Abstract
The prognosis of metastatic endometrial carcinoma (EC), one of the most
common gynecological malignancies worldwide, remains poor, and the
underlying driver of metastases is poorly understood. Dysregulation in
estrogen-related signaling and inactivation of tumor suppressor PTEN
are two essential risk factors of EC. However, whether and how they are
interconnected during EC development remains unclear. Here, we
demonstrate that the deacetylase SIRT7 is upregulated in EC patients
and mouse models, facilitating EC progression in vitro and in vivo.
Mechanistically, in an estrogen-dependent fashion, SIRT7 mediates PTEN
deacetylation at K260, promoting PTEN ubiquitination by the E3 ligase
NEDD4L, accelerating PTEN degradation and, consequently, expediting EC
metastasis. Additionally, SIRT7 expression strongly correlates with
poor survival in EC patients with wild-type PTEN, though no significant
correlation is observed in PTEN mutation patients. These results lay
the foundation for the study of targeting estrogen-SIRT7-PTEN axis, to
restore PTEN abundance, offering potential avenues for EC therapy.
Subject terms: Endometrial cancer, Ubiquitylated proteins,
Tumour-suppressor proteins
__________________________________________________________________
Loss of the tumor suppressor PTEN is often observed during endometrial
cancer (EC) progression. Here the authors show that the deacetylase
SIRT7 mediates PTEN deacetylation in an estrogen-dependent manner,
leading to increased ubiquitination and degradation of PTEN to promote
EC metastasis.
Introduction
Endometrial cancer (EC) is the most frequent gynecological malignancy
in developed countries, with continuously increasing incidence and
mortality rates in recent 10 years^[52]1,[53]2. Surgery is the main
initial management of the primary EC tumor, with or without
post-operative radio- or chemotherapy. EC is broadly divided into two
subtypes, endometrioid and non-endometrioid carcinoma. The endometrioid
type contributes to up to 80% of EC and is known as a hormone-dependent
disease, with exposure to estrogen unopposed to progesterone as one of
the major risk factors^[54]3. Overall, the outcome of the early-stage
patients after surgery and radio- or chemotherapy is favorable,
however, for patients who have more advanced diseases with metastasis
to lymph nodes or other organs, available treatments have limited
effects, and the 5-year survival rate is reported to be less than
20%^[55]4. Meanwhile, no therapeutic regimen has been globally
accepted^[56]5,[57]6. Hence, identifying the metastasis drivers of EC
and understanding the EC metastatic mechanisms are critical to benefit
the patients.
The phosphatase and tensin homolog deleted on chromosome ten (PTEN) is
described as a haplo-insufficient tumor suppressor and studies have
demonstrated that partial loss of PTEN is sufficient to promote several
types of malignancies, including breast cancer, prostate cancer, and
uterine cancer, etc.^[58]7–[59]10. As a dual phosphatase, PTEN exerts
tumor suppressor function to modulate PIP3 concentration on
membrane^[60]11, negatively mediating PI3K-AKT pathway. Meanwhile, PTEN
participates in a variety of biological processes through its
substrates such as FAK^[61]12, IRS1^[62]13, CREB^[63]14 , etc. For
endometrial cancer, PTEN is one of the most frequently altered and
inactivated genes^[64]15. Intriguingly, a study has shown the loss of
PTEN protein without mRNA level alterations in endometrioid endometrial
cancer^[65]16. In poor prognosis EC subtypes, serous carcinoma, 45% of
patients have PTEN protein level loss by immunohistochemistry (IHC)
while the PTEN mutation rate is only about 7%^[66]17. More importantly,
the negative PTEN IHC staining is strongly related to a poor survival
rate of patients with advanced EC^[67]18. Collectively, these clues
have suggested that PTEN loss at protein level is a crucial and
independent cancer-promoting event in EC. Furthermore, these findings
indicate that the post-translational regulation of PTEN may play a
critical role in the progression of EC, especially under the PTEN
intact genetic context. However, how PTEN protein is regulated in EC
process remains poorly investigated.
Lysine acetylation and deacetylation have emerged as essential
post-translational modifications involved in various biological
functions^[68]19. Studies have revealed that the acetylation levels of
non-histone proteins can influence the protein function, abundance,
localization, and affinity to other proteins or DNA^[69]20. Notably,
recent studies have shown that crosstalk of the acetylation or
deacetylation with proteomic pathways could precisely tune the cellular
protein abundance. For instance, the acetylation of MOB1, a protein
involved in the Hippo pathway, could promote its stability by limiting
its binding affinity to E3 ligase Praja2 and affecting the subsequent
ubiquitination^[70]21. SIRT2 could deacetylate and stabilize FGL1
protein in a ubiquitination-dependent manner in liver cancer to promote
immune evasion^[71]22. In addition, accumulating evidence has shown
that changes in acetylation level could serve as one of the underlying
mechanisms of cancer development through modifying its oncogenic or
tumor suppressing substrates^[72]20. Thus, acetyltransferases and
deacetylases are recognized as critical potential targets in cancer
therapy. In EC, previous proteogenomic characterization^[73]23 has
pointed out the heterogeneity of the acetylome and suggested that
acetylation regulators play critical roles in the occurrence and
development of endometrial carcinoma. Characterizing such
acetyltransferases and deacetylases will help us to understand the
tumor biology of EC and more importantly, develop promising therapeutic
approaches.
In mammals, sirtuins (SIRT1-7) are a family of nicotinamide adenine
dinucleotide (NAD^+) dependent, evolutionarily conserved deacetylases.
Sirtuin family has seven members sharing similar catalytic cores yet
having various substrates and different subcellular localizations. They
are involved in numerous processes during ageing, cancer, and other
diseases^[74]24,[75]25. Previous quantitative mass spectrometry
analysis implied that in mouse embryonic fibroblasts cells, SIRT7, the
least studied sirtuin protein, might be a potential deacetylase of
PTEN^[76]26, which is the key tumor suppressor in EC described above.
Studies have manifested that, as a mainly nucleolus localized protein,
SIRT7 exhibits deacetylation, desuccinylation and deglutarylation
activities^[77]27. SIRT7 deacetylates H3K18 to drive oncogenic
transformation^[78]28 and its deacetylating function on H3K36 is
associated with heterochromatin silencing and genomic stability^[79]29.
It is also reported that SIRT7 is involved in chromatin remodeling by
desuccinylating H3K122^[80]30. The non-histone substrates of SIRT7
include ATM^[81]31, SMAD4^[82]32, NPM^[83]33 , etc., which makes SIRT7
a critical modulator in genome stability, cellular stress response, as
well as tumor development^[84]34,[85]35. It is lately demonstrated by
several studies that SIRT7 could promote tumorigenesis or aggravate
tumor progression in different cancer types such as pancreatic ductal
adenocarcinoma^[86]36, prostate cancer^[87]37, thyroid cancer^[88]38
and colorectal cancer^[89]39,[90]40. However, the effect of SIRT7 in EC
development is currently unclear.
Herein, we identify the tumor suppressor PTEN as a substrate of SIRT7
in endometrial cancer, and elaborate SIRT7 facilitating endometrial
cancer metastasis. Utilizing human patient samples and two established
spontaneous EC murine models, we find that SIRT7 exhibits higher
expression level in tumor samples. In detail, SIRT7 deacetylates PTEN,
promoting its degradation mediated by E3 ligase NEDD4L in an
estrogen-dependent manner. More importantly, we identify that K260 is a
deacetylation site of PTEN, and the acetylation level of PTEN-K260 is
significantly associated with EC progression both in human and mice.
Our study illustrates one key node of PTEN stability regulation network
and its deacetylation in tumor progression process. We elucidate the
basic pathogenesis of high expressed SIRT7 endometrial cancer and the
possibility of achieving PTEN expression recovery to slow down cancer
progression by suggesting SIRT7 as a promising therapeutic target for
endometrial cancer.
Results
SIRT7 high expression is positively associated with EC progression
Aging is an important and independent risk factor of EC, with
increasing incidence, mortality, and recurrence rate^[91]41. Sirtuins,
known as longevity proteins, yet the links between these longevity
proteins and endometrial cancer are still poorly understood^[92]42.
Since clues suggested that the level of tumor suppressor PTEN was
strongly associated with the development of EC, initially, we aim to
screen that whether the members of Sirtuins could regulate PTEN. The
co-IP results suggested that SIRT1, 6, 7 could interact with PTEN
(Supplementary Fig. [93]1a). Moreover, from the The Cancer Genome Atlas
(TCGA) database of Uterine Corpus Endometrial Carcinoma (UCEC), we
found that the expression level of SIRT6 and SIRT7 were upregulated
while other Sirtuin family members such as SIRT1, SIRT2, SIRT3, SIRT4
were downregulated in UCEC tumor tissue compared with normal uterus
tissue, and SIRT5 showed no difference (Fig. [94]1A and Supplementary
Fig. [95]1b). Of note, SIRT7 was relatively high expressed in high
grade EC samples, a more aggressive subset of ECs with relatively
unfavorable clinical outcome^[96]43, compared to grade 1/2 EC tumors.
The similar pattern was seen in database [97]GSE115810^[98]44, an
expression profiling of EC in different grades carried out previously
(Fig. [99]1B). Moreover, further analysis of TCGA revealed that EC
patients with high SIRT7 expression were associated with poor overall
survival (Supplementary Fig. [100]1c), indicating that SIRT7 expression
might be related with the malignancy and poor prognosis of endometrial
cancer, while EC patients with SIRT6 high expression, oppositely,
having a relatively good survival condition (Supplementary
Fig. [101]1d). These all suggest that among Sirtuin members, SIRT7
might be a PTEN regulator and its high expression might be associated
with EC poor progression.
Fig. 1. High expression of SIRT7 is positively associated with endometrial
cancer progression.
[102]Fig. 1
[103]Open in a new tab
A SIRT7 expression analysis of normal endometrium tissue (n = 35) and
endometrial cancer tissue (n = 544), data from TCGA. B SIRT7 expression
analysis of endometrial cancer in grade 1–2 and high grade, data from
TCGA (left, G1/G2 n = 218, High Grade n = 326), data from
[104]GSE115810 (right, G1/G2 n = 18, High Grade n = 6). C SIRT7
expression analysis of human EC tissue by IHC. Endometrial cancer and
paired para-tumor tissues (tumor n = 21, para-tumor n = 18) were
stained with SIRT7 and quantified by IHC score. Scale bar = 50 μm. D
SIRT7 expression analysis by Western blot. Tissues of paired
endometrial cancer with adjacent normal tissue (n = 10) were analyzed
for SIRT7 expression and quantified. The relative expression level was
calculated by SIRT7/GAPDH. ImageJ software was used. E SIRT7 expression
analysis of human EC tissue by IHC. Primary tumor tissue of EC and
paired metastatic tissues in ovary or fallopian tube (n = 13) were
stained with SIRT7 and quantified by IHC score. Scale bar = 50 μm. F
SIRT7 expression analysis of human EC tissue by IHC. Primary tumor
tissue of EC and paired metastasis tissues in lymph nodes (n = 15) were
stained with SIRT7 and quantified by IHC score. Scale bar = 50 μm. G
The diagram of Lkb1^f/f mice crossed with Pgr-Cre mice. H The
representative photograph, H&E staining and immunohistochemistry
staining of uterus tissue of Lkb1 CKO (Lkb1^f/f Pgr-Cre) mice and
control mice (Lkb1^f/f) at different weeks post birth (n = 5). Scale
bar = 50 μm. I Western blot analysis of the uterus tissue of Lkb1 CKO
mice and control mice at six weeks post birth. Different EC tumor
nodules of Lkb1 CKO mice were isolated and analyzed (n = 6). Relative
Sirt7 expression was quantified by Sirt7/Tubulin. The P value was
calculated with two tailed unpaired t test for (A–C, I) and calculated
with two tailed paired t test for (D–F). Data are presented as mean
values ± s.e.m.
