. 5c). Consistent
with the strong immune phenotype and massive accumulation of phagocytic histiocytes in males, multiple immune-related signatures as well as phagosome formation were among the top activated pathways (Fig. 5b). This was similar to the later-stage response that we observed in females at 12 weeks (Extended Data Fig. 5a). When we assessed the male cytokine profile (Fig. 5d and Extended Data Fig. 5b), we found
Fig. 5 | Senescence-mediated immune activation in male Znrf3-cKO mice is characterized by higher SASP induction. a, Heatmap of the top 1,000 DEGs in adrenals from Znrf3-cKO animals at 4, 6 and 9 weeks of age (Padi < 0.05). IPA identified the top inhibited (a) and activated (b) canonical pathways, transcriptional regulators (c) and cytokines (d) in adrenals from male Znrf3-cKO animals at 6 versus 4 weeks (P < 0.05). Data are representative of four biological replicates per group, and statistical analysis in IPA was performed using a right- tailed Fisher’s exact test. BTG, B cell translocation gene; HIF-la, hypoxia-inducible factor 1-alpha; TREM1, triggering receptor expressed on myeloid cells 1. e, GSEA for inflammatory signatures (IL-6-JAK-STAT3 signaling, chemokine signaling, the inflammatory response and the innate immune system) identifies positively enriched pathways in adrenals from male Znrf3-cKO animals at 6 versus 4 weeks. f, Heatmap representing supervised hierarchical clustering of the 40 most significant DEGs that encode secreted proteins in adrenals from male Znrf3-cKO
animals compared to those of controls; statistical analysis was performed using the Wald test (Padj<0.05). g, Comparison of the top SASP genes in males and females identified 12 shared factors induced in both sexes. h, Pharmacological inhibition of galectin 3 (encoded by Lgals3) with the selective inhibitor TD139 significantly reduced phagocytic activity in adrenal tissue from male Znrf3-cKO mice containing the R26RmTmG lineage reporter. Quantification of internalized phagocytosis beads (asterisks) was performed using QuPath digital analysis based on the number of fluorescent beads normalized to DAPI (nuclei) values. Auto-fluorescent lipid content (white) is present in large fused clusters. Three adrenal tissue slices from four independent animals per group were analyzed. i.p., intraperitoneally. Box- and-whisker plots indicate the median (line) within the upper (75%) and lower (25%) quartiles, and whiskers represent the range. Statistical analysis was performed using two-tailed Student’s t-test. Scale bars, 20 um.
substantial overlap with females, with eight of the top ten activated cytokines shared in both sexes. However, although the profile was similar, males consistently displayed a higher magnitude of induc- tion, including IL-1B (Z score =5.92), IFN-y, (Z score = 7.54) and TNF (Z score = 6.83). This was further confirmed with enhanced GSEA enrich- ment of immune-related signatures (Fig. 5e and Extended Data Fig. 5c). Moreover, unbiased analysis of the top DEGs that encode secreted proteins (Fig. 5f) revealed that the male SASP shared several common factors with that of females (Fig. 5g). In sum, these data suggest that both male and female Znrf3-cKO animals develop a functional SASP following senescence activation. However, males undergo the pheno- typic switch earlier and display a higher magnitude of SASP induction, ultimately leading to an enhanced immune response.
The activation of SASP factors common to both sexes suggested that these components were likely to play a key role in senescence- driven immune remodeling. To prioritize candidates to functionally interrogate, we looked for factors that (1) were expressed in the adrenal cortex in our scRNA-seq data, (2) have a known function in phagocytosis and (3) are targetable using a small-molecule inhibitor. Based on these criteria, our top candidate was Lgals3, which encodes galectin 3, a mem- ber of the ß-galactoside-binding family of lectins. Moreover, galectin 3 is known to facilitate macrophage phagocytosis50 and can be selectively inhibited using TD139 (ref. 51). We administered TD139 to male Znrf3- cKO mice beginning at 6 weeks when immune cells had just begun to infiltrate. We then isolated adrenals from vehicle- and TD139-treated animals at 12 weeks and performed an ex vivo phagocytosis assay using fresh slices of adrenal tissue. This approach allowed us to preserve the native tissue architecture and avoid disruption of critical cell-to-cell interactions. We incubated adrenal tissue slices with pH-sensitive, dye-conjugated Escherichia coli particles. These particles are non- fluorescent outside the cell and only fluoresce once inside a phagosome with the shift from neutral to acidic pH52. Using this new assay system, we found a significant reduction in the number of internalized particles in adrenal slices from mice treated with TD139 as compared to those treated with vehicle control (Fig. 5h). These results demonstrate that
inhibition of galectin 3 significantly impairs phagocytic activity in our model.
Androgens amplify the SASP-mediated immune response
The elevated SASP in males implicated a potential male- and/or female- specific factor capable of enhancing or repressing SASP target gene expression, respectively. However, prior studies in the normal adrenal cortex have shown that androgens suppress proliferation of stem and pro- genitor cells after puberty53, which is consistent with a role for androgens in promoting mechanisms related to cell cycle arrest. To functionally test whether androgens promote senescence, we castrated 4-week-old male control and Znrf3-cKO mice to lower androgen levels (Fig. 6a). At 9 weeks, we observed more than a threefold increase in adrenal weight in cas- trated Znrf3-cKO animals compared to that in sham controls (Fig. 6b,c). This robust increase in adrenal size was associated with both retention of proliferative adrenal cortex cells (Fig. 6d) as well as a near-complete block in the myeloid response, including neutrophils (Extended Data Fig. 6a) as well as both CD11c-positive and CD11b-positive cells (Fig. 6e). Consistent with the significant reduction in antigen-presenting cells, T cell infiltration was significantly lower in castrated animals than in sham controls (Extended Data Fig. 6b). Tissue-resident macrophages marked by F4/80 (ref. 49) were androgen responsive yet remained rela- tively abundant following castration (Fig. 6e). However, their distribution was more dispersed, similar to that of the normal adrenal cortex. Intrigu- ingly, some senescent markers, including p16 and p21, were unchanged or even elevated in adrenals from castrated animals compared to those of sham controls (Fig. 6d). These results suggest that the dominant impact of androgens on senescence occurs in the later immune-mediated stages as opposed to in the initial cell cycle arrest. To further tease apart the role of androgens in early versus late senescent responses, we castrated male Znrf3-cKO mice at 6 weeks, when hyperproliferation has normally been suppressed. We analyzed mice 6 weeks later and found a significant block in the myeloid response yet no increase in proliferation (Extended Data Fig. 6c-g). These experiments suggest that androgens primarily regulate the late senescent response.
Fig. 6 | Androgen deprivation restricts immune infiltration in male adrenals. a, Male control or Znrf3-cKO mice were castrated at 4 weeks of age to lower androgen levels. b,c, At 9 weeks of age, adrenal glands from castrated mice were significantly larger than those from sham-operated controls. b, Representative gross histology images are shown. Scale bars, 1 mm. c, Normalized adrenal weight. Each dot represents an individual animal. Box-and-whisker plots indicate the median (line) within the upper (75%) and lower (25%) quartiles, and whiskers represent the range. Relative fold change is indicated below each group. Senescence markers (Ki67, p16 and p21) (d) and myeloid immune markers (CD68, CD11c, CD11b and F4/80) (e) in adrenals from sham-operated versus castrated animals. Representative images are shown for each group. Scale bars, 20 um. Quantification was performed using QuPath digital image analysis based on the number of positive cells normalized to total nuclei. Each dot represents an individual animal. Box-and-whisker plots represent mean with variance
across quartiles. Statistical analysis was performed using two-tailed Student’s t-test (c,d) or two-way ANOVA (e) followed by Tukey’s multiple-comparison test. IPA of RNA-seq data from adrenals of 9-week-old males identified the top inhibited and activated canonical pathways (f) and cytokines (g) in adrenals from castrated versus sham-operated Znrf3-cKO animals; statistical analysis in IPA was performed using a right-tailed Fishers exact test (P < 0.05). CREB, cAMP response element-binding protein; FAK, focal adhesion kinase; LXR, liver X receptor; RXR, retinoid X receptor. h, Heatmap representing supervised hierarchical clustering of the 40 most significant DEGs that encode secreted proteins in adrenals from castrated compared to sham-operated Znrf3-cKO mice; statistical analysis was performed using the Wald test (Padj < 0.05). Asterisks indicate genes previously identified as shared SASP components common to both male and female Znrf3- cKO animals. Data are representative of four biological replicates per group, and statistical tests were performed using IPA.
To further examine the mechanism by which androgens regu- late senescence, we performed bulk RNA-seq on whole adrenals from 9-week-old castrated and sham-operated Znrf3-cKO animals. We found that castration reversed many of the top canonical pathways (Fig. 6f), cytokines (Fig. 6g) and genes encoding secreted proteins (Fig. 6h) that were previously enhanced in males compared with females. In sum,
these experiments demonstrate that androgens amplify the SASP- mediated immune response.
Senescence-driven tissue remodeling leads to adrenal tumors Based on high levels of immune infiltration and adrenal degenera- tion, we predicted that Znrf3-cKO animals would succumb to adrenal
a
b
Sham
Castrated
Control
Znrf3 cKO
c
Normalized adrenal weight (mg) per body weight (g)
4
₫ 9 weeks
*** P=1.68 x 10-4
3
Sham
Castrated
2
1
*** P = 2.26 ×10-4
0
Control
Znrf3 cKO 3.10
Fold change:
1.62
d Senescence markers (9-week-old Znrf3-cKO mice)
2
**** P=9.37 x 10°
-5
Ki67
Percent Ki67+ (log2 normalized)
1
Sham
0
-1
-2
Castrated
Sham
Castrated
p16
Percent p16+ (log2 normalized)
3
P = 0.07
2
Sham
1
0
-1
Castrated
-2
Sham
Castrated
p21
3
Percent p21+ (log2 normalized)
*P =1.21 ×10
-2
Sham
2
1
Castrated
0
-1
Sham
Sham
Castrated
Castrated
e
Immune markers (9-week-old control and Znrf3-cKO mice)
CD68
CD11c
CD11b
F4/80
Sham
Control mice
Castrated
Sham
Znrf3- cKO mice
Castrated
*** P = 8.2 ×10-4
P = 0.76
**** P = 6.31 ×10-7
**** P = 2.45 × 10-13
P = 0.69
**** P = 1.18 × 10-8
*** P = 1.58 × 10-4
*P = 1.12 ×10-2
Percent CD68+ (log2 normalized)
6
**** p = 7.03 × 10-9
(log2 normalized)
**** P = 5.55 × 10-6
Percent CD11b+ (log2 normalized)
4
**** P = 3.59 x 10-8
Percent CD11c+
6
Percent F4/80+
(log2 normalized)
6.5
**** P = 4.21 × 10-6
** P =
4
** P = 6.85
2
6.0
× 10-3
4
P = 0.22
5.5
1.17 x10
3
2
0
P = 0.99
2
5.0
·
-2
0
4.5
0
-4
4.0
-2
Znrf3 cKO
-2
-6
3.5
Control
Control
Znrf3 cKO
Control
Znrf3 cKO
Control
Znrf3 cKO
Sham
Castrated
Sham
Castrated
f
Pathway analysis: 9-week-old Znrf3-cKO castrated vs sham mice
Inhibited
Activated
S100 family signaling pathway
P value
Phagosome formation
1 × 10 -12
1 × 10
-9
CREB signaling in neurons
1 × 10
-6
1 × 10
3
FAK signaling
Pathogen-induced cytokine storm signaling
pathway
G protein-coupled receptor signaling
Breast cancer regulation by stathmin 1
Inhibition of matrix metalloproteases
Apelin adipocyte signaling pathway
Cholesterol biosynthesis I
Cholesterol biosynthesis II
(via 24,25-dihydrolanosterol)
Cholesterol biosynthesis III (via desmosterol)
LXR/RXR activation
Antioxidant action of vitamin C
-7.0
-6.0
-5.0
2.0
2.4
2.8
3.2
Activation Z score
g
Cytokine analysis: 9-week-old Znrf3-cKO castrated vs sham mice
IL-33
Inhibited
Activated
TNF
IL-4
P value
1 × 10
-14
CSF2
1 × 10
-10
IL-13
1 × 10
6
TNFSF11
· 1 x 10
2
IFN-Y
IL-6
CSF1
CCL20
PF4
IL-13
IL-18
IL-2
IL-17a
IL-1a
IL-3
CNTF
SCGB1A1
IL-37
IL-1RN
-5
-4
-3
1.5
2.0
2.5
Activation Z score
h
Top genes encoding senescence-associated secreted proteins
Fabp5
NTF.