We then evaluated SIRT7 protein level in tumor versus para-tumor
tissues from EC patients in different stages and different pathological
subtypes. Tumor (n = 21) and para-tumor (n = 18) tissues of EC patients
were collected and examined by immunohistochemistry (IHC), significant
higher levels of SIRT7 were detected in tumor samples (Fig. [105]1C)
compared to para-tumor samples. In addition, the western blot analysis
manifested that, in 9 out of 10 patients, SIRT7 was upregulated in
tumor tissues (Fig. [106]1D), collectively suggesting a potential
tumor-promoting role of SIRT7 in endometrial cancer. To further
investigate SIRT7’s role in cancer development in EC patients,
metastatic tumor tissues were collected, including metastasis in
ovaries, fallopian tubes (n = 13), and in lymph nodes (n = 15). SIRT7
expression level, detected by immunohistochemistry, showed a remarkable
increase in metastatic tumor tissues compared to paired primary tumor
tissues (Fig. [107]1E, F). These results imply that SIRT7 has the
potential to facilitate tumor progression and may contribute to tumor
metastasis in EC.
To further investigate the oncogenic function of SIRT7 in EC in vivo,
two spontaneous EC murine models previously validated were utilized.
Given that Lkb1 knockout in uterus by the Cre recombinase could lead to
invasive endometrial malignancies, and low expression levels of Lkb1 in
endometrial cancers correlate with invasiveness^[108]45,[109]46, we
crossed the Lkb1-flox mice with Pgr-cre mice^[110]47,[111]48 to obtain
ablation of Lkb1 in mice uterus (Fig. [112]1G and Supplementary
Fig. [113]1e). The Lkb1-CKO (Lkb1^f/f Pgr-cre) mice developed
spontaneous tumors consistent to what was previously described^[114]45
(Fig. [115]1H), and Sirt7 level was detected at 6 weeks post birth or
12 weeks post birth. As examined by immunohistochemistry and Western
blot, we found that the Sirt7 levels were preferentially upregulated in
uterus tumor tissue of Lkb1 CKO mice compared with normal uterus tissue
in control mice (Fig. [116]1H, I). Moreover, by the time when the Lkb1
CKO mice reached the age of 12 weeks, the uterus carcinoma had invaded
and infiltrated nearly all layers of the uterine muscle, accompanied by
adhesions to surrounding tissues (Fig. [117]1H). At this point, we
observed a significant increase in the Sirt7 staining in the tumor
tissues compared to that in 6-week-old Lkb1 CKO mice, suggesting
Sirt7’s potential role in promoting tumor progression within the murine
model. Furthermore, we verified our findings in another EC mouse model,
Pten^f/f Pgr-Cre spontaneous model (PC model)^[118]49,[119]50. This PC
model could develop uterus carcinoma spontaneously as early as 30 days
post birth and myometrial invasion in 90 days post birth^[120]51. Sirt7
level was detected by Western blot and immunohistochemistry at 10 weeks
post birth, and consequently, upregulation of Sirt7 was also exhibited
in the uterus tumor tissue in Pten CKO mice compared with normal uterus
tissue in control mice (Supplementary Fig. [121]1f, g). Taken together,
these results illustrate that SIRT7 facilitates the progression of
endometrial cancer.
SIRT7 accelerates endometrial cancer metastasis
In order to explore the mechanism of SIRT7-mediated EC tumor
progression, we analyzed the transcriptional profiles of parental
endometrial cancer cells HEC-1B cells or SIRT7 knockdown HEC-1B cells
by RNA-sequencing (Fig. [122]2A). With pathway enrichment analysis, we
found that several tumor migration related pathways were remarkably
downregulated in SIRT7 KD HEC-1B cells (Fig. [123]2B), such as ‘ECM
receptor interaction’, and ‘Focal adhesion’. Consistently, gene set
enrichment analysis (GSEA) also demonstrated that the signatures of
tumor migration ‘epithelial–mesenchymal transition’ (EMT) is one of the
most altered pathways in SIRT7 KD cells compared to control cells
(Fig. [124]2C). Similarly, in TCGA database, when we compare the
differentially expressed genes between SIRT7 high expression and low
expression groups, the epithelial mesenchymal transition (EMT) gene set
is significantly enriched in GSEA analysis (Supplementary
Fig. [125]2a). These findings indicated that SIRT7 might play a crucial
role in EC metastasis. Hence, we separately generated SIRT7 KD cell
lines with two non-overlapping shRNAs in two EC cell lines, HEC-1B and
KLE (Supplementary Fig. [126]2b). As shown in Fig. [127]2D–G, SIRT7
knockdown reduced the migratory and invasive capacity of HEC-1B and KLE
cells by wound healing and transwell assays. Particularly, SIRT7
knockdown downregulated more than 85% of the invasion rate of HEC-1B
cells (Fig. [128]2E). However, no significant changes were observed in
cell growth and viability in vivo and in vitro (Supplementary
Fig. [129]2c–g). We also did not see obvious morphological changes in
HEC-1B cells after SIRT7 knockdown (Supplementary Fig. [130]2h). Also,
some EMT markers were examined in HEC-1B shRFP/shSIRT7 cells, and we
found the downregulation of N-cadherin, ZEB1 and upregulation of ZO-1
after SIRT7 knockdown, no changes were observed in the protein levels
of Vimentin, β-catenin or SLUG (Supplementary Fig. [131]2i).
Fig. 2. SIRT7 accelerates endometrial cancer metastasis.
[132]Fig. 2
[133]Open in a new tab
A The schematical process of RNA-Sequencing. Differentially expressed
genes based on the RNA-Seq results were analyzed by KEGG enrichment (B)
and GSEA enrichment (C). D Representative images of wound healing assay
for HEC-1B cells with or without SIRT7 knockdown. The area of migration
was calculated in six random views and normalized by shRFP group, scale
bar = 500 μm, n = 6, cell cultures from three independent experiments.
E Representative images of transwell assay and invasion assay for
HEC-1B cells with or without SIRT7 knockdown. Cell number of six random
views and normalized to control for migration or invasion ability
quantification, scale bar = 200 μm, n = 6, cell cultures from three
independent experiments. F Wound healing assay for KLE cells with or
without SIRT7 knockdown, n = 3, cell cultures from three independent
experiments, scale bar = 500 μm. G Transwell assay and invasion assay
for KLE cells with or without SIRT7 knockdown, n = 4, cell cultures
from four independent experiments, scale bar = 200 μm. H Diagram of the
construction of a liver metastasis mouse model. Created in BioRender.
Hu, Z. (2025) [134]https://BioRender.com/f78p401. I Representative
images of the mice liver, spleen and H&E-stained liver demonstrating
the normal liver tissue and metastasis nodules in spleen injection
assay. The metastasis area was marked with asterisks and arrows, scale
bar = 100 μm. Tumor burden analysis was calculated by tumor area/liver
area identified by H&E staining, n = 6. J Diagram of the uterus
orthotopic injection mouse model. Created in BioRender. Hu, Z. (2025)
[135]https://BioRender.com/m69e532. K Representative images of the mice
liver, uterus and H&E-stained liver in uterus orthotopic injection
assay. The incidence of metastasis nodules occurred on liver was shown
in the table and the tumor burden was calculated. shRFP group n = 11,
shSIRT7-1 group n = 9, shSIRT7-2 group n = 9, scale bar = 100 μm. Data
are presented as mean values ± s.d. for (D–G), and ±s.e.m. for (I, K).
The P value was calculated with two tailed unpaired t test.
Next, to explore the in vivo function of SIRT7 in tumor metastasis, we
utilized the hepatic metastasis model by intrasplenic
injection^[136]52,[137]53 in nude mice. SIRT7 proficient (WT) and
deficient (KD) HEC-1B cells were injected under the spleen capsule
after the mice were anesthetized. Eight weeks post inoculation, the
metastasis nodules on liver were photographed and measured by H&E
staining (Fig. [138]2H). Astonishingly, SIRT7 knockdown could eliminate
the tumor metastasis process to liver in mice, with reduction of tumor
burden from 8% to less than 1% (Fig. [139]2I). To better simulate the
metastasis context and pattern of endometrial cancer in situ, we also
established the orthotopic tumor model in nude mice. HEC-1B cells with
or without SIRT7 knockdown were injected into the uterus lumen of nude
mice after anesthesia. The surgical orthotopic injection was performed
as indicated^[140]54–[141]56 and particular attention was paid to
ensure no visible leakage occurred (Fig. [142]2J). It is manifested
that in the shRFP group, seven out of eleven mice developed diffused
liver metastases following in utero injection, whereas only 1/9 and 0/9
mice in the SIRT7 knockdown group exhibited liver metastasis
(Fig. [143]2K). The significant decrease in tumor burden observed after
SIRT7 knockdown in the orthotopic tumor model suggests the effective
inhibition of EC cell metastasis in nude mice by SIRT7 deletion.
Collectively, these results demonstrate that SIRT7 knockdown could
abolish tumor metastasis ability in endometrial cancer both in vivo and
in vitro.
SIRT7 interacts with and directly deacetylates PTEN
As mentioned above, PTEN, the key tumor suppressor in endometrial
cancer, is closely associated with metastasis and could interact with
SIRT7. To investigate the relationship and connection of PTEN and SIRT7
in EC, we hypothesized that SIRT7 might functionally regulate PTEN via
deacetylation. Thus, further co-immunoprecipitation assays were
performed in cell lysates of HEK293FT cells ectopically expressing
FLAG-SIRT7 and GFP-HA-PTEN. The western blot analysis showed the
interaction between the ectopically expressed SIRT7 and PTEN, validated
with anti-HA or anti-FLAG antibodies (Fig. [144]3A, B). Moreover,
endogenous interactions were further verified in HEK293FT and HEC-1B
cell lines, both precipitated wild type PTEN (Fig. [145]3C). To further
confirm the direct interaction between SIRT7 and PTEN, GST pull-down
assay was conducted with HIS-SIRT7 and GST-PTEN proteins purified from
E. coli, and correspondingly, purified SIRT7 and PTEN showed a direct
interaction in vitro (Fig. [146]3D). Furthermore, binding domain of the
two proteins were investigated with the aim of detailing the interplay
pattern of SIRT7 and PTEN. We fragmented SIRT7 based on its functional
domains, and found out both SIRT7-FL and its enzymatic core fragment
(amino acid 90-331) interacted with PTEN (Fig. [147]3E). We also
fragmented PTEN into PTEN- N terminus (amino acid 1-185) and C terminus
(amino acid 186-403), and as a result, the N terminus, including the
dual-specificity phosphatase (DUSP) domain and a PIP2-binding motif
(PBM)^[148]57, was identified to interact with SIRT7 (Fig. [149]3F).
Fig. 3. SIRT7 interacts with and directly deacetylates PTEN.
[150]Fig. 3
[151]Open in a new tab
A, B Co-IP assays of the interaction of GFP-HA-PTEN and FLAG-SIRT7 in
HEK293FT cells. GFP-HA-PTEN and FLAG-SIRT7 were overexpressed in
HEK293FT cells and immunoprecipitants were enriched with anti-HA (A)
antibodies or anti-FLAG (B) antibodies. C Endogenous Co-IP assays of
the interaction between SIRT7 and PTEN using antibodies against PTEN in
HEK293FT cells (upper) and HEC-1B cells (lower). D GST-pulldown assay.
Purified GST-PTEN, GST proteins were harvest with purified HIS-SIRT7
and precipitated with anti-GST antibodies then subjected to
immunoblots. E Truncations of GST-SIRT7 were constructed and the
truncation proteins were harvested with FLAG-PTEN and precipitated with
anti-GST antibodies. F Truncations of GST-PTEN were constructed and the
truncation proteins were harvested with FLAG-SIRT7 and precipitated
with anti-GST antibodies. G Acetylation level of PTEN in HEK293FT cells
while SIRT7-WT or SIRT7-H187Y was overexpressed. The pan-Acetyl
antibodies were used to detect the acetylation level of PTEN after
immunoprecipitation. H Acetylation level of PTEN in HEK293FT cells with
or without SIRT7 knockdown in HEK293FT cells. I Acetylation level of
PTEN in HEK293FT cells. For HEK293FT cells, SIRT7 were knockdown and
then FLAG-SIRT7 was re-expressed. J, K The immunoblots showing the
acetylation level of PTEN in vitro. FLAG-SIRT7 and HA-PTEN was purified
by antibodies from cell lysates. PTEN acetylation level was detected by
pan-acetylation antibodies after harvesting with or without SIRT7,
NAD^+ and NAM as indicated. L Immunoblots showing acetylation level of
WT-PTEN and KR mutants (K60R, K223R, K260R, K344R, 4KR), detected by
pan-acetyl antibodies after immunoprecipitation. To ensure similar
expression levels among various PTEN-mut-plasmids, plasmids of
different volumes are transfected. M Amino acid sequences of PTEN
(aa251-264) of human, mouse, rat, and dog. N Immunoblots showing
acetylation levels of WT PTEN and PTEN-K260R in vitro after incubation
with or without SIRT7 and NAD^+, detected by pan-AcK antibodies. O
Immunoblots showing PTEN-K260 acetylation levels of PTEN-WT/K260R/K260Q
after immunoprecipitation. The PTEN-K260 acetylation level was detected
by AcK260-PTEN specific antibody after immunoprecipitation. P
Immunoblots showing the K260 acetylation level of endogenous PTEN after
immunoprecipitation in sgSIRT7 (pool) or sgScramble HEK293FT cells. The
PTEN-K260 acetylation level was detected by AcK260-PTEN specific
antibody.
To gain further insights into how SIRT7 regulates PTEN, we hypothesized
that SIRT7 could deacetylate PTEN. The acetylation level of ectopic
PTEN was decreased when wild type SIRT7 (SIRT7-WT) is overexpressed in
HEK293FT cells but not the enzymatic dead mutants, SIRT7-H187Y
(Fig. [152]3G). Consistently, knockdown of SIRT7 by two independent
short hairpin RNAs led to a marked increase in the ectopic PTEN
acetylation level (Fig. [153]3H), and such increase could be blocked by
reintroduction of the SIRT7 construct (Fig. [154]3I). Moreover, for the
purpose to identify the direct role of SIRT7 in deacetylating PTEN, in
vitro deacetylation assays were performed. The acetylation level of
PTEN almost vanished in the presence of SIRT7 along with NAD^+
(Fig. [155]3J), while remained unchanged if the reaction was inhibited
by nicotinamide (NAM) (Fig. [156]3K), which is a general sirtuin
inhibitor^[157]28. Collectively, these results demonstrate that SIRT7
directly interacts and deacetylates PTEN.
To further map the deacetylation sites of PTEN by SIRT7, we performed a
mass spectroscopy analysis (Supplementary Fig. [158]3a), and 4 sites
were identified as the potential deacetylation sites, K60, K223, K260,
K344 (Supplementary Fig. [159]3b). To determine the PTEN deacetylation
sites, we replaced each lysine site with amino arginine separately
(K60R, K223R, K260R, K344R), or together (4KR). WT-PTEN and the mutant
PTEN plasmids were transiently transfected into HEK293FT cells, and
acetylation levels of WT and mutant PTEN were determined by Western
blot (Fig. [160]3L). The acetylation level was faint only when K260R-
or 4KR- PTEN was transfected, indicating K260, an evolutionarily
conserved lysine residue (Fig. [161]3M), might be the predominant
acetylation site of PTEN. On top of that, in vitro deacetylation assays
showed that FLAG-SIRT7 purified from cell lysates could lead to the
significantly decreased acetylation level of WT-PTEN in the presence of
NAD^+, but little alteration was observed when expressing the
K260R-PTEN mutant (Fig. [162]3N), which suggested that K260-PTEN could
be the deacetylation site regulated by SIRT7.
To precisely monitor the acetylation level at PTEN-K260 and to
ascertain the deacetylation of PTEN modulated by SIRT7, we generated a
specific antibody for PTEN-K260 acetylation (AcK260-PTEN). The
specificity and efficiency of this antibody were demonstrated by dot
blot experiments with modified or unmodified peptides (Supplementary
Fig. [163]3c). As a result, the acetylation levels of PTEN-K260 mutant
and WT-PTEN were evaluated by AcK260-PTEN antibody, which furthermore
indicated that PTEN was deacetylated at K260 site (Fig. [164]3O).
Furthermore, the acetylation level of endogenous PTEN was examined
using AcK260-PTEN antibody and consequently, the acetylation level of
PTEN-K260 remarkably elevated after SIRT7 depletion (Fig. [165]3P). In
in vitro assay, the acetylation level of PTEN-K260 almost abolished
after incubating PTEN with SIRT7 and NAD^+ (Supplementary
Fig. [166]3d). These results suggest that SIRT7 could regulate PTEN
deacetylation at K260. Of note, the KQ mutation of PTEN-K260 did not
change the SIRT7-PTEN interaction (Supplementary Fig. [167]3e). In
addition, we assessed the acetylation status of Pten-K260 in mouse
uterine tissue following immunoprecipitation and the findings revealed
that in EC tumor tissues from Lkb1 CKO mice, the acetylation level of
Pten-K260 was significantly reduced compared to that in the uterine
tissue of control mice (Supplementary Fig. [168]3f), which was
consistent with the elevation of Sirt7 level in EC tumor mice as
demonstrated in Fig. [169]1. Altogether, these results suggest that
PTEN is a bona fide substrate of SIRT7 and it can be deacetylated by
SIRT7 at lysine 260.
SIRT7-mediated PTEN deacetylation is associated with PTEN instability
We then proceeded to further explore the function of PTEN deacetylation
by SIRT7. Astonishingly, we noticed that in SIRT7 knockdown HEC-1B
cells, endogenous protein level of PTEN was significantly higher than
that in control cells (Fig. [170]4A) and the phenomenon was confirmed
in another EC cell line KLE (Supplementary Fig. [171]4a). Also, similar
pattern was exhibited when SIRT7 was knocked out in HEC-1B cells with
three different sgRNAs (Supplementary Fig. [172]4b). Correspondingly,
the level of PTEN showed a dose-dependent decrease while overexpressing
SIRT7-WT (Fig. [173]4B and Supplementary Fig. [174]4c), however,
overexpressing enzymatic dead mutant SIRT7-H187Y abolished the PTEN
decrease (Fig. [175]4C) in HEC-1B cells. To determine whether SIRT7
regulates PTEN expression by impeding its transcription, we examined
the mRNA level and the promoter activity of PTEN, and there were no
significant changes when we overexpressed or knocked down SIRT7
(Supplementary Fig. [176]4d–g), suggesting that SIRT7 could regulate
PTEN expression through a post-transcriptional manner.
Fig. 4. SIRT7-mediated PTEN deacetylation is associated with PTEN
instability.
[177]Fig. 4
[178]Open in a new tab
A–C PTEN protein level shown by immunoblots and quantification analysis
of three independent repeats (n = 3). SIRT7 was knocked down (A),
otherwise SIRT7 (B) or SIRT7-WT/H187Y (C) was overexpressed in HEC-1B
cells. The relative expression level was analyzed in Image J,
calculated by PTEN/Tubulin, and normalized by the control lane. The P
value was calculated with two tailed unpaired t test. D The immunoblots
showing the half-life of protein PTEN in HEC-1B cells with or without
SIRT7 knockdown and the quantitative analysis of three independent
repeats (n = 3). HEC-1B cells were treated with CHX (100 μg/ml) for
indicated time and subjected for immunoblot analysis. HEC-1B shSIRT7
stable cell line was used and relative PTEN protein level was analyzed
in Image J, calculated by PTEN/Tubulin, and normalized by the lane
‘0 h’. E Immunoblots showing the half-life of protein PTEN in HEC-1B
cells over-expressing SIRT7-WT or SIRT7-H187Y and quantitative analysis
(n = 3). F Immunoblots showing the half-life of protein PTEN-WT or
PTEN-K260Q overexpressed in HEC-1B cells and quantitative analysis
(n = 3). Two-way ANOVA was used for the P value calculation for (D–F).
G Representative images of the mice uterus in immunofluorescence assay.
The frozen sections of mice uterus tissue from control mice (Sirt7^f/f)
and Sirt7 uterus CKO mice (Sirt7^f/f Pgr-cre) (n = 3) were stained with
SIRT7, PTEN and DAPI. Scale bar = 50 μm. H Representative images of IHC
staining of the uterus tissue from control mice (Sirt7^f/f) and Sirt7
uterus CKO mice (Sirt7^f/f Pgr-cre) (n = 5). The uterus tissue of the
indicated mice was stained with SIRT7, PTEN, AcK260-PTEN antibodies.
Scale bar = 50 μm. Data are presented as mean values ± s.e.m.