Anxa1
Castrated
Pla2g7
Creg2
0.5
LA
Hpse Mim.
0
Mmp12.
F10
-0.5
Lgals3 + Timp1
-1.0
Trem2 .
lgf1
Čd63
-1.5
Pilra
Ccl3
Csf3r Ctss
Cd200r1
Creg1
Apoe Nia
Prss.35
Nxph1
Hmgb1 .
Ephb6
Č4b
Scube1
Masp1
Tafa5
Btcl
Ccdc126
Tulp3
Pdgfd
Hdgf
Tfpi
clu
Tnfsf13b + Col4a5 .
-
4
0
[
C
0
0
0
-
1.5
Condition Sham
1.0
| Castration Sham control procedure or | Harvest adrenals |
|---|---|
| P0 4 weeks | 9 weeks |
a
c
Adrenal weight (mg) per body weight (g)
Female mice
. Control
30
· Znrf3 cKO
Control
25
Normal aged adrenal
Male
Female
20
15
n = 9 n = 11
10
6
4
2
0
-
T
Atrophic
9
Male 3/12
Female 3/21 (14.3%)
4
6
12
24
44
52
78
Age (weeks)
(25%)
b
Adrenal weight (mg) per body weight (g)
Male mice
Znrf3 cKO
. Control
Benign
Male 8/12 (66.7%)
Female
4
Znrf3 cKO
5/21
(23.8%)
3
2
…:
1
Malignant
Male 1/12
Female
0
$
.
&
.
+
2
9
?
13/21 (61.9%)
4
6
9
12
24
44
52
78
(8.3%)
Age (weeks)
d
78-week female mice
e
78-week male mice
f
H&E
GFP DAPI
· Atrophic
?
**** P = 4.10 x 10
-B
**** P =2.98 x10
5
8
Benign
· Atrophic
8
· Benign
Proliferation
Ki67 proliferation index (log2 transformed)
Ki67 proliferation index (log2 transformed)
Primary tumor
· Malignant
6
*P=0.038
6
*P= 0.038
· Malignant
P = 0.99
** P = 2.18 x10
-3
4
4
2
.
2
.
0
0
.
Liver metastasis
-2
.
-2
Control
Znrf3 cKO
Control
Znrf3 cKO
g
78-week female mice
h
78-week male mice
Normalized adrenal weight (log2 transformed)
**** P = 8.70 ×10
-5
· Atrophic
Normalized adrenal weight (log2 transformed)
j
· Atrophic
9 benign tumors
**** P= 3.02 ×10
-11
Benign
P= 0.28
+ Benign
Ki67 proliferation index (log2 transformed)
6
· Malignant
6
· Malignant
4
** P = 9.58 × 10-3
(log2 transformed)
8
** P = 4.85 × 10-3
** P = 7.09 x 10
-3
**** P = 8.25 x10
9
4
*P=0.046
4
Weight
CD68 index
P = 3.01 ×10
-5
3
6
2
2
¥
2
4
0
.
0
-2
-2
1
2
-4
-4
0
0
Control
Znrf3 cKO
Control
Znrf3 cKO
Nodule
Gland
Nodule
Gland
i
H&E
CD68
Ki67
k
§ malignant tumors
30
Control
Normal aged adrenal
Ki67 proliferation index (%)
R = 0.54
**** P = 6.73 ×10
-5
20
Atrophic
10
0
Nodule
Nodule
Nodule
0
10
20
30
Znrf3 cKO
Benign
CD68 index (%)
Region A Region B Region C
Malignant
Region A
Region A
Region A
CD68
Region B
Region B
Region B
Ki67
Region C
Region C
Region C
insufficiency. To test this hypothesis, we generated late-stage cohorts at 44, 52 (1 year) and 78 weeks (18 months) (Fig. 7a,b). Surprisingly, rather than adrenal failure, a large proportion of Znrf3-cKO animals
ultimately developed adrenal tumors (Fig. 7c). Notably, tumor develop- ment and progression were highly sex dimorphic, with females exhibit- ing increased malignancy.
Fig. 7| Following senescence-mediated remodeling of the tissue microenvironment, metastatic ACC tumors arise in a sex-dimorphic manner. Adrenal weight measurements in cohorts of female (a) and male (b) control and Znrf3-cKO mice, including at 44, 52 and 78 weeks of age. After 52 weeks, adrenal tumors of varying sizes begin to form. Each dot represents an individual animal. The black line indicates the mean. c, Representative images of the gross histology (left) and H&E staining (middle and right) from each pathology observed at 78 weeks of age. Blinded samples were scored as atrophic, benign or malignant. The respective proportion of each pathology in male compared to female cohorts is indicated. Scale bars, 1 mm (left), 200 um (middle);, 500 um (right). d,e, Proliferation as measured by the Ki67 index is significantly higher, with progression to benign and malignant tumors in Znrf3-cKO mice. Each symbol represents an individual animal. The black line indicates the mean. Statistical analysis was performed using one-way ANOVA followed by Tukey’s post hoc test. f, Malignant tumors metastasize to distant sites, including the liver. Representative images are shown from mice expressing the R26RmTmG lineage reporter, which permanently labels adrenal cortex cells with GFP. Scale bars,
100 um. g,h, Tissue weight is significantly higher, with progression to benign and malignant tumors in Znrf3-cKO mice. Each symbol represents an individual animal. The black line indicates the mean. Statistical analysis was performed using one-way ANOVA followed by Tukey’s post hoc test. i, Representative H&E (left), CD68 (middle) and Ki67 (right) staining in atrophic, benign and malignant adrenals from Znrf3-cKO 78-week-old mice. Scale bars, 500 um (whole tissue), 100 um (high-magnification inset). j, In Znrf3-cKO adrenals, benign nodules have a significantly higher Ki67 index and lower CD68 index than the background gland. Each dot represents an individual nodule. Box-and-whisker plots indicate the median (line) within the upper (75%) and lower (25%) quartiles, and whiskers represent the range. Statistical analysis was performed using two-tailed Student’s t-test. k, In malignant tumors, there is a significant inverse correlation between the Ki67 index and the CD68 index. Each dot represents a distinct tumor region. Colors represent distinct regions from the same tumor. Statistical analysis was performed using simple linear regression (n = 23 tumor regions from six animals).
To evaluate tumors without bias, we performed H&E and Ki67 staining. Blinded samples were then pathologically reviewed based on criteria derived from the Weiss system conventionally used to evaluate adrenal cortical neoplasms in adult patients54,55. This analysis revealed three main phenotypes in Znrf3-cKO adrenals: atrophic glands, benign lesions and malignant tumors (Fig. 7c). Atrophic glands were character- ized by high histiocyte accumulation (Fig. 7c) and an overall Ki67 index at or below that of age-matched controls (Fig. 7d,e). This pathology was observed in 14.3% of females (three of 21 mice) and 25% of males (three of 12 mice). Benign tumors were most prevalent in males (66.7%, eight of 12 mice). These lesions were characterized by formation of discrete nodules contained within the encapsulated adrenal cortex (Fig. 7c) with a significantly higher Ki67 index than that of matched controls (Fig. 7d,e). Finally, malignant tumors were most prevalent in females (61.9%, 13 of 21 mice). We conservatively classified tumors as malignant strictly based on the presence of metastases, which occurred to clini- cally relevant sites including the lung and the liver (Fig. 7f). However, malignant tumors also had a significantly higher Ki67 index (>10%) (Fig. 7d,e) and were significantly larger in size (Fig. 7g,h). Overall, these findings demonstrate that Znrf3 loss is permissive for ACC in a sex-dependent manner with advanced aging.
Reduced myeloid cells are predictive of poor patient outcome Our Znrf3-cKO model presented two striking sexual dimorphisms. Males exhibited an earlier, more robust immune response following senescence and were less likely to develop malignant tumors. Con- versely, females displayed a dampened immune response and were sig- nificantly more likely to develop metastatic tumors. To further examine the relationship between the immune microenvironment and tumor progression, we quantified CD68-positive myeloid cells in aged animals at 44 and 52 weeks of age before the onset of tumors and in atrophic, benign and malignant tumors at 78 weeks. With increased aging, female Znrf3-cKO animals continued to accumulate histiocytes (Extended Data Fig. 7a,b) yet retained a greater proportion of adrenal cortex cells than males (Extended Data Fig. 7c,d). Control adrenals also displayed
an age- and sex-dependent accumulation of CD68-positive cells at the cortical-medullary boundary (Extended Data Fig. 7e-h), which is where the longest-lived adrenal cortex cells reside. These results are consist- ent with myeloid cell recruitment in the adrenal gland being driven by age-related mechanisms. At 78 weeks, we observed a heterogeneous immune landscape, dependent on the tumor stage. Atrophic glands exhibited a high CD68 index, consistent with the accumulation of histio- cytes noted in H&E staining (Fig. 7i and Extended Data Fig. 7i). In benign cases, there was an inverse correlation between proliferation and mye- loid cell infiltration. Nodules displayed a significantly higher Ki67 index and lower CD68 index than those of the background gland (Fig. 7i,j and Extended Data Fig. 7j). This apparent exclusion of myeloid cells in more proliferative regions was also evident in malignant tumors. Here, we took advantage of high intertumoral and intratumoral heterogeneity and analyzed CD68-positive myeloid cells across different tumors as well as subpopulations. We found that regions with the highest Ki67 index had the lowest CD68 index (R2 = 0.54, **** P = 6.73 x 10-5) (Fig. 7i,k). In sum, these results suggest that, as adrenal tumors become more aggressive, the immune microenvironment is remodeled to exclude tumor-suppressive myeloid cells.
Given these observations, we next sought to assess myeloid cell infiltration in ACC patient tumors. A myeloid response score (MRS) was recently established in hepatocellular carcinoma that predicts patient prognosis and therapy response56. We evaluated expression of each hepatocellular carcinoma MRS gene in our scRNA-seq dataset and identified four genes (Cd33, Cd68, Itgam, Msr1) that were myeloid specific (Fig. 8a-d) and for which the combined expression represented the full adrenal myeloid compartment (Fig. 8e). We then evaluated the adrenal MRS (AMRS) in ACC tumors from The Cancer Genome Atlas (TCGA) (TCGA-ACC) in conjunction with established markers of poor prognosis (high MKI67 (ref. 57) and low G0S2 (ref. 58); Fig. 8f). Consistent with our Znrf3-cKO model, females had a significantly lower AMRS than males (Fig. 8g). Moreover, low AMRS was associated with significantly reduced overall (Fig. 8h) and progression-free survival (Fig. 8i), even in females alone (Extended Data Fig. 8). To compare
Fig. 8 | A low myeloid response is associated with worse patient outcome in ACC. UMAP of scRNA-seq data from adrenal tissue of female Znrf3-cKO animals. Expression of individual myeloid cell markers Cd33(a), Cd68 (b), Itgam (also known as CD11b) (c) and Msr1 (also known as CD204) (d) is shown. e, The combined expression of these four markers was used to generate an AMRS. Exp, expression. f, Analysis of TCGA-ACC data, including patient demographics and outcome, in accordance with mRNA expression of AMRS genes (CD33, CD68, ITGAM, MSR1) and established prognostic markers (high MKI67 (proliferation) and low G0S2 (recurrent disease)). AMRS is significantly lower in female than in male TCGA patients with ACC (g), and low AMRS is associated with shorter overall (h) and progression-free (i) survival.j, In primary human ACC tumors, CD68+
histiocytes are found in a subset of cases. k, A representative example region is shown relative to a Znrf3-cKO mouse adrenal tumor. l,m, The infiltration of CD68+ myeloid cells in primary human ACC tumors is heterogeneous. Regions with a high Ki67 index (>15%) have a significantly lower CD68 index than regions of low proliferation (Ki67 <5%). IHC quantification was performed using QuPath digital analysis based on the number of positive cells normalized to total nuclei. Tumor- associated stroma and highly necrotic regions were excluded from analysis. Each dot represents a distinct tumor region (n = 50 tumor regions from 38 patients). Statistical analysis was performed using two-tailed Mann-Whitney test (g), log- rank Mantel-Cox test (h,i) or two-tailed Student’s t-test (m). Scale bars, 100 um.
the AMRS to that of individual immune cell populations, we derived signatures for populations of interest based on established cell type- specific markers that were robustly expressed in TCGA-ACC data. We
found that signatures of individual myeloid cell populations, includ- ing macrophages, DCs and neutrophils, failed to predict both overall and progression-free survival (Supplementary Table 3). By contrast, a
a
b
c
d
e
AMRS
Cd33
4.0
Cd68
6.0
Itgam
Msr1
5.0
Combined
6.0
6.0
0
0
0
0
0
log2 exp
log2 exp
log2 exp
log2 exp
log2 exp
UMAP2
UMAP1
f
TCGA-ACC
Sex
Overall survival status
Progression-free status
CD33
Adrenal
ITGAM
myeloid
MSR1
response genes
CD68
Prognostic markers
G0S2
MKI67
Sex Female
Male
Overall survival status Living
Deceased
Progression-free status
Censored
Progression
Expression heatmap
-3
3
g
Probability of survival (%)
TCGA-ACC: overall survival
i
AMRS
TCGA-ACC: progression-free survival
30
100
100
Expression Z score
AMRS
Probability of survival (%)
*P = 2.30 × 10-2
AMRS
20
- High
- High
Low
Low
*P = 1.89 × 10-2
10
50
50
** P =1.98 ×10-3
0
-10
0
0
Male Female
0
50
100
150
0
50
100
150
Time (months)
Time (months)
j
H&E
CD68
k
H&E
CD68
Primary human ACC tumor
Znrf3-cKO mouse ACC tumor
l
H&E
CD68
Ki67
m
high-proliferation region
8
P = 1.63 x 10
4
Representative
CD68 index
(log2 transformed)
6
4
2
0
-2
Proliferation index
Representative
low-proliferation region
High Low (>15%) (<5%)
signature of monocyte-derived cells was strongly predictive of patient outcome and was significantly higher in males than in females, consist- ent with our AMRS and the collective importance of multiple recruited myeloid cell populations.