Next, we wonder how SIRT7 effects on the turnover of PTEN protein. We
assayed the half-life of PTEN under the treatment of the protein
synthesis inhibitor cycloheximide (CHX). The half-life of PTEN extended
from ~20 hrs in control cells to more than 36 hrs in SIRT7 knockdown
cells (Fig. [179]4D and Supplementary Fig. [180]4h). Moreover,
overexpression of SIRT7 could expedite PTEN turnover with CHX
inhibition while the SIRT7-H187Y enzymatically dead mutants could not
(Fig. [181]4E and Supplementary Fig. [182]4i), which implied that SIRT7
accelerated the degradation of PTEN via its deacetylase activity.
Hence, we deduced that SIRT7 might regulate the instability of PTEN by
regulating PTEN deacetylation. To further confirm the association of
stability and acetylation level of PTEN, the turnover of PTEN mutants
at K260 was examined. As a result, the half-life of PTEN proteins was
remarkably extended when PTEN-K260Q was overexpressed, mimicking a
constant acetylated PTEN version (Fig. [183]4F). Moreover, the
half-life of PTEN-K260Q did not change as SIRT7 was overexpressed in
HEC-1B cells (Supplementary Fig. [184]4j). These data strongly suggest
that SIRT7 mediates PTEN deacetylation at K260 and accelerates the
degradation of PTEN.
To further explore SIRT7-modulated function of PTEN in vivo, we
established the conditional SIRT7 knockout mice in uterus
(Supplementary Fig. [185]4k). The immunofluorescence staining of the
cross section of mouse uterus showed that SIRT7 was knocked out in most
uterus tissues due to Pgr expressing and relatively strong PTEN
staining was observed in SIRT7 knocked out samples (Fig. [186]4G),
indicating a negative correlation between SIRT7 and PTEN expression.
Such correlation was also seen in IHC staining of Sirt7 CKO and WT
mice. Most importantly, the acetylation level of K260-PTEN was
manifested to significantly arise in Sirt7 CKO mice (Fig. [187]4H),
highlighting that the SIRT7-mediated PTEN downregulation was dependent
on the deacetylation of PTEN by SIRT7. In all, these findings present a
SIRT7-PTEN modulation pattern that SIRT7 deacetylates PTEN at K260,
consequently favoring PTEN degradation in vitro and in vivo.
PTEN-K260 deacetylation by SIRT7 leads to NEDD4L-mediated ubiquitination
We next sought out to explore whether proteasomes or lysosome pathways
are involved in SIRT7 mediated PTEN degradation. HEC-1B cells
overexpressing SIRT7 or empty vector were treated by proteasome
inhibitor MG132, lysosome inhibitor chloroquine (CHQ), or DMSO
respectively. Consequently, PTEN level reduced by ~40% after SIRT7
overexpression, and such effect was blocked by MG132 treatment
(Fig. [188]5A), while lysosome inhibition only exerted faint changes on
PTEN level, indicating that SIRT7 mediated PTEN degradation was mainly
dependent on proteasome. We next investigated how SIRT7 affects the
ubiquitination levels of PTEN. It was shown by the immunoprecipitation
assays that the ubiquitination levels of PTEN were substantially
elevated in cells with GFP-SIRT7 overexpression (Fig. [189]5B).
Conversely, the ubiquitination level of PTEN dramatically declined in
SIRT7 deficient HEK293FT cells, when PTEN was expressed both
ectopically and endogenously (Fig. [190]5C and Supplementary
Fig. [191]5a). These pieces of evidence pointed out that SIRT7 could be
involved in promoting PTEN ubiquitin-proteasomes dependent degradation.
Furthermore, in line with the state of SIRT7 downregulation, the
ubiquitination of PTEN-K260Q mutant maintained a lower level compared
to PTEN WT, and conversely, the ubiquitination level of PTEN-K260R
mutant was higher than PTEN WT (Fig. [192]5D), indicating that the
PTEN-K260 acetylation level regulated by SIRT7 was crucial for its
ubiquitination.
Fig. 5. PTEN-K260 deacetylation by SIRT7 contributes to NEDD4L-mediated
ubiquitination.
[193]Fig. 5
[194]Open in a new tab
A Immunoblot manifesting PTEN protein level and the quantitative
analysis for three repeats (n = 3). SIRT7 was overexpressed in HEC-1B
cells, and the cells were treated with DMSO or MG132 (10 μM, 8 h) or
CHQ (20 μM, 16 h) before being collected. The error bars indicate the
s.e.m. values. The P value was calculated with two tailed unpaired t
test. B Ubiquitination level of PTEN with or without SIRT7
overexpression in HEK293FT cells. C Ubiquitination level of PTEN in
sgSIRT7 (pool) or sgScramble HEK293FT cells. D Ubiquitination level of
PTEN-WT/ K260Q/ K260R. For B–D, the HA-Ub and FLAG-EV,
FLAG-PTEN-WT/K260Q/K260R were overexpressed in HEK293FT cells as
indicated, and the ubiquitination level was detected after
immunoprecipitation. E Mass spectrum results suggesting the E3 ligase
interacting with PTEN and ranked by number of unique peptides. F PTEN
protein level shown by immunoblots after NEDD4L knockdown. NEDD4L was
deleted in HEC-1B cells by short hairpin RNAs. G Ubiquitination level
of FLAG-PTEN after NEDD4L knockdown. HA-Ub and FLAG-PTEN were over
expressed, and the ubiquitination level was detected after
immunoprecipitation. H Ubiquitination level of FLAG-PTEN after NEDD4L
knockdown and SIRT7-Myc overexpression. HA-Ub and FLAG-PTEN were over
expressed, and the ubiquitin level was detected after
immunoprecipitation. I Co-IP manifesting the interaction between
endogenous NEDD4L and PTEN after Myc-SIRT7 overexpression in HEK293FT
cells. J Co-IP manifesting the interaction between endogenous NEDD4L
and PTEN-WT/K260Q after Myc-SIRT7 overexpression in HEK293FT cells.
We then sought out to figure out the corresponding E3 ligase in this
SIRT7-PTEN axis. With Co-IP and mass spectrum, we identified several E3
ligases associated with PTEN (Fig. [195]5E). However, only NEDD4L and
TRIM25 were negatively correlated with PTEN in endometrial cancer at
the protein levels as analyzed in CPTAC database (Supplementary
Fig. [196]5b), inferring that they might be the potential E3 ligases of
PTEN in endometrial cancer. To investigate whether NEDD4L or TRIM25 is
involved in the SIRT7-mediated PTEN degradation, NEDD4L and TRIM25 were
knocked down respectively with short hairpin RNAs in EC cell line
HEC-1B. Our results showed that PTEN level was significantly increased
by NEDD4L depletion while TRIM25 knockdown only brought about a slight
change (Fig. [197]5F and Supplementary Fig. [198]5c). In addition,
ubiquitination assay confirmed that ubiquitination level of exogenous
and endogenous PTEN both declined in NEDD4L knock down HEK293FT cells
(Fig. [199]5G and Supplementary Fig. [200]5d), and SIRT7 overexpression
would no longer elevate PTEN ubiquitination in NEDD4L-deficient
condition (Fig. [201]5H and Supplementary Fig. [202]5e). Moreover,
NEDD4L or SIRT7 overexpression could increase the ubiquitination level
of PTEN-WT, however, could not change the ubiquitination of PTEN-K260Q
(Supplementary Fig. [203]5f, g). Taken together, these results
demonstrates that NEDD4L is responsible for PTEN ubiquitination,
following its deacetylation by SIRT7 in endometrial cancer.
Next, we stepped deeper to investigate the regulation pattern of SIRT7
in NEDD4L mediated PTEN ubiquitination and found out that SIRT7
overexpression could enhance the interaction between PTEN and NEDD4L
(Fig. [204]5I and Supplementary Fig. [205]5h). Notably, the interaction
between PTEN and NEDD4L almost vanished when PTEN was mutated to
PTEN-K260Q, while SIRT7 overexpression could no longer enhance the
interaction between PTEN-K260Q and NEDD4L (Fig. [206]5J). These
results, collectively, imply that SIRT7-mediated PTEN deacetylation
could facilitate PTEN ubiquitination by promoting the interaction
between NEDD4L and PTEN. In summary, these findings suggest that
PTEN-K260 deacetylation mediated by SIRT7 promotes PTEN degradation by
favoring its ubiquitination, which could be mediated by the E3 ligase
NEDD4L in endometrial cancer.
PTEN-K260 deacetylation by SIRT7 promotes endometrial cancer metastasis
To further explore the biological and functional relevance of SIRT7
mediated PTEN deacetylation and degradation, we sought to decipher
whether SIRT7 facilitates tumor metastasis via promoting PTEN-K260
deacetylation. To assess whether SIRT7’s role in tumor migration is
PTEN related, the transwell assay was performed in ISHIKAWA cells, an
EC cell line with PTEN deletion^[207]58, and it turned out that SIRT7
had minimal effect on migratory and invasive abilities of ISHIKAWA
cells because of the depletion of PTEN (Supplementary Fig. [208]6a).
Furthermore, we established PTEN depleted HEC-1B cells with two
separated short hairpin RNAs and found that SIRT7 knockdown did not
give impetus to the migration or invasion in HEC-1B cells with the
absence of PTEN (Supplementary Fig. [209]6b, c), whereas there was a
drastic downregulation in HEC-1B cells with adequate WT PTEN expression
as we previously shown (Fig. [210]2D–G), which demonstrates that SIRT7
promotes endometrial cancer metastasis through a PTEN-dependent manner.
To further explore the effect of the SIRT7-mediated PTEN-K260
deacetylation during endometrial cancer metastasis, we re-expressed
shRNA target sequence-modified plasmids rPTEN-WT or rPTEN-K260Q in
HEC-1B PTEN knocked down cells (Supplementary Fig. [211]6d). With wound
healing and transwell assays, we found that rPTEN-WT inhibited cell
migration and invasion (Fig. [212]6A, B), corresponding to the previous
report that PTEN modulates cell migration negatively^[213]59–[214]62.
More importantly, rPTEN-K260Q exerted an even stronger
migration-inhibiting activity than rPTEN-WT, while rPTEN-K260R had a
less inhibitory effect on migration and invasion compared to rPTEN-WT
(Fig. [215]6A, B), suggesting that the K260 acetylation level of PTEN
regulated by SIRT7 could modulate cell migratory ability in EC. In
addition, cell migration and invasion were exacerbated when
overexpressing SIRT7, which could be rescued by the expression of
PTEN-WT, and the expression of PTEN- K260Q further abolished the cell
migration and invasion activity (Fig. [216]6C, D). To further explain
the SIRT7-PTEN axis in vivo, we then injected HEC-1B cells expressing
rPTEN-WT/ 260Q using the spleen injection metastasis mouse model. The
results showed that cells with rPTEN-WT expressing developed less liver
metastasis tumors in mice than PTEN deficient HEC-1B cells. Meanwhile,
the metastasis process was further blocked in mice that were injected
with HEC-1B rPTEN-K260Q cells (Fig. [217]6E). Moreover, in uterus
orthotopic injection model, HEC-1B cells expressing empty vector,
rPTEN-WT, or rPTEN-260Q were injected into the uterus lumen of nude
mice. It was manifested that all the mice in the EV group develop
metastatic nodules in the liver after injection, however, only four out
of six mice injected with HEC-1B cells expressing rPTEN-WT afflicted
with liver metastasis. More importantly, only one out of six mice
injected with HEC-1B cells expressing rPTEN-K260Q mutant emerged
disseminated masses on liver, and the tumor burden of K260Q group
significantly decreased as well, compared to rPTEN-WT group
(Fig. [218]6F), all inferring that the high acetylation status of
PTEN-K260 could exhibit a more metastasis inhibitory role than WT-PTEN
both in vitro and in vivo.