Our results implicated myeloid cells as a key component of the ACC tumor microenvironment, yet ACC tumors are classically described as immune poor59. However, this notion is primarily based on evalu- ation of lymphoid cells60, and myeloid cells have thus far remained unexplored. To extend the translational relevance of our findings, we obtained formalin-fixed, paraffin-embedded (FFPE) primary tumor samples from 38 patients with ACC. We found that all primary tumor specimens contained CD68-positive myeloid cells, with an average overall index of 11.52% (Supplementary Table 4). Moreover, a subset of cases (seven of 38 tumors or 18.4%) contained distinct histiocytes, similar to what we observed in tumors from Znrf3-cKO mice (Fig. 8j,k). In addition, we found that tumor regions with a high Ki67 index (>15%) had a significantly lower CD68 index than regions of low prolifera- tion (Ki67 <5%) (Fig. 8l,m), mirroring what we observed in our mouse model. In sum, our results reveal an important role for myeloid cells in restraining ACC tumor progression with substantial prognostic value.
Discussion
Our data demonstrate that loss of the Wnt inhibitor ZNRF3 is permis- sive for metastatic adrenal cancer with advanced aging. However, genetic Znrf3 loss is not completely sufficient and depends on addi- tional extrinsic factors, including sex and the aged immune microen- vironment. These factors are closely linked to cellular senescence, an age-associated stress response activated in ZNRF3-deficient adrenals. Our in vivo characterization of senescence-induced tissue remodeling reveals a prominent role for myeloid-derived immune cells during the phenotypic switch from hyperplasia to regression. Males activate senescence earlier and mount a greater immune response, mediated in part by androgens, which ultimately results in fewer metastatic tumors. We translate these findings to human patients with ACC, in whom a high myeloid response is associated with better outcome. Collectively, our model recapitulates the age and sex dependence of ACC and reveals a new role for myeloid cells in restraining adrenal cancer progression.
Sex-specific differences in stem cell renewal, the tissue micro- environment and innate and adaptive immune responses have been increasingly acknowledged as key factors in both aging61 and cancer62, including immunotherapy response63. Our model, in which males exhibit an advanced senescent response and subsequently become more protected from malignancy, provides a system to interrogate sex-dependent mechanisms of immunosurveillance. While there are multiple underlying differences between males and females, we dem- onstrate a specific role for male androgens in amplifying the SASP to enhance recruitment and activation of monocyte-derived mac- rophages and DCs. These results are consistent with prior work showing that androgens activate cellular senescence in vitro64,65 and recently published observations that androgens enhance the phagocytic activ- ity of macrophages66. In our model, we found that androgen depriva- tion had no effect on the initial cell cycle arrest but instead blocked manifestation of a full senescent phenotype by suppressing the SASP. This not only reduced macrophages but also DCs and other myeloid cells. The adaptive immune response and T cell infiltration were also impaired, likely due to the loss of antigen-presenting cells, and consist- ent with recent reports of senescence promoting anti-tumor immunity in other cancers67,68.
Following extensive immune remodeling and advanced aging, Znrf3-cKO animals ultimately develop adrenal tumors at 12-18 months. This corresponds to peak ACC incidence at an age of 45-55 years7,8. Moreover, metastatic tumors in our model were more common in females, which is analogous to the higher prevalence of ACC in women7. In Znrf3-cKO animals, tumor development requires escape from senes- cence-mediated tissue degeneration. While our studies focus on ZNRF3
as a common ACC tumor suppressor, we expect that tumorigenesis is influenced by additional genetic mutations, particularly within genes that bypass cellular senescence and are recurrently altered in patients (for example, TP53, CDKN2A and RB1 (refs. 10,11)). However, in addition to mutational events, we also expect that age-dependent changes in the tissue microenvironment substantially impact the func- tional consequence of acquired mutations. In particular, we observe a striking exclusion of myeloid cells with advanced tumor progression, suggesting an age-dependent conversion of the SASP from an anti- tumorigenic to a pro-tumorigenic process. These findings translate to human patients and align with the immune-poor nature of ACC59 as well as the modest efficacy of immunotherapy thus far69. Given that a strong intratumor IFN-y signature correlates with better immunotherapy response in other cancers70-72, exploiting underlying mechanisms that regulate immune cell recruitment in our model may ultimately help enhance the efficacy of immune-based therapy for ACC and other immune-cold tumors.
Overall, our Znrf3-cKO model is a highly unique in vivo system that highlights the pleiotropic effects of cellular senescence in can- cer. Traditionally, cellular senescence has been studied as a tumor- suppressive mechanism that halts proliferation of damaged cells at risk for neoplastic transformation73. However, more recent studies reveal that some senescent cells may be released from their growth arrest and actually promote more aggressive phenotypes in the long term74,75. In our model, tumors may potentially arise from cells that evaded senescence or from cells released from senescence. Newly developed lineage-tracing tools to track senescent cells will provide further insight on the origin of these tumors as well as metastatic outgrowths. This has important clinical implications given that conven- tional chemotherapy agents, including EDP (etoposide, doxorubicin and cisplatin), frequently used as standard of care for advanced ACC76, are known to act in part by triggering ‘therapy-induced senescence’ (ref. 77). Consequently, EDP therapy may favor long-term maintenance of senescent tumor cells primed for late-stage recurrence78. Senolytic agents that selectively kill senescent cells may be a promising strategy to eliminate tumor cells following therapy-induced senescence, which is an active area of ongoing clinical investigation79 and future study.
Methods Mice
All animal procedures were approved by the Institutional Animal Care and Use Committee at the University of Michigan (protocol 00010217) and the University of Utah (protocol 21-02009). Mouse strains used in this study have been previously described: SF1Cre-high17, Znrf3-floxed13, R26RmTmG (ref. 23). All animals were maintained on the C57Bl/6J back- ground with a 12-h light-12-h dark cycle and ad libitum access to food and water. Littermate control animals were used in all experiments. Tumor-bearing mice were monitored based on body weight and con- dition. Mice that exhibited pain, loss of appetite, distress, abnormal behavior or loss of condition were evaluated by a veterinarian and euthanized accordingly.
Ultrasound
Mice were anesthetized using 3% isoflurane in O2 (1 1 min-1) in an induc- tion chamber until reaching the surgical plane. Following induction, mice were weighed and moved to a heated platform in the prone posi- tion in a modified Trendelenburg position. To maintain a surgical plane of anesthesia, 1-1.5% isoflurane in O2 was supplied via a nose cone. Eye lubricant was applied to prevent corneal damage during prolonged anesthesia. ECG and respiration were monitored via non-invasive rest- ing ECG electrodes. Hair was removed from the left dorsal area with depilatory cream. Two-dimensional (B-mode) and three-dimensional ultrasound images were recorded using a VisualSonics Vevo 2100 in vivo micro-imaging system with an MS-550D transducer, which has a center frequency of 40 MHz and a bandwidth of 22-55 MHz.
Surgical castration
Male control or Znrf3-cKO mice at 4 or 6 weeks of age were anesthetized using 3% isoflurane in O2 (1 1 min-1) in an induction chamber until reach- ing the surgical plane. To maintain a surgical plane of anesthesia, 1-1.5% isoflurane in O2 was supplied via a nose cone. Eye lubricant was applied to prevent corneal damage during prolonged anesthesia. A single inci- sion was made at each scrotum, and testes were exposed and removed. Pain relief was administered using bupivacaine (2 mg ml-1) subcutane- ously and lidocaine (4 mg per kg) at the surgical site. Sham operations followed the same protocol, without exposing and removing the testes. Mice were housed singly after the operation and monitored in accord- ance with Institutional Animal Care and Use Committee protocols. Mice were euthanized at 9 or 12 weeks, respectively, and prostate regression was visually inspected to confirm complete castration.
TD139 study
TD139 (15 mg ml-1) or DMSO was administered intraperitoneally to 6-week-old male Znrf3-cKO mice every 3 d for 6 weeks. Upon termi- nation of the study, adrenals were removed, cleaned of excess fat, embedded in 4% low-melt agarose (Lonza, 50101) and cut into 150- um tissue sections using a Leica VT1200S vibratome. Tissue slices were cleaned of excess agarose and incubated in complete medium (DMEM/F12, 1× GlutaMAX, 1× ITSX and 1% penicillin-streptomycin) at 37 ℃ with 0% CO. Phagocytosis beads (pHrodo Deep Red E. coli BioParticles; Thermo Fisher, P35360) reconstituted at 1 mg ml-1 in sterile 1× PBS were added to adrenal slices (5 µg ml-1) and incubated for 90 min with agitation. Upon assay termination, slices were washed with 1x ice-cold PBS and fixed for 15 min using 4% paraformaldehyde (Fisher Scientific, NC1537886). Slices were then stained with DAPI (Thermo Fisher, D21490) and mounted using a glycerol-based mount- ing medium. Images were acquired using a Leica SP8 laser scanning confocal microscope with a ×40,1.10 water-immersion objective (HC PL APO CS2). Volume acquisition was performed in unidirectional xyz scan mode to obtain z stacks using a slice interval of 1 um and a resolution of 0.284 um per voxel. The pinhole was set to 1 Airy unit. Serial acquisition was used to minimize cross-talk for 405-, 488-, 561- and 633-nm lasers. The number of phagocytosis beads per image was quantified using QuPath digital image-analysis software and normalized to DAPI values.
Adrenal weight
Adrenals were cleaned of excess fat, weighed and snap frozen in liquid nitrogen or fixed as described for histology. In control animals, we observed an age-dependent increase in adrenal weight across our time series that significantly correlated with an age-dependent increase in body weight (Supplementary Fig. 3a-c). To account for these normal aging differences, adrenal weights in control and Znrf3-cKO mice were compared based on the adrenal-to-body weight ratio calculated as the sum of both adrenal glands (mg) normalized to body weight (g). Raw adrenal weight data are provided for reference in Supplementary Fig. 3d-g.
Mouse adrenal histology
Adrenals were fixed in 10% normal buffered formalin (Fisher Scientific, 23-427098) for 24 h at room temperature, embedded in paraffin and cut into 5-um sections. Antibody information and staining conditions are listed in Supplementary Table 5. Endogenous peroxidase activity was blocked with 0.3% H2O2 for 30 min at room temperature. For bright- field IHC, primary antibodies were detected with HRP polymer solution (Vector Laboratories, MP-7401 (anti-rabbit), MP-7402 (anti-mouse) or MP-7405 (anti-goat)) and DAB EqV peroxidase substrate (Vector Labo- ratories, SK-4103-100), nuclei were counterstained with hematoxylin (Sigma, GHS132), and slides were mounted using VectaMount (Vector Laboratories, H-5000-60). For immunofluorescence, primary antibod- ies were detected with HRP polymer solution and Alexa Fluor tyramide reagent (Thermo Fisher, B40953 (488), B40955 (555) or B40958 (647)),
nuclei were counterstained with Hoechst (Thermo Fisher, H3570), and slides were mounted using ProLong Gold (Thermo Fisher, P36930). For multiplexed immunofluorescence, consecutive staining of pri- mary antibodies and the corresponding HRP-conjugated polymers followed by signal detection with Alexa Fluor tyramide reagents was performed using the same conditions as described for brightfield IHC. Antibody stripping was performed after detection of each primary antibody by boiling in the antigen-retrieval buffer of the next primary antibody. Single-stain and fluorescence-minus-one controls were run to confirm complete stripping. Imaging was performed on a Zeiss Apotome microscope with an AxioCam MRm camera, a Pannoramic MIDI II (3DHISTECH) digital slide scanner or a Zeiss Axio Scan.Z1 digi- tal slide scanner. Quantification was performed using QuPath digital pathology software. Multiplex immunofluorescence quantification was performed using QuPath on manually annotated fused cell clusters.