Fig. 6. PTEN-K260 deacetylation by SIRT7 promotes endometrial cancer
metastasis.
[219]Fig. 6
[220]Open in a new tab
A Wound healing assay in HEC-1B-rescue cells. For HEC-1B-rescue cells,
PTEN was knocked down and empty vector, rPTEN-WT, rPTEN-K260Q or
rPTEN-K260R was expressed. Relative area of migration was measured and
normalized to the ‘EV’ group, n = 6, cell cultures from three
independent experiments, scale bar = 500 μm. B Transwell and invasion
assay in HEC-1B-rescue cells. Relative cell number was quantified,
n = 8, cell cultures from four independent experiments, scale
bar = 200 μm. C Wound healing assay in HEC-1B cells. SIRT7 was
overexpressed and then PTEN-WT, PTEN-K260Q or empty vector was then
expressed, n = 6, cell cultures from three independent experiments,
scale bar = 500 μm. D Transwell and invasion assay in HEC-1B cells.
SIRT7 was overexpressed and then PTEN-WT, PTEN-K260Q or empty vector
was then expressed, n = 8, cell cultures from three independent
experiments, scale bar = 200 μm. E Representative images of the mice
liver, spleen, and H&E-stained liver after the injection of
HEC-1B-rescue cells into the spleen. Tumor burden was calculated, EV
group n = 12, rPTEN-K260Q/rPTEN-K260R group n = 10, scale bar = 100 μm,
metastasis nodules marked with asterisks. F Representative images of
the mice liver, uterus, and H&E-stained liver, n = 6, scale
bar = 100 μm. The incidence of liver metastasis nodules was shown and
tumor burden was calculated. G AcK260-PTEN level analysis of human EC
tissue by IHC. Primary tumor and paired metastasis tissue (in
ovary/fallopian, n = 13) or metastatic lymph nodes (n = 15) were
stained with AcK260-PTEN and quantified by IHC score. Scale bar = 50 μm
H AcK260-PTEN level analysis of human EC tissue by IHC. Tumor (n = 21)
and paired para-tumor tissue (n = 18) were stained with AcK260-PTEN,
and quantified by IHC score. Scale bar = 50 μm. The P value was
calculated with two tailed unpaired t test for (A–F, H), and was
calculated with two tailed paired t test for (G). I Kaplan–Meier
overall survival analysis of EC patients with or without PTEN mutation,
data from the TCGA, n = 175 for PTEN-WT (SIRT7 high = 117; SIRT7
low = 58), n = 332 for PTEN-mut (SIRT7 high = 222; SIRT7 low = 110).
Cut-off = 0.33. Data are presented as mean values ± s.d. for (A–D), and
± s.e.m. for (E–H).
Next, we stepped to explore how NEDD4L was involved in the SIRT7-PTEN
regulated cell migratory function. It was shown in transwell assay that
NEDD4L deficiency cast a negative influence on cell migration and
invasion in HEC-1B cells, however such effect was abolished in PTEN
deficient cells (Supplementary Fig. [221]7a), implying NEDD4L could
promote EC cell migration in a PTEN dependent manner. SIRT7
overexpression could promote cell migration and invasion in control
cells, detected by transwell assay, however, in cells with NEDD4L
depletion, such significant change was not observed. Moreover, further
knockdown of PTEN on this basis could once again enhance the cells’
migration and invasion abilities, indicating SIRT7’s role in promoting
cell migration is PTEN and NEDD4L dependent. (Supplementary
Fig. [222]7b). Furthermore, in transwell assays, we observed that
NEDD4L overexpression could lead to an increase of cell migration and
invasion with PTEN-K260 in a low acetylation status (mimicked by
K260R), while could no longer cause significant alterations to cell
migration while PTEN-K260 is in a high acetylation level (mimicked by
KQ mutant) (Supplementary Fig. [223]7c), emphasizing that deacetylation
of PTEN-K260 regulated by SIRT7 is crucial to NEDD4L involved cell
migration regulation process.
Furthermore, to explore whether the SIRT7-mediated PTEN deacetylation
correlated with EC progression in patients, we detected PTEN-K260
acetylation level with the AcK260-PTEN antibody in human tissues. The
IHC staining showed that AcK260-PTEN was remarkably decreased in
metastatic EC tumor tissues in ovary, fallopian tube and lymph node
compared to primary tumor tissues (Fig. [224]6G) and was also
diminished in tumor tissues compared to para-tumor tissues
(Fig. [225]6H). These results were corresponding to the findings above
that SIRT7 expression was elevated in EC tumor and metastatic tumor
tissues (Fig. [226]1). In addition, the PTEN staining also showed a
significant decrease in metastatic tumor tissues compared to primary
tumor tissues, and showed reduced expression in tumor tissues compared
to para-tumor tissues (Supplementary Fig. [227]7d, e), indicating a
negative correlation between SIRT7 and PTEN, which could be due to the
deacetylation of PTEN-K260. Collectively, these results suggest that
SIRT7-mediated PTEN-K260 deacetylation is functional in cell migration
and invasion and the acetylation level at K260 of PTEN could serve as a
prognosis marker for endometrial cancer metastasis.
As mentioned before, we have shown that SIRT7’s role in promoting tumor
migration is PTEN-dependent. Considering PTEN is the most frequent gene
mutated in EC^[228]63, we next decided to examine whether SIRT7
expression has different influence on PTEN mutated or PTEN wild type EC
patients. It is demonstrated in TCGA database that, SIRT7
overexpression has been remarkably related to a poorer overall survival
condition of PTEN gene wild type EC patients (p = 0.012), however, in
patients with PTEN gene mutated, no significant relevance was observed
(p = 0.77) (Fig. [229]6H). Though study showed that PTEN mutation took
place in about 65.5% of EC samples, these patients were reported to
have relatively favorable prognosis^[230]64. Correspondingly, TCGA data
manifested that EC patients with wild type PTEN expression have a worse
overall survival compared to PTEN mutated patients (Supplementary
Fig. [231]7f). Thus, our findings of the important oncogenic role of
SIRT7 in EC may emerge SIRT7 as a potential target for the treatment of
PTEN non-mutated EC patients, which, especially, provides insights to
improve the prognosis of EC patients without PTEN mutation.
Estrogen acts as the upstream regulator of SIRT7-PTEN axis
As shown above, SIRT7 acts as an oncogene in EC metastasis process, and
relies on an intact PTEN context. Next, we aimed to explore what is the
pathological regulator for this SIRT7-PTEN axis. It is well known that
endometrial cancer is a hormone dependent disease and that estrogen
exposure in the long term without a progesterone antagonism is one of
the most important risk factors of endometrial cancer^[232]65–[233]67.
We then wondered whether estrogen is involved in the upstream switch of
SIRT7-mediated PTEN deacetylation-ubiquitination process. As shown in
Fig. [234]7A, we observed that along with 17β-estradiol stimulation,
PTEN protein level was declined in a concentration dependent manner,
without any significant change of SIRT7 levels in HEC-1B cells
(Fig. [235]7A). However, in SIRT7 deficient HEC-1B cells, the decline
of PTEN level under estrogen stimulation vanished (Fig. [236]7B), which
suggested that estrogen could influence PTEN expression in a
SIRT7-dependent fashion. To investigate the molecular mechanisms by
which estrogen stimulation induces PTEN decaying by SIRT7, we explored
whether the interplay between SIRT7 and PTEN is affected.
Astonishingly, Proximity Ligation Assay (PLA) assay showed a
significant increase of SIRT7-PTEN association when HEC-1B cells were
treated with 17β-estradiol (Fig. [237]7C). Co-immunoprecipitation assay
further confirmed that 17β-estradiol could reinforce the interaction
between SIRT7 and PTEN in a dose-dependent manner, both in HEC-1B
(Fig. [238]7D) and HEK293FT cells (Fig. [239]7E). These suggest that
estrogen stimulation could promote the association between SIRT7 and
PTEN. However, we noted that overexpression of estrogen-receptor ESR1
did not change SIRT7-PTEN interaction (Fig. [240]7F). Next, we stepped
to investigate the deacetylation process under estrogen stimulation and
found out that 17β-estradiol stimulation abated the acetylation level
of PTEN to a lower level when SIRT7 was overexpressed (Fig. [241]7G).
Notably, in HEC-1B cells, estrogen stimulation could decrease the K260
acetylation level of endogenous PTEN, detected by the AcK260-PTEN
antibody (Fig. [242]7H), suggesting that estrogen stimulation promotes
the deacetylation of PTEN regulated by SIRT7. Moreover, 17β-estradiol
treatment also increased PTEN-NEDD4L interaction, as well as the
ubiquitination level of PTEN (Fig. [243]7I, J). Altogether, these clues
indicate that estrogen stimulation could promote the SIRT7-PTEN axis in
endometrial cancer.
Fig. 7. Estrogen acts as the upstream regulator of SIRT7-PTEN axis.
[244]Fig. 7
[245]Open in a new tab
A Immunoblots showing PTEN level after the cells being treated with 17
beta-Estradiol (E2) for 48 hours with indicated concentration in HEC-1B
cells. B Immunoblots showing PTEN level after estrogen treatment for
48 hours with indicated concentration in HEC-1B cells with or without
SIRT7 knockdown. C Representative pictures and quantitative analysis of
PLA assay demonstrating the interaction of endogenous PTEN and SIRT7
protein with or without E2 stimulation. HEC-1B cells were treated with
E2 (100 nM, 48 h) and MG132 (10 μM, 6 h) as indicated. 100 cells were
quantified. Scale bar = 100 μm. D Co-IP assay manifesting the
interaction between endogenous SIRT7 and PTEN in HEC-1B cells with E2
stimulation for 48 hours with indicated concentration. E Co-IP assay
manifesting the interaction between SIRT7 and PTEN in HEK293FT with E2
stimulation for 48 hours with indicated concentration. F Co-IP assay
manifesting the interaction between SIRT7 and PTEN after ESR1
overexpression in HEK293FT cells. G The acetylation level of PTEN in
HEK293FT cells with stimulation of E2 (50 nM, 48 h) with or without
SIRT7 overexpression. H The AcK260 level of endogenous PTEN in HEC-1B
cells with or without stimulation of E2 (100 nM, 48 h) detected by
AcK260-PTEN antibody. I Co-IP assay manifesting the interaction between
endogenous NEDD4L and PTEN in HEK293FT with SIRT7 overexpression and E2
stimulation as indicated in HEK293FT cells. J The ubiquitination level
of PTEN in HEK293FT cells with stimulation of E2 of indicated
concentration. Cells were treated with MG132 (10 μm) for 6 hours before
lysis. AcK260-PTEN (K) and SIRT7 (L) analysis of human uterus
endometrium tissue during menstrual cycle by IHC. Endometrium in
secretory phase (n = 12) and proliferative phase (n = 17) were stained
with AcK260-PTEN and SIRT7 and quantified by IHC score. Scale
bar = 50 μm. M Schematic diagram showing the mechanisms that SIRT7
upregulation facilitates EC progression through PTEN-K260 deacetylation
and subsequent ubiquitination and degradation under long-term
stimulation of estrogen. Created in BioRender. Hu, Z. (2025)
[246]https://BioRender.com/d55g597. The P value was calculated with two
tailed unpaired t test for (C, K, L). Data are presented as mean
values ± s.e.m.
Notably, estrogen has been explicated to periodically fluctuate during
female menstrual cycle, causing several cyclic biological changes,
including the renovation and thickness change of endometrium^[247]68
and studies have pointed out that the estrogen takes part in pathways
facilitating the growing of the endometrium^[248]69,[249]70. Given that
PTEN expression also alters during the menstrual^[250]71, we then ask
if the estrogen-SIRT7-PTEN axis is involved in the renovation in the
menstrual cycle. We collected the endometrium samples of patients with
no hyperplasia or cancer, who were in the proliferative and secretory
period of menstrual. As a result, IHC staining revealed that during the
secretory phase, the acetylation level of PTEN-K260 in the endometrium
was significantly higher compared to the proliferative phase,
corresponding to the dominant status of estrogen in the proliferative
phase, and there was no significant change of SIRT7 staining
(Fig. [251]7K, L). These provided clues that the AcK260-PTEN could be
fluctuated during the menstrual cycle, responding to the estrogen
regulation.
In summary, these findings indicate that estrogen could act as the
upstream regulator of SIRT7-PTEN axis during menstrual cycle in
physiological state. In endometrial cancer, when SIRT7 is
hyper-expressed, prolonged estrogen exposure leads to the exacerbation
of PTEN-K260 deacetylation by SIRT7. The significant low acetylation
level at K260 results in extensive ubiquitination of PTEN mediated by
NEDD4L, subsequently leading to PTEN degradation. Low PTEN protein
level prompt the migratory and invasive potential of endometrial cancer
and result in poor prognosis of the tumor (Fig. [252]7M).
Discussion
Cellular protein level of tumor suppressor PTEN is tightly correlated
with the progression of endometrial cancer. Current studies have
pointed out that PTEN protein loss is related to its dysregulation of
post-translational modification, such as ubiquitination and
phosphorylation^[253]72–[254]75. In this study, we demonstrate
deacetylation of PTEN at K260 is responsible for promoting its
subsequent ubiquitination and degradation. In endometrial cancer, we
identify SIRT7, one of the mammalians sirtuins, as the major regulator
of such deacetylation-ubiquitination axis of PTEN. Anomalous SIRT7
overexpression is associated with PTEN loss, tumor migration and
invasion in endometrial cancer.
Increased expression and a potential oncogenic role of SIRT7 in tumor
is reported in various cancer types. For instance, SIRT7 is
overexpressed in prostate cancer and SIRT7 depletion inhibits cell
proliferation by the androgen receptor signal pathway^[255]37; SIRT7
exhibits higher expression and is associated with a poor prognosis,
promoting tumor development in hepatocellular
carcinoma^[256]76–[257]78; SIRT7 upregulation exhibits an oncogenic
property and could serve as a prognostic factor in colorectal
cancer^[258]39; Increased SIRT7 expression promotes thyroid
tumorigenesis through activation of AKT and p70S6K1^[259]38. On the
other hand, some studies have also manifested that SIRT7 achieves
anti-tumor activities under other contexts. SIRT7 hyperphosphorylation
inhibits tumor progression by preventing K63-linked AKT
polyubiquitination and activation in breast cancer^[260]79, and SIRT7
is also reported to suppress breast cancer migration via TGF-β
signaling^[261]32. In endometrial cancer, data from TCGA indicates that
SIRT7 is upregulated in EC and SIRT7 high expression is correlated to
poor prognosis. Here, in our study, we have shown that SIRT7 is
overexpressed in endometrial cancer both in human tissues and two
established transgenic EC mouse models, presenting prominent evidence
of the potential tumor-facilitating role of SIRT7 in endometrial
cancer.
Mutant PTEN is an oncogenic driver for EC initiation. However, as to
PTEN wild type and intact genetic context, people have little idea
about how this subtype of EC tumor develops and what is the specific
context-dependent role of PTEN. Post-translational modifications (PTMs)
of PTEN are critical for controlling its activity, abundance, and
subcellular localizations, so PTEN modulating pattern regulated by PTMs
presents great complexity and has received substantial concern in
recent years. For example, ubiquitin ligase MARCH8 promotes the
hepatocellular carcinoma progression through modulating PTEN
ubiquitination and degradation^[262]80. Phosphorylation of PTEN on
tyrosine 240 by FGFR2 could promote DNA repair in glioblastoma, and is
tightly associated with the therapeutic resistance to ionizing
radiation^[263]81. Ge et al.^[264]82 identified FBXO22 involved in
selective nuclear PTEN ubiquitination and degradation in colon cancer,
and loss of nuclear PTEN is related to comparatively more advanced
carcinoma^[265]82–[266]84. Here, we have identified the NAD^+-dependent
deacetylase SIRT7 as an upstream negative modulator of PTEN in EC and
SIRT7 mediated PTEN deacetylation leads to PTEN abatement and tumor
progression. Our work provides a missing piece of the function of
post-translational modifications of PTEN and uncovers the critical role
of SIRT7 in PTEN regulation, especially in endometrial cancer with PTEN
wild type context. The conception of “tumor suppressor reactivation” of
PTEN was proposed in prostate cancer years ago and it is suggested that
PTEN reactivation by targeting its critical modifying enzyme could be
sufficient to mitigate PTEN down-regulated tumor process^[267]85. It
was also reported that down-regulation of PTEN expression at protein
level is tightly associated with poor progression in advanced
EC^[268]18. Thus, our findings provided an insight of treatment for EC
patients that targeting SIRT7 might be able to restore PTEN expression
and restrain the EC advancement.
Previous studies have suggested that acetylation level of PTEN is
responsible of adjusting its affinity to proteins and membrane, as well
as regulating its enzymatic function directly. In detail, acetylation
at K125/K128 of PTEN by PCAF negatively regulates PTEN lipid
phosphatase activity^[269]86. Deacetylation at K163 by HDAC6 is
relevant to PTEN membrane affinity^[270]87. Acetylation level on K402
affects PTEN interaction with PDZ-domain-containing proteins, which was
reported to be regulated mainly by CBP and SIRT1^[271]88. Our study
identified SIRT7 as a deacetylase of PTEN, deacetylating PTEN at lysine
260. We also showed the acetylation level regulated by SIRT7 was
tightly associated with the ubiquitination and stability of PTEN.
Lysine 260 locates in the C2 domain of PTEN, which is the functional
domain responsible for the association with membrane, lipid vehicles,
and C terminal tail^[272]89–[273]91. Nguyen et al. demonstrated that K
to A mutations at lysine in the CBR3 loop in C2 domain would decrease
PTEN membrane localization in Dictyostelium cells^[274]92.
Nevertheless, the role of the acetylation level of K260 remains vague.
In our study, we have shown that PTEN with a higher acetylation level
at K260, mimicked by K260Q, is less ubiquitinated and remains more
stable. Meanwhile, further studies are needed to explore whether other
lysine sites in this region are acetylated and what are the
corresponding functions. In addition, whether the deacetylation of
PTEN-K260 will influence its conformation is also uninvestigated,
considering that the intramolecular interplay between C2 domain and C
terminal tail of PTEN is associated with its stability and
accessibility to ubiquitination^[275]93.
PTEN inhibits tumor migration in AKT-dependent or independent pathways.
It is well-established that PTEN could limit cancer metastasis by
PI3K-AKT pathway^[276]61,[277]94, as well as regulate cell polarity and
chemotaxis by sustaining a PIP2-PIP3 gradient^[278]95,[279]96. As a
dual phosphatase, PTEN could also alter cell adhesion and focal
adhesion process by targeting FAK^[280]59,[281]97. Recent studies have
addressed the novel gatekeeper roles of PTEN in tumor metastasis. Jiang
et al.^[282]98 have reported that loss of PTEN enhances
cholangiocarcinoma cell migration by disrupting TFEB-regulated lysosome
formation. Chang et al.^[283]99 illustrated that PTEN knockdown could
favor the invasiveness of pancreatic neuroendocrine tumors via
DUSP19-mediated VEGFR3 dephosphorylation. In addition, another recent
work has pointed out that PTEN loss expedite cell migration through
AMPK activity^[284]100. In our study, we have established that
SIRT7-regulated PTEN downregulation leads to more aggressive EC
migration both in EC cells and mouse models. However, our experiment
showed that phosphorylation level of AKT at Thr308 and Ser473 are not
significantly altered by SIRT7 knockdown (Supplementary Fig. [285]4l).
We supposed that it is due to SIRT7’s sophisticated role in PI3K-AKT
signaling, which could explain that cell proliferation is not
distinctly inhibited, under SIRT7 mediated PTEN downregulation
condition, in HEC-1B cells (Supplementary Fig. [286]2b–g). It has been
reported that SIRT7 favors AKT dephosphorylation by targeting FKBP51
and depletion of SIRT7 significantly increases AKT activity in
mice^[287]101. On the other hand, SIRT7 promotes AKT phosphorylation
and activation in a DBC1/SIRT1 dependent manner in thyroid
cancer^[288]38. Nevertheless, our results, collectively, hint that
SIRT7-PTEN axis presents the major effect on tumor migration in EC with
an AKT-independent manner.
Continuous exposure to estrogen and dysregulation of the hormone
pathways is one of the major risk factors of endometrial
cancer^[289]102. It is well biologically and etiologically demonstrated
that rising and falling of estrogen levels across the menstrual cycle
is one of the key factor dominating series of alterations in menstrual
cycle, including the proliferation, shedding and remodeling of
endometrium^[290]68. The upregulation of estrogen helps renovate the
functional layer of the endometrium. Our data have suggested that
estrogen stimulation augments the interaction between SIRT7 and PTEN,
and downregulates PTEN expression in a SIRT7-dependent manner, which
implies that estrogen could be the upstream switch of SIRT7-PTEN axis.