Senescence-associated ß-galactosidase staining
Freshly isolated adrenal tissue was cryopreserved in optimal cutting temperature compound. Cryosections (10 um) were stained using the Senescence ß-Galactosidase Staining Kit (Cell Signaling Technology, 9860) according to the manufacturer’s protocol with minor modifi- cations. Briefly, cryosections were air dried at room temperature for 30 min, fixed for 15 min and washed with PBS. Sections were stained inside a sealed, humidified chamber at 37 ℃ overnight. Sections were then counterstained with eosin, dehydrated and mounted for imaging.
RNA isolation
Adrenals isolated from mice at the indicated time points were cleaned of excess fat, weighed, snap frozen in liquid nitrogen and stored at -80 ℃. For homogenization, RLT lysis buffer (Qiagen, 74104) contain- ing 1% ß-mercaptoethanol was added to thawed adrenal tissue in high- impact zirconium 1.5-mm bead tubes (Benchmark Scientific, D1032-15). Samples were homogenized for 30 s using the BeadBug 3 homogenizer (Benchmark Scientific, D1030), returned to ice and homogenized for an additional 30 s. Total RNA isolation was completed using the RNeasy Kit (Qiagen, 74104) according to manufacturer guidelines. DNases were removed using the RNase-Free DNase Set (Qiagen, 79254) and subsequently purified using the RNeasy MinElute Cleanup Kit (Qiagen, 74204) according to manufacturer guidelines.
Bulk RNA sequencing
Following RNA isolation and DNase treatment, RNA quality was evalu- ated using an Agilent 2100 Bioanalyzer (Agilent Technologies). All samples used for sequencing had an RNA-integrity number value >8.0. Libraries were prepared using the NEBNext Ultra II Directional RNA Library Prep Kit with the poly(A) mRNA isolation protocol and sequenced using the NovaSeq Reagent Kit version 1.5 150 x 150-bp sequencing protocol. Libraries were sequenced on an Illumina NovaSeq 6000. The mouse GRCm38 genome and gene-feature files were down- loaded from Ensembl release 102, and a reference database was created using STAR version 2.7.6a with splice junctions optimized for 150-bp reads80. Optical duplicates were removed from the paired-end FASTQ files using clumpify version 38.34, and reads were trimmed of adap- tors using cutadapt 1.16 (ref. 81). The trimmed reads were aligned to the reference database using STAR in two-pass mode to output a BAM file sorted by coordinates. Mapped reads were assigned to annotated genes using featureCounts version 1.6.3. Expected gene counts were filtered to remove features with zero counts and features with fewer than ten reads in any sample. The age and genotype were combined into a single column in the design formula, and DEGs were identified using a 5% false discovery rate with DESeq2 version 1.30.1 (ref. 82). Male and female datasets were analyzed in separate DESeq2 models. Pathways were analyzed using the fast gene set enrichment package83 and IPA20. Z scores reported from IPA represent the activation Z score, which was used to infer the activation states of pathways and upstream regulators.
Adrenal gland dissociation
Adrenal glands were obtained from 6- or 9-week-old mice following rapid decapitation to minimize stress-induced transcriptional changes. Adrenals were placed immediately into ice-cold 1× Hank’s Balanced Salt Solution (HBSS) containing calcium and magnesium (Thermo Fisher, 14025134). Tissues were then finely chopped and digested enzymati- cally using enzymes and reagents from the Neural Tissue Dissociation Kit (Miltenyi Biotec, 130-092-628). All steps were carried out at 4 ℃, including centrifugation. All tubes and pipette tips used to handle cell suspensions were pre-coated with 3% BSA in HBSS to prevent cell loss. During the dissociation, the cell suspension was gently agitated with mechanical pipetting every 10 min and visually assessed under a stereomicroscope until the tissue was fully digested. The suspension was then filtered through 70-um filters to obtain a single-cell suspen- sion, and enzymes were neutralized using HBSS containing 10% FBS. Red blood cells were removed using Red Blood Cell Lysis buffer (Roche, 11814389001) according to manufacturer guidelines, and the cells were washed twice with HBSS containing 2% FBS before counting on a Countess Automated Cell Counter (Thermo Fisher).
Single-cell RNA sequencing
Immediately following dissociation, cells were stained with 0.1 µg ml-1 DAPI and 5 µM Vybrant DyeCycle Ruby (Invitrogen, V10309). Viable Ruby-positive, DAPI-negative single cells were sorted based on fluores- cent properties using the BD FACSAria III Cell Sorter at the University of Utah Flow Cytometry Shared Resource. Cells were collected in 1x PBS (free of calcium and magnesium) containing 0.04% BSA and kept on ice. Single-cell droplets with a target capture of 10,000 cells were immediately prepared on the 10x Chromium system at the Huntsman Cancer Institute High-Throughput Genomics Shared Resource (Sup- plementary Fig. 4 Gating Strategy). Single-cell libraries were prepared using the Chromium Next GEM Single Cell 3’ Gene Expression Library Construction Kit version 3.1 according to manufacturer instructions. Sequencing was performed on an Illumina NovaSeq 6000 instrument using an S4 flow cell generating 150-bp paired-end reads at a target depth of 300 million reads per sample. Raw data were converted to demultiplexed FASTQ files with 10x Cellranger mkfastq. Reads were mapped to the prebuilt mm10 reference distributed by 10x Genomics (version mm10-2020-A), and genes were counted with 10x Cellranger count. Samples were also aggregated into a single Loupe file with 10x Cellranger aggr. Filtering parameters for gene-count range (>200 and <5,500), unique molecular identified counts (>200 and <30,000) and mitochondrial transcript fraction (<15%) were used to remove low-quality cells.
scRNA-seq analysis and cluster identification
For analysis of scRNA-seq clusters, major cell types were distinguished using graph-based and k-mean methods using the Loupe Cell Browser Interface. For major cell clusters, core cell type-specific genes were used to validate initial clustering methods: adrenal cortex (Nr5a1, also known as Sf1), immune (Ptprc), endothelial (Plvap) and medulla (Th) (Supplementary Fig. 2). Validation of subclusters was completed by identifying defined transcriptional signatures from previously anno- tated mouse scRNA-seq datasets84-86 and established identification markers (Supplementary Fig. 2). Both intercluster and intra-cluster comparisons were performed for cells of similar lineage. Final cluster- ing was based on both gene expression patterns and the proportion of significant DEGs between respective clusters. Validation of immune- based clusters was completed using Cluster Identity Predictor corre- lation analysis to compare new clusters with the established ImmGen reference dataset41. Enrichment analysis was performed using Enrichr with which DEGs in the dominant myeloid clusters (macrophages, monocytes, DCs and MDSCs) were compared between 9 and 6 weeks. Combined scores were used to indicate statistical significance (P) and ranking (x) based on c = log (P) x x (ref. 87).
TCGA-ACC immune signatures
The relative AMRS in each tumor from TCGA-ACC data was calculated using mRNA expression Z scores obtained from cBioPortal (https:// www.cbioportal.org, PanCancer Atlas). A composite enrichment score was calculated based on the average expression of CD33, CD68, ITGAM (also known as CD11b) and MSR1 (also known as CD204). TCGA-ACC tumors without mRNA expression data were excluded (n =14), resulting in final analysis of 31 male and 47 female tumors. A score cutoff above or below the median, respectively, was used to classify AMRS-high versus AMRS-low tumors. Individual immune cell scores were derived using the same approach for macrophages (C1QA, C1QB, APOE, MERTK), DCs(CD209,ITGAE,IRF8, TLR3), neutrophils (S100A8,S100A9,MMP9, CSF3R), B cells (CD79A, CD79B,RAG1,CD22), natural killer cells (GZMA, GZMB, CCL5, NKG7), T cells (CD3E, CD3D, CCR7,CTLA4) and monocyte- derived cells (ITGAX, CD68, CCR2, ITGAM). Gene markers for each cell population were chosen based on established cell type-specific genes that were also robustly expressed across the TCGA-ACC cohort. MKI67, which is an established prognostic marker in ACC57, as well as a pub- lished gene set for phagocytic macrophages66 (CD68, TREM2, TYROBP) were included for comparison.
Human ACC tumor histology
Primary ACC FFPE tumor samples were obtained from the Huntsman Cancer Institute and the University of Colorado School of Medicine at Colorado Anschutz Medical Campus in accordance with study pro- tocols approved by the Institutional Review Board at each institu- tion. Surgical specimens from each case were reviewed by a clinical pathologist to select representative FFPE blocks enriched for viable tumor tissue. H&E and immunostaining were performed by the ARUP Research Histology Lab and the Huntsman Cancer Institute Bioreposi- tory and Molecular Pathology Shared Resource, respectively, on 5-um serial sections. Immunostaining was performed on the Leica Bond automated staining platform (Leica Biosystems). Antigen retrieval was performed with Epitope Retrieval Solution 2 (Leica Biosystems) for 20 min (CD68) or 30 min (Ki67) at 95 ℃, followed by primary anti- body incubation (anti-CD68, Leica, PA0273, ready to use; anti-Ki67, Abcam, ab1667, 1:400) for 15 min at room temperature and the Bond Polymer Refine Detection kit (Leica Biosystems) for 8 min at room temperature. Slides were imaged on a Zeiss Axio Scan.Z1 slide scanner, and automated quantification was performed using QuPath digital pathology software. Surrounding fat and stroma as well as necrotic tumor regions were excluded from analysis. Tumors with substantial intratumor heterogeneity with respect to the Ki67 proliferation index were scored based on region.
Statistics and reproducibility
Sample size was determined based on power calculations performed using G*Power to obtain >80% power with 5% type 1 error. Statistical analysis was performed using R or GraphPad Prism 9. For comparison of two groups, a two-tailed Student’s t-test was performed. For more than two groups, one-way ANOVA followed by Tukey’s post hoc test was per- formed. When analyzing two independent variables, two-way ANOVA followed by Tukey’s post hoc test was used. Normal distribution was tested with an F-test (two groups) or the Brown-Forsythe test (ANOVA). If normal distribution could not be assumed, a non-parametric analysis was used or data were log2 transformed before statistical analysis. Spe- cific statistical tests used for each analysis are indicated in the figure legends. A Pvalue <0.05 was considered significant, and actual Pvalues for each test are indicated in each figure. Samples were blinded before analysis whenever possible. This included the pathological review of tumors performed on samples from 78-week-old animals. In many experiments, blinding was not feasible, as the genotype was evident from the image data based on differences in adrenal size. Results from the adrenal ultrasound were not replicated. Rather, we reproduced these finding using an expanded orthogonal approach based on adrenal
weight with increased numbers and in both sexes. Results from bulk RNA-seq and scRNA-seq were confirmed using orthogonal approaches. All other laboratory experiments were replicated successfully, and no data were excluded. For further information on experimental numbers for in vivo studies, see the Supplementary Information. For TCGA-ACC analysis using cBioPortal, patients without mRNA expression data were excluded. For RNA-seq experiments, mice were randomly chosen from the respective cohorts. For the TD139 drug study, mice were randomly assigned vehicle control (DMSO) or drug treatment. For castration studies, male mice were randomly assigned castration or sham surgical procedure. For all other experiments, mice were assigned to control or Znrf3-cKO groups based on genotype.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
All sequencing datasets have been deposited to the NCBI Gene Omnibus Database (accession code GSE201127 (GSE201125, bulk RNA-seq; GSE201126, scRNA-seq). Patient data analyzed from TCGA- ACC are publicly available using https://www.cbioportal.org/study/ summary?id=acc_tcga_pan_can_atlas_2018. All other data supporting the findings of this study are available from the corresponding author upon reasonable request.
Code availability
Code used for bulk RNA-seq and scRNA-seq analyses is freely available at https://github.com/HuntsmanCancerInstitute. Additional scRNA- seq packages used for analysis are available at https://github.com/ atakanekiz/CIPR-Package.