Intriguingly, IHC results unraveled that PTEN-K260 level was altered
along menstrual cycles, indicating a biological function of
SIRT7-mediated PTEN deacetylation during menstrual cycles. Though to
further ascertain whether estrogen-SIRT7-PTEN axis exert any functions
in menstrual cycle, more examinations focused on it are urged.
In summary, our study demonstrates that SIRT7 mediates deacetylation of
PTEN and diminishes its expression, unveiling a mechanism of SIRT7
involved in PTEN deacetylation-ubiquitination pathway. We show that
SIRT7 facilitates the EC metastasis by promoting PTEN degradation in a
deacetylation-dependent fashion, which provides a insight to target
SIRT7, subsequently restore PTEN level, as potential therapeutic
strategy, for patients with advanced endometrial cancer.
Methods
Ethical statement
All animal studies were conducted under the NIH Guide for the Care and
Use of Laboratory Animals and approved by Tongji University School of
Medicine Animal Care and Use Committee (Approval Number: TJBG02524101).
The immunohistochemistry and western blot studies of human EC tissue
were approved by Ethics Committee of Shanghai First Maternity and
Infant Hospital (Approval Number: KS22356). Clinical samples were
collected from Shanghai First Maternity and Infant Hospital, with
informed consent obtained from each participant.
Cell lines and cell culture
The human EC cell line HEC-1B (Cat: TCHu115) and human embryonic kidney
cell line HEK293FT (Cat: SCSP-5212) were purchased from Chinese
National Collection of Authenticated Cell Cultures. Human EC cell lines
Ishikawa (Cat: FH0305) and KLE (Cat: FH0304) were purchased from
Shanghai Fuheng Biotechnology and certificated by STR analysis. ISK,
HEC-1B cells were cultured in DMEM/F12 medium (BasalMedia, #L310KJ).
KLE and HEK293FT cells were cultured in DMEM medium (BasalMedia,
#L110KJ). The medium was supplemented with 10% FBS (Gibco, #16000-044)
and 1% penicillin/streptomycin (Gibco, #15140-122). All cells were
cultured in a 37 °C incubator with 5% CO2 and mycoplasma test was
regularly conducted.
Transgenic mouse models
The Sirt7^f/f mouse was a gift from professor Baohua Liu of Shenzhen
University Health Science Center, Shenzhen, China. The Lkb1^f/f
(C57BL/6J-Stk11^em1(flox)Cya) mouse was purchased from Cyagen,
Guangzhou, China. The Pten^f/f mouse (B6.129S4-Pten^tm1Hwu/J, #006440)
and Pgr-cre (B6.129S(Cg)-Pgr^tm1.1(cre)Shah/AndJ, #017915) mouse were
purchased from JAX lab. Sirt7^f/f, Lkb1 ^f/f or Pten^f/f mice were
crossed with Pgr-cre mice to get the uterus conditional KO mice of
indicated genes and the mice were sacrificed for analysis at indicated
birth weeks. All mice were housed under specific-pathogen-free
conditions and with a 12-hour light/12-hour dark cycle. The ambient
temperature was kept at 22 °C ± 2 °C with a 45% humidity. All mice were
fed with standard Irradiated Diet (Jiangsu Xietong Company, #XTI01ZJ).
Animal studies were conducted under the NIH Guide for the Care and Use
of Laboratory Animals. In all studies, the tumor burden did not exceed
10% of body weight and the maximal tumor size was less than 2 cm³, as
permitted by Tongji University School of Medicine Animal Care and Use
Committee.
Plasmids, antibodies, and reagents
The information of antibodies is provided in Supplementary
Table [291]1.
The SIRT7 plasmids were gifts from Professor Wei-Guo Zhu of Shenzhen
University Health Science Center; The HA-PTEN plasmids were purchased
from Addgene (Cat. #78776) and the CDS was cloned into a pEGFP vector
to get the GFP-HA-PTEN plasmids. The FLAG-PTEN plasmids in a pQCXIN
vector were a gift from Professor Shao-Ming Shen of Shanghai Jiao Tong
University School of Medicine; The mutant and WT PTEN plasmids were
generated based on a PLVX-IRES-Puro backbone and mutants were generated
by the site-specific mutagenesis method (Cat. #FM111-02, Transgen)
following sequencing confirmation. The GST-tagged plasmids were cloned
into a pGEX4T1 vector. The FLIG-NEDD4L plasmids were provided by
professor Jian-Wei Zhou from Nanjing Medical University.
shRNA expressing plasmids was constructed in a pLKO.1 vector and
sequences are as follows:
shSIRT7-1: CCCTGAAGCTACATGGGAA,
shSIRT7-2: AGCCATTTGTCCTTGAGGAA,
shPTEN-1: CAGCATACACAAATTACAAAAGT,
shPTEN-2: TAGAGTTCTTCCACAAACAGAAC,
shNEDD4L-1: TGAGGATCATTTGTCCTAC,
shNEDD4L-2: GCTAGACTGTGGATTGAGT,
shTRIM25-1: ACAACAAGAATACACGGAAAT,
shTRIM25-2: GTGCCCGATTCCTCTTAGAGA.
The information of reagents used is as follows:
Cycloheximide (Biovision, Cat. #1041), MG132 (MCE. Cat. #HY-13259),
Chloroquine (MCE, Cat. #17589 A), β-Estradiol (Sigma, Cat. #E2758).
Cell transfection and RNA interference
HEK293FT cells were transfected with plasmids using PEI (Polysciences,
USA). For endometrial cancer cells, the transient transfection of
plasmids and siRNA was conducted with the EN130 program in a Lonza 4D
instrument (Lonza, Cologne, Germany). Lentivirus infection was used to
obtain stable cell line. 6 μg lentiviral constructions, 4 μg pSPAX2,
2 μg pMD2G were co-transfected into HEK293FT cells. 72 hours after
transfection, the supernatants were collected, and filtered through
0.45μm membrane (Millipore, USA) for cell infection. Puromycin in
appropriate concentration was used for cell selection.
siRNAs sequences used are as follows:
SIRT7 siRNA 1: GGGAGTACGTGCGGGTGTT;
SIRT7 siRNA 2: CCCTGAAGCTACATGGGAA.
CRISPR-Cas9-based gene editing
LentiCRISPRv2 vectors were used in CRISPR-Cas9-based gene editing. The
plasmids were constructed as described^[292]103,[293]104.
gRNA sequences used are as follows:
sgSIRT7-1: GAGCCGCTCCGAGCGCAAAG,
sgSIRT7-2: GCGTCTATCCCAGACTACCG,
sgSIRT7-3: CGAGAGCGCGGACCTGGTAA.
RNA extraction and real-time quantitative PCR
Total RNA was extracted with kit (TIANGEN, #DP419) according to the
instructions. mRNA was reverse transcribed into cDNA with Transgen
EasyScript® One-Step gDNA Removal and cDNA Synthesis SuperMix
(Transgen, #AE311-03). RT-qPCR was performed with SYBR Green qPCR Mix
(Roche, #04913914001). The primers used are as follows:
PTEN-F: TGGATTCGACTTAGACTTGACCT,
PTEN-R: GGTGGGTTATGGTCTTCAAAAGG,
SIRT7-F: ATGAGCAGAAGCTGGTGC,
SIRT7-R: CTGTCTGGTGTCTGTGGA.
RNA sequencing and GSEA analysis
RNA Seq was performed by Majorbio, Shanghai, China. Significant
differentially expressed genes were analyzed by DEseq2 according with a
significance level P adjust <0.05, |log[2]FC | >=1. Data from
RNA-Sequencing were subjected to GSEA. GSEA was performed by the GSEA
v4.2.3 program comparing shRFP vs shSIRT7 HEC-1B cells. The gene sets
were downloaded from MSigDB.
TCGA data analysis
The transcriptome, simple nucleotide variation and clinical data of
TCGA-UCEC was downloaded from the TCGA websites. Transcript per million
was used to present SIRT7 expression. For the survival rate of PTEN
mutant/wild-type UCEC patients (Fig. [294]6), samples were classified
based on PTEN mutation status, and then were classified as either SIRT7
high or SIRT7 low based on SIRT7 expression level using a 33% cutoff.
Kaplan-Meier overall survival analysis was conducted by survival (v
3.5.8) and survminer (v 0.4.9) packages in R studio. The P value was
calculated using the two-sided log-rank test.
The survival rate of TCGA-UCEC dataset in Supplementary Fig. [295]1c, d
was analyzed by GEPIA 2.0^[296]105
([297]http://gepia2.cancer-pku.cn/#survival) with a 50% cutoff. P value
was calculated by log-rank method. Overall survival rate was analyzed.
Mass spectrum
The mass spectrum analysis was conducted at the School of Life
Sciences, Fudan University, Shanghai, China. For sample preparation,
FLAG-PTEN plasmids were overexpressed in HEK293FT cells. No control
sample was analyzed. One protein sample (number of replicates =1) was
prepared after enrichment with anti-FLAG antibodies and subjected to
SDS-PAGE. Gel bands of interest were excised and subjected to 10 mM
dithiothreitol (DTT) for 30 minutes at 56 °C, followed by alkylation
with 50 mM iodoacetamide (IAA) for 45 minutes in the dark at room
temperature. Then, the gel lane was excised and treated with 5 ng/μl of
sequencing-grade modified trypsin (Promega) overnight at 37 °C for
protein digestion. The digestion was terminated by adding 10% formic
acid (FA). The supernatants were collected, and the peptides were
extracted using a solution of 30% acetonitrile (ACN). The resulting
peptide mixtures were dried and reconstituted in 0.1% formic acid for
mass spectrometry analysis.
For LC-MS analysis, a nanoflow EASYnLC 1200 system (Thermo Fisher
Scientific, Odense, Denmark) coupled with an Orbitrap Exploris480 mass
spectrometer (Thermo Fisher Scientific, Bremen, Germany) was employed.
A one-column system was adopted for all analyses. Samples were analyzed
on a home-made C18 analytical column (75 µm i.d. × 25 cm, ReproSil-Pur
120 C18-AQ, 1.9 µm (Dr. Maisch GmbH, Germany))^[298]106. The mobile
phases consisted of Solution A (0.1% formic acid) and Solution B (0.1%
formic acid in 80%ACN).The derivatized peptides were eluted using the
following gradients: 5–8% B in 2 min, 8–44% B in 38 min, 44–70% B in
8 min, 70–100% B in 2 min, 100% B for 10 min, at a flow rate of 200 nL
min. High-field asymmetric-waveform ion mobility spectrometry (FAIMS)
was enabled during data acquisition with compensation voltages set as
−40 and −60 V.MS1 data were collected in the Orbitrap (60,000
resolution). Charge states between 2 and 7 were required for MS2
analysis, and a 45 s dynamic exclusion window was used. Cycle time was
set at 1 s. MS2 scans were performed in the Orbitrap with HCD
fragmentation (isolation window 1.6; 15,000 resolutions; NCE 30%).