References
1. National Cancer Institute. Age and Cancer Risk https://www. cancer.gov/about-cancer/causes-prevention/risk/age (2021).
2. Siegel, R. L., Miller, K. D. & Jemal, A. Cancer statistics, 2018. CA Cancer J. Clin. 68, 7-30 (2018).
3. Rozhok, A. & DeGregori, J. A generalized theory of age-dependent carcinogenesis. eLife 8, e39950 (2019).
4. Laconi, E., Marongiu, F. & DeGregori, J. Cancer as a disease of old age: changing mutational and microenvironmental landscapes. Br. J. Cancer 122, 943-952 (2020).
5. Phillip, J. M., Aifuwa, I., Walston, J. & Wirtz, D. The mechanobiology of aging. Annu. Rev. Biomed. Eng. 17, 113-141 (2015).
6. Crona, J. & Beuschlein, F. Adrenocortical carcinoma-towards genomics guided clinical care. Nat. Rev. Endocrinol. 15, 548-560 (2019).
7. Else, T. et al. Adrenocortical carcinoma. Endocr. Rev. 35, 282-326 (2014).
8. Flurkey, K., Mcurrer, J. & Harrison, D. Mouse models in aging research. In The Mouse in Biomedical Research Vol. 3, 637-672 (Elsevier, 2007).
9. Nusse, R. & Clevers, H. Wnt/B-catenin signaling, disease, and emerging therapeutic modalities. Cell 169, 985-999 (2017).
10. Assie, G. et al. Integrated genomic characterization of adrenocortical carcinoma. Nat. Genet. 46, 607-612 (2014).
11. Zheng, S. et al. Comprehensive pan-genomic characterization of adrenocortical carcinoma. Cancer Cell 29, 723-736 (2016).
12. Hao, H. X. et al. ZNRF3 promotes Wnt receptor turnover in an R-spondin-sensitive manner. Nature 485, 195-200 (2012).
13. Koo, B. K. et al. Tumour suppressor RNF43 is a stem-cell E3 ligase that induces endocytosis of Wnt receptors. Nature 488, 665-669 (2012).
14. Berthon, A. et al. Constitutive ß-catenin activation induces adrenal hyperplasia and promotes adrenal cancer development. Hum. Mol. Genet. 19, 1561-1576 (2010).
15. Heaton, J. H. et al. Progression to adrenocortical tumorigenesis in mice and humans through insulin-like growth factor 2 and ß-catenin. Am. J. Pathol. 181, 1017-1033 (2012).
16. Borges, K. S. et al. Wnt/B-catenin activation cooperates with loss of p53 to cause adrenocortical carcinoma in mice. Oncogene 39, 5282-5291 (2020).
17. Bingham, N. C., Verma-Kurvari, S., Parada, L. F. & Parker, K. L. Development of a steroidogenic factor 1/Cre transgenic mouse line. Genesis 44, 419-424 (2006).
18. Basham, K. J. et al. A ZNRF3-dependent Wnt/B-catenin signaling gradient is required for adrenal homeostasis. Genes Dev. 33, 209-220 (2019).
19. Mitani, F., Mukai, K., Miyamoto, H., Suematsu, M. & Ishimura, Y. Development of functional zonation in the rat adrenal cortex. Endocrinology 140, 3342-3353 (1999).
20. ). Krämer, A., Green, J., Pollard, J. & Tugendreich, S. Causal analysis approaches in Ingenuity Pathway Analysis. Bioinformatics 30, 523-530 (2014).
21. Panier, S. & Boulton, S. J. Double-strand break repair: 53BP1 comes into focus. Nat. Rev. Mol. Cell Biol. 15, 7-18 (2014).
22. Gorgoulis, V. et al. Cellular senescence: defining a path forward. Cell 179, 813-827 (2019).
23. Muzumdar, M. D., Tasic, B., Miyamichi, K., Li, L. & Luo, L. A global double-fluorescent Cre reporter mouse. Genesis 45, 593-605 (2007).
24. Coppé, J .- P. et al. Senescence-associated secretory phenotypes reveal cell-nonautonomous functions of oncogenic RAS and the p53 tumor suppressor. PLOS Biol. 6, e301 (2008).
25. Basisty, N. et al. A proteomic atlas of senescence-associated secretomes for aging biomarker development. PLOS Biol. 18, e3000599 (2020).
26. Wiley, C. D. et al. Analysis of individual cells identifies cell-to-cell variability following induction of cellular senescence. Aging Cell 16, 1043-1050 (2017).
27. Kim, S .- J. et al. Endothelial Toll-like receptor 4 maintains lung integrity via epigenetic suppression of p16INK4a. Aging Cell 18, e12914 (2019).
28. Limbad, C. et al. Astrocyte senescence promotes glutamate toxicity in cortical neurons. PLOS ONE 15, e0227887 (2020).
29. Jochems, F. et al. The Cancer SENESCopedia: a delineation of cancer cell senescence. Cell Rep. 36, 109441 (2021).
30. Schiraldi, M. et al. HMGB1 promotes recruitment of inflammatory cells to damaged tissues by forming a complex with CXCL12 and signaling via CXCR4. J. Exp. Med. 209, 551-563 (2012).
31. Lotze, M. T. & Tracey, K. J. High-mobility group box 1 protein (HMGB1): nuclear weapon in the immune arsenal. Nat. Rev. Immunol. 5, 331-342 (2005).
32. Lau, L., Porciuncula, A., Yu, A., Iwakura, Y. & David, G. Uncoupling the senescence-associated secretory phenotype from cell cycle exit via interleukin-1 inactivation unveils its protumorigenic role. Mol. Cell. Biol. 39, e00586-18 (2019).
33. Galanos, P. et al. Mutational signatures reveal the role of RAD52 in p53-independent p21-driven genomic instability. Genome Biol. 19, 37 (2018).
34. Li, Y. et al. Senescent mesenchymal stem cells promote colorectal cancer cells growth via galectin-3 expression. Cell Biosci. 5, 21 (2015).
35. Mosteiro, L., Pantoja, C., de Martino, A. & Serrano, M. Senescence promotes in vivo reprogramming through p16INK4ª and IL-6. Aging Cell 17, e12711 (2018).
36. Coppé, J .- P., Desprez, P .- Y., Krtolica, A. & Campisi, J. The senescence-associated secretory phenotype: the dark side of tumor suppression. Annu. Rev. Pathol. 5, 99-118 (2010).
37. Kale, A., Sharma, A., Stolzing, A., Desprez, P .- Y. & Campisi, J. Role of immune cells in the removal of deleterious senescent cells. Immun. Ageing 17, 16 (2020).
38. Naeim, F. Histiocytic and dendritic cell disorders. In Hematopathology (eds Naeim, F., Nagesh Rao, P. & Grody, W. W.) 489-512 (Elsevier, 2008).
39. Picarsic, J. L. & Chikwava, K. Disorders of histiocytes. In Hematopathology 3rd edn (ed. Hsi, E. D.) 567-616 (Elsevier, 2018).
40. Dale, D. C., Boxer, L. & Liles, W. C. The phagocytes: neutrophils and monocytes. Blood 112, 935-945 (2008).
41. Ekiz, H. A., Conley, C. J., Stephens, W. Z. & O’Connell, R. M. CIPR: a web-based R/shiny app and R package to annotate cell clusters in single cell RNA sequencing experiments. BMC Bioinformatics 21, 191 (2020).
42. Evrard, M. et al. Developmental analysis of bone marrow neutrophils reveals populations specialized in expansion, trafficking, and effector functions. Immunity 48, 364-379 (2018).
43. Volberding, P. J. et al. Suppressive neutrophils require PIM1 for metabolic fitness and survival during chronic viral infection. Cell Rep. 35, 109160 (2021).
44. Kuleshov, M. V. et al. Enrichr: a comprehensive gene set enrichment analysis web server 2016 update. Nucleic Acids Res. 44, W90-W97 (2016).
45. Bassler, K., Schulte-Schrepping, J., Warnat-Herresthal, S., Aschenbrenner, A. C. & Schultze, J. L. The myeloid cell compartment-cell by cell. Annu. Rev. Immunol. 37, 269-293 (2019).
46. Guilliams, M. et al. Unsupervised high-dimensional analysis aligns dendritic cells across tissues and species. Immunity 45, 669-684 (2016).
47. Wang, H. et al. Role of bone marrow-derived CD11c+ dendritic cells in systolic overload-induced left ventricular inflammation, fibrosis and hypertrophy. Basic Res. Cardiol. 112, 25 (2017).
48. Song, P. et al. Hepatic recruitment of CD11b+Ly6C+ inflammatory monocytes promotes hepatic ischemia/reperfusion injury. Int. J. Mol. Med. 41, 935-945 (2018).
49. Dolfi, B. et al. Unravelling the sex-specific diversity and functions of adrenal gland macrophages. Cell Rep. 39, 110949 (2022).
50. Sano, H. et al. Critical role of galectin-3 in phagocytosis by macrophages. J. Clin. Invest. 112, 389-397 (2003).
51. Hirani, N. et al. Target inhibition of galectin-3 by inhaled TD139 in patients with idiopathic pulmonary fibrosis. Eur. Respir. J. 57, 2002559 (2021).
52. Lindner, B., Burkard, T. & Schuler, M. Phagocytosis assays with different pH-sensitive fluorescent particles and various readouts. Bio Techniques 68, 245-250 (2020).
53. Grabek, A. et al. The adult adrenal cortex undergoes rapid tissue renewal in a sex-specific manner. Cell Stem Cell 25, 290-296 (2019).
54. Weiss, L. M. Comparative histologic study of 43 metastasizing and nonmetastasizing adrenocortical tumors. Am. J. Surg. Pathol. 8, 163-169 (1984).
55. Weiss, L. M., Medeiros, L. J. & Vickery, A. L. Pathologic features of prognostic significance in adrenocortical carcinoma. Am. J. Surg. Pathol. 13, 202-206 (1989).
56. Wu, C. et al. Myeloid signature reveals immune contexture and predicts the prognosis of hepatocellular carcinoma. J. Clin. Invest. 130, 4679-4693 (2020).
57. Beuschlein, F. et al. Major prognostic role of Ki67 in localized adrenocortical carcinoma after complete resection. J. Clin. Endocrinol. Metab. 100, 841-849 (2015).
58. Mohan, D. R. et al. Targeted assessment of GOS2 methylation identifies a rapidly recurrent, routinely fatal molecular subtype
of adrenocortical carcinoma. Clin. Cancer Res. 25, 3276-3288 (2019).
59. Thorsson, V. et al. The immune landscape of cancer. Immunity 48, 812-830 (2018).
60. Landwehr, L .- S. et al. Interplay between glucocorticoids and tumor-infiltrating lymphocytes on the prognosis of adrenocortical carcinoma. J. Immunother. Cancer 8, e000469 (2020).
61. Hägg, S. & Jylhävä, J. Sex differences in biological aging with a focus on human studies. eLife 10, e63425 (2021).
62. Clocchiatti, A., Cora, E., Zhang, Y. & Dotto, G. P. Sexual dimorphism in cancer. Nat. Rev. Cancer 16, 330-339 (2016).
63. Guan, X. et al. Androgen receptor activity in T cells limits checkpoint blockade efficacy. Nature 606, 791-796 (2022).
64. Roediger, J. et al. Supraphysiological androgen levels induce cellular senescence in human prostate cancer cells through the Src-Akt pathway. Mol. Cancer 13, 214 (2014).
65. Mirzakhani, K. et al. The androgen receptor-IncRNASAT1-AKT-p15 axis mediates androgen-induced cellular senescence in prostate cancer cells. Oncogene 41, 943-959 (2022).
66. Wilmouth, J. J. et al. Sexually dimorphic activation of innate antitumor immunity prevents adrenocortical carcinoma development. Sci. Adv. 8, eadd0422 (2022).
67. Chen, H .- A. et al. Senescence rewires microenvironment sensing to facilitate antitumor immunity. Cancer Discov. 13, 432-453 (2023).
68. Marin, I. et al. Cellular senescence is immunogenic and promotes antitumor immunity. Cancer Discov. 13, 410-431 (2023).
69. Jimenez, C. et al. Endocrine and neuroendocrine tumors special issue-checkpoint inhibitors for adrenocortical carcinoma and metastatic pheochromocytoma and paraganglioma: do they work? Cancers 14, 467 (2022).
70. Ayers, M. et al. IFN-y-related mRNA profile predicts clinical response to PD-1 blockade. J. Clin. Invest. 127, 2930-2940 (2017).