The data were processed with UniProt human protein database (22,045
entries, download in 02/02/2020) and the using Protein Discoverer
(version 2.4, thermo Fisher Scientific) with Mascot (version 2.7.0,
Matrix Science)^[299]107. The mass tolerances were 10 ppm for precursor
and fragment Mass Tolerance 0.05 Da. Up to two missed cleavages were
allowed. Minimum number of unique peptides for protein identification
was 1. The search engine set cysteine carbamidomethylation as a fixed
modification, and set N-acetylation in the proteins and oxidation of
methionine as variable modifications.
Cell viability and colony formation assay
Cells were seeded with indicated number in 6-well-plates and cell
number was calculated every 24 hours. For colony formation assay, 1000
cells were seeded, incubated for 2 weeks, fixed with paraformaldehyde
then stained with crystal violet. Colonies with at least 50 cells were
counted.
Wound healing assay
Endometrial cancer cells were seeded in 6-well plates at approximately
90% confluence. A wound was created at the well center using a pipette
tip, and any debris was removed with PBS washes. Cells were then
incubated in serum-free medium for the specified time period before
imaging. Image J software was used to measure migration area. Views
from three independent experiment were quantified.
Migration and invasion assay
For migration assays, 50,000 cells were seeded in the upper chamber of
a transwell plate (Cat. #3422, Corning, USA) with serum-free medium.
For invasion assays, 250,000 cells were seeded after pre-coating the
membrane with 1:10 diluted Matrigel (Cat. #356234, Corning, USA) and
allowing it to solidify. The lower chamber was filled with complete
medium containing 10% FBS. After the indicated incubation period, the
membrane was washed twice with PBS and any remaining cells on the upper
surface were removed with a cotton swab. Cells were then fixed with
paraformaldehyde and stained using crystal violet. Views from
three/four independent experiment were quantified. Image J software was
used to quantify cell number.
Luciferase assay
For promoter-firefly luciferase plasmid construction, a 2 kb DNA
segment upstream of transcription start site of PTEN was cloned into
the pGL3-Basic vector and the genomic DNA template was extracted from
HEC-1B cells. HEC-1B cells were transfected with 2 μg PTEN
promoter-firefly luciferase plasmids along with 3 ng Renilla luciferase
plasmids for normalization of the transfection efficiency. For
overexpressing situation, empty vectors or FLAG-SIRT7 plasmids were
co-transfected. Cells were collected and lysed 72 hours post
transfection. Luciferase activity was measured with a Dual-Luciferase
Reporter System (Cat. #E1910, Promega, USA) and a GloMax luminometer
(Cat. #E5311, Promega, USA).
GST pull-down
GST pull down assays were conducted as previously described^[300]108.
GST and GST-fusion proteins were induced in Escherichia coli using
0.1 mM isopropyl-β-D-thiogalactopyranoside and incubated overnight at
16 °C. These proteins were then purified with glutathione Sepharose 4B
beads (GE Healthcare, USA). The beads and GST proteins were combined in
reaction buffer [10 mM Tris-HCl (pH 8.0), 1 mM EDTA, and 100 mM NaCl]
and co-incubated at 4 °C with either His-tagged proteins isolated from
bacteria or Flag-tagged proteins extracted from cell lysates. Following
overnight incubation, the beads were washed three times using the
specified buffer and subsequently analyzed by western blotting.
Immunoprecipitation
Cells transfected with indicated plasmids were collected and lysed in
IP buffer [20 mM HEPES pH = 8.0, 0.2 mM EDTA pH = 8.0, 5% glycerol,
150 mM NaCl, 1% NP40 and protease inhibitor] on ice for 30 min, and
then the cell lysates were sonicated 10 times on ice at 35% amplitude,
1 s per time, centrifuged at 18000 g (13500 rpm) for 15 min at 4 °C.
The supernatant was afterwards incubated with indicated antibodies or
IgG overnight at 4 °C on a rotation. Same amounts protein A/G beads
(Cat. #A10001M, Abmart, China) were added and the incubation was
continued for another 2 hours. The beads were washed with IP buffer for
3 times (centrifuged at 100 g) before being boiled and subjected to
western blotting.
In vitro deacetylation assay
In vitro deacetylation assay was performed as described^[301]31.
Briefly, HEK293FT cells transfected with HA-PTEN or FLAG-SIRT7,
respectively, were collected and lysed with IP buffer. For in vitro
deacetylation reaction, HA-PTEN or FLAG-SIRT7 proteins were enriched
from the cell lysates by antibodies and protein A/G. The protein
components were incubated in the reaction buffer [100 mM KCl, 20 mM
Tris-HCl (pH 7.9), 5 mM MgCl2, 0.2 mM EDTA, 10% glycerol, 0.5 mM DTT]
at 30 °C for 1 h in the presence or absence of 5 mM NAD^+. The
acetylation of PTEN was analyzed by western blotting.
In situ Proximity Ligation Assay assay
HEC-1B cells were seeded on confocal dishes and treated for estrogen
(100 nM, 48 h) and MG132(10 μL, 6 h) as indicated. Cells were washed
with PBS and followed by fixation with 4% paraformaldehyde (PFA) for
15 minutes. Then, HEC-1B cells were washed with PBS twice and underwent
a subsequent permeabilization with 0.5% Triton X-100 for 15 minutes and
were then blocked using 3% BSA for 1 hour at room temperature.
Subsequently, cells were co-stained with anti-PTEN (mouse) and
anti-SIRT7 (rabbit) primary antibodies. Proximity ligation assay (PLA)
staining was conducted utilizing the Duolink In Situ Red Starter Kit
(mouse/rabbit) in strict accordance with the instructions provided. In
short, cells were stained with Duolink In Situ PLA Probe Anti-Rabbit
PLUS and Duolink In Situ PLA Probe Anti-Mouse MINUS for 1 hour at
37 °C. Following washed by wash buffer A for 2 times, samples were
treated with ligation solution added with ligase for 30 minutes at
37 °C for ligation and followed by wash. Then, the amplification
polymerase solution was applied for 100 minutes at 37 °C. At last, the
dishes were sealed with Duolink In Situ Mounting Medium containing
DAPI. Fluorescent images were captured and PLA signals were quantified
using ImageJ software.
Xenograft and metastasis mouse model
Five-week-old female BALB/c Nude mice were divided into three group
randomly for both xenograft assay and spleen injection. HEC-1B cells
were injected subcutaneously at day 0 (2
[MATH: × :MATH]
10^6 cells in 100 μl PBS). Tumor volume was recorded every two or three
days by measuring diameters, and the volume was calculated as
volume = width^2 × length/2.
The mouse liver metastasis model was established as previously
described^[302]53. In brief, the female BALB/c Nude mice at 6-week post
birth and similar body weight were grouped randomly and anesthetized by
Avertin (0.1 ml/10 g) before injection. A superficial incision about 1
centimeter was made at the middle left of abdomen to expose the spleen.
HEC-1B cells were injected gently into the splenic capsule at the
middle of the spleen (2
[MATH: × :MATH]
10^5 cells in 50 μl PBS), and after that, the injection sites were
pressed for 30 s to avoid the leakage and intraperitoneal
dissemination. After the incision was closed, the mice were warmed on a
warming plate with close attention until they were awake. The mice were
sacrificed 8 weeks post inoculation for measurements of the tumor
nodules on the liver.
Orthotopic injection mouse model
Female nude mice in 8 weeks old were applied for orthotopic injection.
The orthotopic injection was conducted as previously
described^[303]55,[304]56 with some modifications. The nude mice were
divided randomly into different groups and anesthetized by Avertin
(0.1 ml/10 g) before injection. a 0.5–1 cm incision was made in the
lower abdomen for exposure of the uterine horn. The distal portion of
the uterine horn was identified and pulled out. HEC-1B cells (2
[MATH: × :MATH]
10^6 cells in 25 μl PBS) were injected into the lumen of uterine horn.
The uterine was carefully returned to the abdomen and the injection
site was closely monitored, ensuring no visible spillage taken place
and then the incision was closed with staples. Mice were monitored
every day and were sacrificed at 4 weeks after injection or they
exhibited signs of severe distress or morbidity, including significant
weight loss (>20% of initial body weight), severe lethargy, hunched
posture, restricted mobility, or severe ascites.
Immunohistochemistry
This study was approved by Ethics Committee of Shanghai First Maternity
and Infant Hospital. Clinical samples were collected from Shanghai
First Maternity and Infant Hospital, with informed consent obtained
from each participant. None of the patients had received radiotherapy,
endocrine therapy, or chemotherapy before the surgery. The IHC staining
were performed using methods described in previous study^[305]109.
Briefly, tissue slides were deparaffinized, rehydrated, and treated
with 3% H2O2 to quench endogenous peroxidase activity. This was
followed by antigen retrieval and BSA blocking. Slides were then
incubated with diluted primary antibodies overnight at 4 °C.
Subsequently, the DAB detection kit (Cat. #KIT9710, MXB
Biotechnologies, Fuzhou, China) was utilized. The stained images were
scored by the intensity (1: low; 2: weak; 3, moderate; 4, strong) and
the percentage of positive cells (1: 0–25%; 2: 26–50%; 3: 51–75%; 4:
76–100%) and under a double-blind protocol. The final
score = intensity × percentage.
Generation of polyclonal rabbit anti-human AcK260-PTEN antibody
The polyclonal antibody specific for the acetylated PTEN at K260 was
produced by PTM Biolab, Hangzhou, China. Rabbits were immunized with
the acetylated PTEN-K260 peptide (EFFH-(Acetyl)K-QNKMLKKD C) as well as
a non-acetylated peptide (EFFH-K-QNKMLKKD C). Prior to immunization,
both peptides were synthesized and analyzed using MS analysis. The sera
obtained from immunized rabbits was then filtered with the unmodified
peptide, followed by affinity purification using the corresponding
acetyl-peptide.
Statistics and reproducibility
Statistical significances were carried out in Prism 8 by Student’s two
tailed t test for comparison between two groups, Pearson’s χ2 test for
correlation analysis, and two-way ANOVA followed by Bonferroni’s
multiple comparisons test for comparison among groups at multiple time
points. Error bars indicate s.e.m. or s.d. as indicated in figure
legends. Sample sizes were estimated based on our experience with
similar experiments and reference to previously published studies. For
RNA-seq data, three biological replicates were set for each group. For
animal and clinical sample experiments, more than 5 biological
replicates were used. For Western blot assays, experiments were
repeated three times independently with similar results and
representative images were given.
Reporting summary
Further information on research design is available in the [306]Nature
Portfolio Reporting Summary linked to this article.
Supplementary information
[307]Supplementary Information^ (13.4MB, pdf)
[308]Reporting Summary^ (250KB, pdf)
[309]Transparent Peer Review file^ (4.1MB, pdf)
Source data
[310]Source Data^ (6.7MB, zip)
Acknowledgements