71. Prat, A. et al. Immune-related gene expression profiling after PD-1 blockade in non-small cell lung carcinoma, head and neck squamous cell carcinoma, and melanoma. Cancer Res. 77, 3540-3550 (2017).
72. Riaz, N. et al. Tumor and microenvironment evolution during immunotherapy with nivolumab. Cell 171, 934-949 (2017).
73. Campisi, J. Cellular senescence as a tumor-suppressor mechanism. Trends Cell Biol. 11, S27-S31 (2001).
74. Roberson, R. S., Kussick, S. J., Vallieres, E., Chen, S .- Y. J. & Wu, D. Y. Escape from therapy-induced accelerated cellular senescence in p53-null lung cancer cells and in human lung cancers. Cancer Res. 65, 2795-2803 (2005).
75. Milanovic, M. et al. Senescence-associated reprogramming promotes cancer stemness. Nature 553, 96-100 (2018).
6. Fassnacht, M. et al. Combination chemotherapy in advanced adrenocortical carcinoma. N. Engl. J. Med. 366, 2189-2197 (2012).
77. Prasanna, P. G. et al. Therapy-induced senescence: opportunities to improve anti-cancer therapy. J. Natl Cancer Inst. 113, 1285-1298 (2021).
78. Saleh, T. et al. Therapy-induced senescence: an ‘old’ friend becomes the enemy. Cancers 12, 822 (2020).
79. Myrianthopoulos, V. et al. Senescence and senotherapeutics: a new field in cancer therapy. Pharmacol. Ther. 193, 31-49 (2019).
30. Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15-21 (2013).
81. Martin, M. Cutadapt removes adapter sequences from high- throughput sequencing reads. EMBnet j. 17, 10-12 (2011).
82. Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550 (2014).
83. Korotkevich, G. et al. Fast gene set enrichment analysis. Preprint at bioRxiv https://doi.org/10.1101/060012 (2016).
84. ImmGen Consortium et al. The neutrotime transcriptional signature defines a single continuum of neutrophils across biological compartments. Nat. Commun. 12, 2856 (2021).
85. Han, X. et al. Construction of a human cell landscape at single- cell level. Nature 581, 303-309 (2020).
86. Zilionis, R. et al. Single-cell transcriptomics of human and mouse lung cancers reveals conserved myeloid populations across individuals and species. Immunity 50, 1317-1334 (2019).
87. Chen, E. Y. et al. Enrichr: interactive and collaborative HTML5 gene list enrichment analysis tool. BMC Bioinformatics 14, 128 (2013).
Acknowledgements
We thank H. Clevers and B .- K. Koo for providing Znrf3-floxed mice and the late K. Parker for providing SF1Cre transgenic mice. Research reported in this publication used the High-Throughput Genomics and Bioinformatic Analysis Shared Resource and the Biorepository and Molecular Pathology Shared Resource at the Huntsman Cancer Institute at the University of Utah and was supported by the National Cancer Institute of the National Institutes of Health under award P30CA042014. We also used the University of Utah Flow Cytometry Shared Resource supported by the Office of the Director of the NIH (award S10OD026959 and NCI award 5P30CA042014-24) and the Cell Imaging Core at the University of Utah. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. We especially thank J. Marvin, O. Allen and M. Bridge for respective technical assistance with flow cytometry, 10x Genomics library preparation and confocal imaging. We also thank J. Gertz, B. Myers and S. Holmen for helpful scientific discussions and comments on the manuscript. This work was supported by funding from a Cancer Center Support Grant (P30CA040214, K.J.B.), the V Foundation (V2021-021, K.J.B.) and 5 For The Fight (K.J.B.).
Author contributions
Conceptualization, K.M.W., G.D.H., K.J.B .; experimentation, K.M.W., L.J.S., L.L., P.W.W., J.L.A., G.C .- H., S.O.I., C.Z., K.D.J., K.C .- B., K.J.B .; methodology, L.J.S .; data analysis, K.M.W., L.J.S., J.L.A., G.C .- H., S.O.I., K.C .- B., K.J.B .; bioinformatic analysis, K.M.W., C.J.S., B.K.L.,
H.A.E .; data curation, K.M.W., C.J.S., K.K .- V., K.J.B .; resources (patient tissue), M.B., M.R.C., K.K .- V .; pathologic review, M.B., M.R.C., T.J.G .; funding acquisition, G.D.H., K.J.B .; project administration, K.J.B .; data visualization, K.M.W., K.J.B .; writing (original draft), K.M.W., K.J.B .; writing (review and editing), all authors.
Competing interests
The authors declare no competing interests.
Additional information
Extended data is available for this paper at https://doi.org/10.1038/ s43587-023-00420-2.
Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s43587- 023-00420-2.
Correspondence and requests for materials should be addressed to Kaitlin J. Basham.
Peer review information Nature Aging thanks Curtis Henry and Ashani Weeraratna for their contribution to the peer review of this work.
Reprints and permissions information is available at www.nature.com/reprints.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
@ The Author(s), under exclusive licence to Springer Nature America, Inc. 2023
a
Adrenal Area
Adrenal Volume
Gross Histology
Example #1
8.22mm2
19.08mm3
0.0199g
Example #2
17.65mm2
64.23mm3
0.0741g
Example #3
29.94mm2
142.29mm3
0.196g
b
2.0
Adrenal Area (mm2)/ Body Weight (g)
1.5
1.0
R2=0.90
** P=1.08x10-3
0.5
0.0
0
2
4
6
8
10
Adrenal Weight (mg)/Body Weight (g)
c
8
Adrenal Volume (mm3)/ Body Weight (g)
6
4
R2=0.89
2
** P=1.35x10-3
0
0
2
4
6
8
10
Adrenal Weight (mg)/Body Weight (g)
9-Weeks
P=0.32
P=0.31
:
6-Weeks
9-Weeks
Znrf3 cKO
a
Female Mice: Cleaved Caspase 3 (CC3)
b
Male Mice: Cleaved Caspase 3 (CC3)
4-Weeks
6-Weeks
9-Weeks
4-Weeks
6-Weeks
Control
Control
Znrf3 cKO
Znrf3 cKO
0.6
P=0.26
P=0.99
P=0.99
0.6
P=0.99
% CC3 Positive
% CC3 Positive
0.4
0.4
0.2
0.2
0
0
4-Weeks Control
6-Weeks
9-Weeks
4-Weeks Control
Znrf3 cKO
c
Male Mice
4-Weeks
6-Weeks
*
6
*** P=1.36x104
** P=3.79x103
Control
% 53BP1 Foci
*
4
*
53BP1 *
53BP1
2
*
*
*
*
*
*
*
Znrf3 cKO
*
*
*
*
*
0
*
4-Weeks
6-Weeks
*
*
*
*
53BP1
53BP1
Control
Znrf3 cKO
d
Male Mice
6-Weeks
9-Weeks
*P=4.50x102
40
P=0.99
** P=5.49x109
Control
p21+ Cells/HPF
30
20
p21
p21
10
Znrf3 cKO
0
6-Weeks
9-Weeks
p21
p21
Control
Znrf3 cKO
e
Male Mice
6-Weeks
9-Weeks
60
*** P=1.51x104
Control
P=0.86
*** P=6.42x104
p16+ Cells/HPF
40
p16
p16
20
Znrf3 cKO
0
6-Weeks
9-Weeks
p16
p16
Control
Znrf3 cKO
f
o Control (9-Weeks)
d’ Znrf3 cKO (9-Weeks)
SA-ß-gal
SA-ß-gal
Extended Data Fig. 2 | Znrf3 cKO adrenals activate cellular senescence during adrenal regression. (a, b) Apoptotic cell death as measured by cleaved caspase 3 (CC3) is not significantly increased in female or male Znrf3 cKO adrenals compared to controls during the initial phase of tissue regression. Quantification of CC3-positive cells was performed using QuPath digital image analysis based on the number of positive cells and normalized to total adrenal cortex nuclei. Arrows indicate CC3-positive cells. Dashed line indicates histological boundary between adrenal cortex and medulla. Scale bars, 100 um. (c) DNA damage as measured by 53BP1 foci is significantly increased in male Znrf3 cKO adrenals compared to controls at 4- and 6-weeks of age. Quantification of 53BP1 foci was performed
using QuPath digital image analysis based on the number of positive foci and normalized to total nuclei. Asterisks indicate 53BP1-positive foci. Scale bars, 10 um. (d) p21, (e) p16INK4a, and (f) senescence-associated beta-galactosidase (SA- B-gal) are significantly increased in 9-week male Znrf3 cKO adrenals compared to controls. Scale bars, 100 um. Quantification of p21 and p16INK4ª IHC was performed using QuPath digital image analysis based on the number of positive cells per high powered field (HPF). Representative SA-ß-gal images obtained from analysis of 3 independent mice is shown. Each dot represents an individual animal. Error bars represent mean ± s.e.m. Statistical analysis was performed using two-way ANOVA followed by Tukey’s multiple comparison’s test.
a
Cytokine Analysis: 9 9 vs 6-Week Znrf3 cKO
inhibited
activated
CSF2
IL 10
TNFSF11-
TNF
OSM.
P-value
1e-13
IL6
1e-09
1e-05
IFNB1.
TNFSF13
IFNG
IL1B
-6
-4
-2
1.5
2.0
2.5
3.0
Activation Z-Score
b Cytokine Analysis: 9 12 vs 6-Week Znrf3 cKO
inhibited
activated
CSF2
IL 10
IL2
IL13
IL5
MIF
CXCL 12
C5
P-value
IL33
1e-18
IL22
1e-13
IL 1B
1e-08
IL6
1e-03
TNFSF12
IL1A
IL4
IFNG
TNF
OSM
-4
-3
-2
1.0
1.5
2.0
2.5
3.0
Activation Z-Score
c
Growth Factors: 9 9 vs 6-Week Znrf3 cKO
inhibited
activated
HGF
AREG
P-value
1e-14
EGF
1e-10
1e-06
1e-02
IGF1
TGFB3
-5
-4
-3
-2
1.56
1.60
1.64
1.68
1.72
Activation Z-Score
d
Growth Factors: 9 12 vs 6-Week Znrf3 cKO
inhibited
activated
AREG
HGF
NRG1
TGFA
KITLG
P-value
GRP
1e-10
1e-07
PDGFB
1e-04
FGF10
MST1
GDNF
BDNF
-3.0
-2.5
-2.0
-1.5
-1.0
1.2
1.6
2.0
Activation Z-Score
a
2
9-Week Control
d
Hyperplasia: 6-Week Female vs 4-Week Male
Female
Male
2
Female
Male
1
1
Cxadr
C3
0
Akric18
Sbk1
Mr1
Pik3c2g
0
Ggt6
Cyp212
Akrid1
Tmem178b
Mgatác
-1
Lama4
-1
Acadsb
Sic24a3
Tmem178b
Heca
Sbk1
Ube3a
Lpp
-2
400
-2
Pde7a
Sic25a51
Scube1
Deptor
Lrm1
Ak3
Cdh13
Ggt6
Fmo1
Sic26a2
Gas6
Bnip3
Deptor
Chot6l
Nrep
Sic25a16
Tcp1112
Gm7463
Ankrd35
Akrid1
Bloc1s5
Tct24 12124
Dopia
Mania
Susd3
Prr16
Tent5b
At2
Stard10
Map6d1
Tmem52
Acadsb
Als2cl
Fam210a
D630003M21Rik Ncr2
Tmem135
Psd3
Tspan11
Prir
Hira3
C3
Cdh22
Pon3
Var
Mgst1 Cd47 Ssbp4
Ddah2
Evaib
Baiap2
Inafm1 Susd3
Npr2
Grb14
Stard10
Rbpms2
Tle5
9530077C05Rik
Níatc4
Gmpr
Kcnmb4
Tead2
Gapdh
Cables1
Reep2
St3gal5
Gm1673
Rogdi
Sic44a2
Cs12ra
And2
Reep2
Vegfb
Alkbh6
Whip
Rbpms2
Etnk2
Tpm2
Tub
b
6-Week Znrf3 cKO
e
Early Regression: 9-Week Female vs 6-Week Male
2
Female
Male
2
Female
Male
1
Lrrn1
1
Oxtr
Stard10
2010007H06Rik
0
Sertad4
Sbk1
0
Susd3
Cdh2
Als2cl
C3
Npr2
-1
Mr1
D630003M21Rik
Pde7a
-1
Rbpms2
Tmem178b
Tpm2
Ggt6
Smagp
-2
Acadsb
And2
Scube1
-2
Tspan11
Mgatác
Cables1
Fmo1
Tcf7l1
Rims4
Sic44a2
Baiap2
Snph
Konk1
Adcy6
Eva1b
Esam
Grb14
Níatc4
Lama5
Egil7
Dgkz
Evaib
Egil7
Balap2
Reep2
Ssbp4
Níatc4
Reep2
Tle2
Tle2
Ushbp1
Ushbp1
Ddah2
Ddah2
Npr2
Gmpr
Tspan11
Uch/1 Syne4 Nor2
2010007H06Rik
D630003M21Rik
Arhgap27
Mgatác
Uchit
Fmo1
Int2
C3
Stard10
Scube1
Trim47
Ankrd35
Als2cl
Nat8f1
Rbpms2
Tmem135
Susd3 Ampd3
Acadsb
AK3
And2
Ggt6
Impf
Pde7a
Procr
Tmem178b
Esam
Cdh2
Etnik2
Sbk1
Sic44a2
Pipp3
Cables1
Lim1
Smagp
Oxtr
Pri
Mr1
Eno2
Cdh13
Sertad4
c
9-Week Znrf3 cKO
f
Late Regression: 12-Week Female vs 9-Week Male
2
Female
Male
2
Female
Male
1
Sbk1
1
Oxtr
Tmem135
Trp53inp2
Bmpria
0
Scube1
Tnfrsf19
Pde7a
0
Mitrex
Mirex
Tmem178b
Sntb1
Cdv3
Sic39a3
Acadsb
Mr1
-1
Sic25a16
Mgatác
-1
Min2
Fmot
Pipp3
-2
Gm7463
Ppp1cb
Ggt6
2
Sntb1
Gm50321
Fbx03
Nat8f1
Gmfb
Cdh13
Lor4
AK3
Pde7a
Tmem178b
Trp53inp2 Acadsb
Tmem135
Sic25a16 Ak3
Pppicb
Sic3983
Plpp3
Sec24d
Thirst19
Sbk1 Oxtr
Lernt
Cdh2
Gm7463
Mrt
Fscn1
G916
Susd3
Mgatác
St6galnac2
Lrm1
D630003M21Rik Procr
Cdh2
Fmo1
Sterd10
Scube1
Baiap2
Gphn
Sic44a2
Nat8f1
St6galnac2
Cdh13
Npr2
Reep2
Dgkz
Ddah2
Syne4
Kcnmb4
Dock6
Reep2
Nor2
Rhbdf1
Als2cl
Als2cl
Stard10
Rbpms2
D630003M21Rik
Ssbp4
Fes
Susd3
Tle2
Ssbp4
Sic22a17
Ushbp1
Hdgli2
Egil7
Ubxn11
Ubxn11
Egil7
Etnik2
Baiap2
Tpm2
a
Canonical Pathway Analysis:
9 12- vs 6-Week Znrf3 cKO
b
Cytokines: o 9- vs 4-Week Znrf3 cKO
inhibited
activated
Kinetochore Metaphase Signaling Pathway
inhibited
activated
WNT5A
P-value
Cyclins and Cell Cycle Regulation -
IL1RN
CSF1
☒ 1e-29
Estrogen-mediated S-phase Entry
EDN1
☒ 1e-21
Mitotic Roles of Polo-Like Kinase
IL4
☒ 1e-13
CD40LG
1e-05
Coagulation System -
IL15
Salvage Pathways of Pyrimidine Ribonucleotides
IL13
IL6
Pyridoxal 5’-phosphate Salvage Pathway
P-value
1e-12
TNFSF12
Production of Nitric Oxide and
IL18
Reactive Oxygen Species in Macrophages
1e-08
IL2
Role of CHK Proteins in Cell Cycle Checkpoint Control -
1e-04
IL1A
IL33
Nicotine Degradation II -
IL1B
LXR/RXR Activation -
TNF
IFNG
Cell Cycle: G2/M DNA Damage Checkpoint Regulation -
3.0
2.9
-2.8
-2.7
-2.6
3
4
5
6
7
Breast Cancer Regulation by Stathmin1-
Activation Z-Score
Phagosome Formation
-3.5-3.0-2.5-2.0-1.5
2.0 2.5 3.0 3.5 4.0
Activation Z-Score
GSEA: /9- vs 4-Week Znrf3 cKO
c
Il6 Jak Stat3 Signaling
Chemokine Signaling Pathway
0.6-
0.5.
Enrichment Score
NES = 2.
FDR = 0.1
P = 0.0006
Enrichment Score
0.4.
NES =1.9
FDR = 0.1
0.4.
P = 0.001
0.3.
0.2.
0.2.
0.1.
0.0-
0.0-
0
5000
10000
15000
0
5000
rank
10000
15000
rank
0.6.
Inflammatory Response
Innate Immune System
Enrichment Score
NES 2.4
FDR = 0.1
0.4.
NES =1.9
P = 0.0007
Enrichment Score
FDR = 0.1
0.4.
0.3.
P = 0.006
0.2.
0.2.
0.1.
0.0.
0.01.
0
5000
10000
15000
0
5000
10000
rank
15000
rank
p< 0.05. (c) GSEA for inflammatory signatures (IL6/JAK/STAT3 signaling, chemokine signaling, the inflammatory response, and the innate immune system) identifies positively enriched genes in 9- vs 4-week male Znrf3 cKO mice. Data is representative of 4 biological replicates per group and, statistical analysis in IPA was performed using a right-tailed fishers exact test, p < 0.05.
a
9-week-old Znrf3 cKO Mice
b
9-week-old Znrf3 cKO Mice
0-
P=0.09
Ly6G
2
**** P= 4.05x10-5
CD3e
(Log2 Normalized)
Sham
Castrated
% CD3e
(Log2 Normalized)
Sham
Sham
Sham
Castrated
1.
Castrated
% Ly6g
-2.
Castrated
0
-4.
-1
-6
-2
Sham Castrated
Sham Castrated
c
d
o
Castrated
e
Study Timeline: o Control or Znrf3 cKO Mice
Sham
Control
Normalized Adrenal Weight (mg)/ Body Weight (g)
12-Weeks
2.0
*** P=4.85x10-4
Castration or Sham control procedure
1.5
Sham
Castrated
Harvest adrenals
Znrf3 cKO
1.0
P0
*** P=8.50x10-4
6-Weeks
12-Weeks
0.5
0.0
Control
Znrf3 cKO
Fold Change:
1.48
2.27
f
Senescence Markers: 12-week-old Znrf3 cKO mice
g
Immune Markers: 12-week-old Znrf3 cKO mice
Ki67
** P=3.52x10-3
*** P=2.93x10-4
1.5
P=0.55
Sham
8
*P=1.95x10-2
6
5
7
% CD68+ (Log2 Normalized)
*P=3.82x10-2
1.0
4
Sham
6
% Ki67+
(Log2 Normalized)
Castrated
% CD11c+
(Log2 Normalized)
(Log2 Normalized)
(Log2 Normalized)
6
4.
% CD11b+
3
0.5
% F4/80+
4-
5
0.0
2
2
2-
4
-0.5
Castrated
1
-1.0
0
0
0
3
Sham Castrated
Sham Castrated
Sham Castrated
Sham Castrated
Sham Castrated
p16
Sham
Castrated
Sham
Castrated
P=0.31
4
CD68
CD11c
CD11b
F4/80
% p16+
(Log2 Normalized)
3
Sham
2.
Sham
1
Castrated
0
Sham Castrated
Castrated
p21
2.0
P=0.39
1.5
% p21+
(Log2 Normalized)
1.0-
Sham
0.5-
0.0-
-0.5
Castrated
-1.0
Sham Castrated
weight. Relative fold change is indicated below each group. (f) Senescence markers (Ki67, p16, and p21) and (g) myeloid immune markers (CD68, CD11c, CD11b, and F4/80) in adrenals from sham versus castrated animals. Representative images are shown for each group. Scale bars 20 uM. Quantification was performed using QuPath digital image analysis based on the number of positive cells normalized to total nuclei. Each dot represents an individual animal. Box and whisker plots indicate the median (line) within the upper (75%) and lower (25%) quartiles, and whiskers represent the range. Statistical analysis was performed using two-tailed Student’s t-test.
a
Female Mice
b
4wk
6wk
9wk
12wk
24wk
44wk
52wk
44-Weeks
52-Weeks
100
Histiocyte
Proportion of Mice (%)
Area
Control
80
one
2%
60
-10%
-
-
40
0-20%
Female
0-50%
20
>50%
0
Control
Znrf3 cKO
Control Znrf3 cKO
Control
Znrf3 cKO
Control
Znrf3 cKO
Control
Znrf3 cKO
Control
Znrf3 cKO
Control
Znrf3 cKO
Znrf3 cKO
-
-
c
Male Mice
d
4wk
6wk
9wk
12wk
24wk
44wk
52wk
44-Weeks
52-Weeks
100
Proportion of Mice (%)
Histiocyte Area
Control
80
one
2%
60
-10%
0-20%
-
-
40
20-50%
Male
20
>50%
0
Control Znrf3 cKO
Control Znrf3 cKO
Control
Znrf3 cKO
Control
Znrf3 cKO
Control
Znrf3 cKO
Control
Znrf3 cKO
Control
Znrf3 cKO
Znrf3 cKO
-
-
e
Female Mice
f
44-Weeks
52-Weeks
8
Control
6
*P=1.87x10-2
*P=1.26x10-2
Log2 (%CD68+)
*P=4.18x102
** P=1.83x10-3
4
P=0.95
P=0.23
-
-
2
Female
0
0
Znrf3 cKO
-2
6wk
9wk
12wk
24wk
44wk
52wk
Control
Znrf3 cKO
-
-
g
Male Mice
h
44-Weeks
52-Weeks
8
*** P=1.67x104
P=0.067
P=0.47
** P=5.04x10-3
Control
6
** P=1.20x103
Log2 (%CD68+)
P=0.95
9
4
-
-
2
Male
0
Znrf3 cKO
-2
6wk
9wk
12wk
24wk
44wk
52wk
Control
Znrf3 cKO
-
-
i
78-Weeks: Atrophic Glands
j
o
o
78-Weeks: Benign Tumors
Ki67 Proliferation Index (Log2 Transformed)
2
** P=4.91x10-3
8
*** P=9.65x10-4
CD68 Index
(Log2 Transformed)
Ki67 Proliferation Index (Log2 Transformed)
3
*P=2.27x10-2
**** P=6.78x10-7
CD68 Index
(Log2 Transformed)
8
1
6
2
·
6
0
4
1
4
-1
2
0
2
-2
0
-1
0
Control
Atrophic
Control
Atrophic
Nodule
Gland
Nodule
Gland
Extended Data Fig. 7 | Immune cell recruitment in the adrenal gland is sex- and age-dependent. (a, b) Histological evaluation of female control and Znrf3 cKO adrenal tissue based on H&E. Female Znrf3 cKOs continue to accumulate histiocytes with advanced aging at 44- and 52-weeks of age. (c-d) Male Znrf3 cKOs sustain high levels of histiocytes previously observed as early as 24-weeks of age. Quantification was performed using QuPath digital analysis based on the proportion of histiocyte area normalized to total adrenal cortex area. Scale bars, 100 um. Data from 4- to 24-weeks of age was previously shown in Fig. 3f-i, and is included as reference. (e-f) In situ validation of myeloid cell accumulation based on IHC for CD68 in control and Znrf3 cKO adrenal tissue from female and (g-h) male cohorts at 44- and 52-weeks of age. Data from 4- to 24-weeks of age was previously shown in Fig. 4e-h, and is included as reference. Quantification was
performed using QuPath digital analysis based on the number of positive cells normalized to total nuclei. Each dot represents an individual animal. Box and whisker plots indicate the median (line) within the upper (75%) and lower (25%) quartiles, and whiskers represent the range. Statistical analysis was performed on log2 transformed data using two-way ANOVA followed by Tukey’s multiple comparison’s test. Scale bars, 100 um. (i) At 78-weeks of age, atrophic Znrf3 cKO adrenal glands have a significantly lower Ki67-index and higher CD68-index compared to age-matched controls. (j) In 78-week-old benign Znrf3 cKO adrenals, nodules have a significantly higher Ki67-index and lower CD68-index compared to the background gland. Each dot represents an individual animal. Box and whisker plots represent mean with variance across quartiles. Statistical analysis was performed using two-tailed Student’s t-test.
a
Female Only TCGA-ACC: Progression-Free Survival
100
Probability of Survival (%)
Adrenal Myeloid
Response Score
High
Low
50
*P=3.18x10-2
0
0
50
100
150
200
Months
nature portfolio
| Corresponding author(s): | Kaitlin J. Basham |
| Last updated by author(s): | Mar 31, 2023 |
Reporting Summary
Nature Portfolio wishes to improve the reproducibility of the work that we publish. This form provides structure for consistency and transparency in reporting. For further information on Nature Portfolio policies, see our Editorial Policies and the Editorial Policy Checklist.
Statistics
For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section. n/a Confirmed ☐
☒ The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement ☐
☒ A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly ☐
The statistical test(s) used AND whether they are one- or two-sided
☒ Only common tests should be described solely by name; describe more complex techniques in the Methods section. ☒ ☐ A description of all covariates tested
☐ ☒ A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons ☐ A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient)
☒ AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals) ☐ ☒ For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted Give P values as exact values whenever suitable. ☒ ☐ For Bayesian analysis, information on the choice of priors and Markov chain Monte Carlo settings ☒ ☐ For hierarchical and complex designs, identification of the appropriate level for tests and full reporting of outcomes ☒
☐ Estimates of effect sizes (e.g. Cohen’s d, Pearson’s r), indicating how they were calculated Our web collection on statistics for biologists contains articles on many of the points above.
Software and code
Policy information about availability of computer code
Data collection Ultrasound: VisualSonics Vevo-2100 in vivo micro-imaging system with MS 550D transducer BulkRNAseq: Agilent Bioanalyzer, NEBNext Ultra II Directional RNA Library prep with poly(A) mRNA Isolation protocol and NovaSeq Reagent Kit v1.5 150 x 150bp sequencing; NovaSeq6000 Instrument (Illumina)
scRNA-seq: Countess Automated Cell Counter, Chromium Single Cell 3’ Gene Expression Library Construction Kit v3.1; NovaSeq6000 instrument (Illumina)
Imaging: Zeiss Apotome microscope with AxioCam MRm camera, Pannoramic MIDI II (3DHISTECH) digital slide scanner, Zeiss Axio Scan.Z1 digital slide scanner, or Leica SP8 laser scanning confocal microscope
Histology: Leica Bond Automated Staining Platform Flow Cytometry: BD FACSAria III Cell Sorter
Data analysis
Loupe Browser (V 6.0), QuPath digital pathology software (V 0.3.0), Graph Pad Prism (V 9.3.1), Ingenuity Pathway Analysis (V 01.20.04), G*Power (V3.1), CaseViewer (v 2.4), RStudio (Build 461), StAR (V2.7.6a), Ensemble (Release 102), Clumpify (v38.34), Cutadapt (v1.16), Feature Counts (v1.6.3), DESeq (v1.30.1), Enrichr (v3.1) (https://maayanlab.cloud/Enrichr/), Cell Ranger (V 6.1.2), Cluster Identify PRedictor (available at https://github.com/atakanekiz/CIPR-Package), all other code used is available at https://github.com/HuntsmanCancerInstitute.
For manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors and reviewers. We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Portfolio guidelines for submitting code & software for further information.
Data
Policy information about availability of data
All manuscripts must include a data availability statement. This statement should provide the following information, where applicable:
- Accession codes, unique identifiers, or web links for publicly available datasets
- A description of any restrictions on data availability
- For clinical datasets or third party data, please ensure that the statement adheres to our policy
All sequencing datasets have been deposited to GEO which will be made available upon publication using the following accession code: GSE201127 (GSE201125 Bulk RNAseq; GSE201126 scRNAseq). Patient data analyzed from TCGA-ACC is publicly available using https://www.cbioportal.org/study/summary? id=acc_tcga_pan_can_atlas_2018.
Field-specific reporting
Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection. ☒ Life sciences ☐ Behavioural & social sciences ☐ Ecological, evolutionary & environmental sciences
For a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summary-flat.pdf
Life sciences study design
All studies must disclose on these points even when the disclosure is negative.
| Sample size | Sample size was determined based on power calculations performed using G*Power to obtain > 80% power with 5% type I error. |
| Data exclusions | For analysis performed on TCGA-ACC data obtained from cBioPortal, patients without mRNA expression data were excluded. |
| Replication | Results from the adrenal ultrasound study were not replicated. Rather, we reproduced this finding using an expanded orthogonal approach based on adrenal weight using an increase in animal number and both males and females. Results from the bulk RNAseq were confirmed using orthogonal approaches: IHC (p21, p16, SABeta Gal, 53BP1). Results from scRNAseq were confirmed using IHC. All other experiments were successfully replicated in independent cohorts over various time-points. |
| Randomization | Adrenal glands for bulk RNA-seq (n=4 biological replicates per genotype and stage) were randomly chosen from each cohort of samples. Each replicate for bulk RNA-seq was from an independent animal. For TD139 drug study, male mice were randomly assigned drug or vehicle control (DMSO). For castration studies, male mice were randomly assigned a castration or sham surgical procedure. For all other experiments, mice were assigned to control or Znrf3 cKO groups based on genotype. |
| Blinding | Samples were blinded prior to analysis whenever possible. This included the pathological review of tumors performed on samples from 78- week-old animals. In many experiments, blinding was not feasible since the genotype was evident from image data based on the difference in tissue size. |
Reporting for specific materials, systems and methods
We require information from authors about some types of materials, experimental systems and methods used in many studies. Here, indicate whether each material, system or method listed is relevant to your study. If you are not sure if a list item applies to your research, read the appropriate section before selecting a response.
Materials & experimental systems
Methods
n/a Involved in the study ☐ ☒ Antibodies
n/a Involved in the study ☒ ☐ ChIP-seq ☒
☒ ☐ Eukaryotic cell lines ☐ ☒ Flow cytometry ☒ ☐ MRI-based neuroimaging
☐ Palaeontology and archaeology
☒ Animals and other organisms ☐
☐ ☒ Human research participants ☐ ☒ Clinical data ☒
☐ Dual use research of concern
Antibodies
Antibodies used
1. Rabbit monoclonal anti-Cleaved caspase 3 (Cell Signaling #9664) used at 1:200 dilution, lot #22
2. Rabbit monoclonal anti-Ki67 (Thermofisher MA5-14520) used at 1:200, clone SP6, lot VL3160001A
3. Rabbit polyclonal anti-53BP1 (Novus Biologicals, NB 100-304) used at 1:1000, lot G-3
4. Rabbit monoclonal anti-p21 (Abcam, ab188224) used at 1:7500, clone EPR18021, lot GR3289187-2
5. Rabbit monoclonal anti-CDKN2A (p16INK4a) (Abcam, ab211542) used at 1:1000, clone EPR20418, lot GR3232279-11
6. Goat polyclonal anti-GFP (Abcam, ab5450) used at 1:1000, lot GR3371088-1
7. Rabbit polyclonal anti-CD68 (Abcam, ab125212) used at 1:1000, lot GR3384571-3
8. Rat monoclonal anti-Ki67 (Thermofisher, 14-5698-80) used at 1:2000, clone SolA15, lot #2488559
9. Rabbit monoclonal anti-F4/80 (Cell Signaling, 70076) used at 1:300, lot#8
10. Rabbit monoclonal anti-CD11b (Abcam, ab133357) used at 1:200, clone EPR1344, lot#GR3345111-13
11. Rabbit monoclonal anti CD11c (Cell Signaling, 97585) used at 1:300, lot#5
12. Rabbit monoclonal anti-Ly6g (Cell Signaling, E6Z1T) used at 1:200 lot#4
13. Rabbit monoclonal anti-CD3e (Cell Signaling, 99940) used at 1:250 lot#4
Validation
1. https://www.cellsignal.com/datasheet.jsp?productId=9664&images=1
3. https://www.novusbio.com/products/53bp1-antibody_nb100-304
4. https://www.abcam.com/p21-antibody-epr18021-ab188224.html
5. https://www.abcam.com/cdkn2ap16ink4a-antibody-epr20418-ab211542.html
6. https://www.abcam.com/gfp-antibody-ab5450.html
7. https://www.abcam.com/cd68-antibody-ab125212.html
8. https://www.thermofisher.com/order/genome-database/dataSheetPdf
producttype=antibody&productsubtype=antibody_primary&productId=14-5698-82&version=270
9. https://www.cellsignal.com/datasheet.jsp?productId=70076&images=1 10. https://www.abcam.com/cd11b-antibody-epr1344-ab133357.html
11. https://www.cellsignal.com/datasheet.jsp?productId=97585&images=0 12. https://www.cellsignal.com/datasheet.jsp?productId=87048&images=0 13.https://www.cellsignal.com/datasheet.jsp?productId=99940&images=0
All antibody staining conditions were validated using positive and negative control tissue. All antibodies except 3 and 6 were validated for immunohistochemistry use in mouse tissue. Antibodies 3 and 6 were validated for autofluorescence use in mouse tissue.
Animals and other organisms
Policy information about studies involving animals; ARRIVE guidelines recommended for reporting animal research
| Laboratory animals | Mouse strains used in this study have been previously described: SF1-Cre-high17, Znrf3-floxed13, R26R-mTmG24. All animals were maintained on the C57BI/6J background with a 12-h light/12-h dark cycle and ad lib access to food and water. Littermate control animals were used in all experiments. This study includes males and female mice of 4, 6, 9, 12, 24, 44, 52 and 78 weeks of age. Mice are housed at 68-79F and between 30-70% humidity. |
| Wild animals | No wild animals were used in this study. |
| Field-collected samples | No field collected samples were used in this study. |
| Ethics oversight | All animal procedures were approved by the Institutional Animal Care & Use Committee at the University of Michigan (Protocol #00010217) and the University of Utah (Protocol #21-02009). |
Note that full information on the approval of the study protocol must also be provided in the manuscript.
Human research participants
Policy information about studies involving human research participants
Population characteristics
Describe the covariate-relevant population characteristics of the human research participants (e.g. age, gender, genotypic information, past and current diagnosis and treatment categories). If you filled out the behavioural & social sciences study design questions and have nothing to add here, write “See above.”
Recruitment
Describe how participants were recruited. Outline any potential self-selection bias or other biases that may be present and how these are likely to impact results.
Ethics oversight
Identify the organization(s) that approved the study protocol.
Note that full information on the approval of the study protocol must also be provided in the manuscript.
Clinical data
Policy information about clinical studies
All manuscripts should comply with the ICMJE guidelines for publication of clinical research and a completed CONSORT checklist must be included with all submissions.
Clinical trial registration
Provide the trial registration number from ClinicalTrials.gov or an equivalent agency.
Study protocol
Note where the full trial protocol can be accessed OR if not available, explain why.
Data collection
Describe the settings and locales of data collection, noting the time periods of recruitment and data collection.
Flow Cytometry
Plots
Confirm that:
☒ The axis labels state the marker and fluorochrome used (e.g. CD4-FITC). ☒ The axis scales are clearly visible. Include numbers along axes only for bottom left plot of group (a ‘group’ is an analysis of identical markers). ☒ All plots are contour plots with outliers or pseudocolor plots.
☒ A numerical value for number of cells or percentage (with statistics) is provided.
Methodology
Sample preparation
Adrenal glands were obtained from 6- or 9-week-old mice following rapid decapitation in order to minimize stress-induced transcriptional changes. Adrenals were placed immediately into ice-cold 1X Hank’s Balanced Salt Solution (HBSS) containing calcium and magnesium (Thermo Fisher, 14025134). Tissues were then finely chopped and digested enzymatically using enzymes and reagents from the Neural Tissue Dissociation Kit (Mitenyi Biotec, 130-092-628). All steps were carried out at 4oC, including centrifugation. All tubes and pipette tips used to handle cell suspensions were pre-coated with 3% BSA in HBSS to prevent cell loss. During the dissociation, the cell suspension was gently agitated with mechanical pipetting every 10 minutes and visually assessed under a stereo microscope until the tissue was fully digested. The suspension was then filtered through 70Em filters to obtain a single cell suspension and enzymes were neutralized using HBSS containing 10% Fetal Bovine Serum (FBS). Red blood cells were removed using Red Blood Cell Lysis buffer (Roche, 11814389001) according to manufacturer guidelines and the cells were washed twice in HBSS containing 2% FBS before counting on a Countess Automated Cell Counter (Thermo Fisher).
Instrument
BD FACSAria III Cell Sorter
Software
FACS Diva
Cell population abundance
Our final sorted fraction represented approx 65-75% of the parent population.
Gating strategy
We gated single cells initially using forward scatter-height and forward scatter-width, we then gated on side scatter-height and side scatter-width to ensure a truly single cell population. From our singlet gate we isolated live cells that were DAPI negative (450/50) and Vybrant DyeCycle Ruby (660/20) positive.
☒ Tick this box to confirm that a figure exemplifying the gating strategy is provided in the Supplementary Information.