1
Tribbles 2 confers enzalutamide resistance in prostate cancer by promoting lineage plasticity
Authors: Jitender Monga
1
, Indra Adrianto
2
, Craig Rogers
1,4
, Shirish Gadgeel
4
, Dhananjay Chitale
3,4,
, Joshi J.
Alumkal
5
, Himisha Beltran
6
, Amina Zoubeidi
7
, and Jagadananda Ghosh
1,4,*
Affiliations:
1
Vattikuti Urology Institute, Henry Ford Health System, Detroit, MI 48202, USA.
2
Public Health Sciences, Henry Ford Health System, Detroit, MI 48202, USA.
3
Department of Pathology, Henry Ford Health System, Detroit, MI 48202, USA.
4
Henry Ford Cancer Institute, Henry Ford Health System, Detroit, MI 48202, USA.
5
Univeristy of Michigan Rogel Cancer Center, Ann Arbor, MI 48109, USA.
6
Dana Farber Cancer Institute and Harvard Medical School, Boston, MA 02115, USA.
7
Vancouver Prostate Center, Vancouver, BC, Canada.
*Correspondence to: jghosh1@hfhs.org
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted March 28, 2021. ; https://doi.org/10.1101/2021.03.26.437250doi: bioRxiv preprint
2
Abstract: Second-generation anti-androgen, such as enzalutamide (Xtandi), is commonly prescribed for prostate
cancer therapy, but enzalutamide-resistant, lethally incurable disease invariably develops. To understand the
molecular basis of enzalutamide resistance, we comprehensively analyzed prostate tumors and clinically relevant
models. These studies revealed that enzalutamide resistant prostate cancer cells overexpress Tribbles 2 (Trib2), a
pseudokinase. Expression of Trib2 is negatively regulated by androgen receptor signaling. Overexpression of
Trib2 makes prostate cancer cells completely resistant to clinically relevant doses of enzalutamide. Trib2
downregulates expression of luminal markers but upregulates the neuronal transcription factor, BRN2, and the
stemness factor, SOX2, to induce neuroendocrine differentiation. Our findings indicate that Trib2 confers
resistance to enzalutamide therapy via a mechanism involving increased cellular plasticity and lineage switching.
Main Text:
Common forms of prostate cancer cells bear luminal characteristics and depend on androgenic signaling for
growth. However, it has been realized that men with prostate cancer who were treated with anti-androgenic
therapies, frequently develop aggressive and deadly forms of prostate cancer which are no longer responsive to
primary therapies. Enzalutamide, an inhibitor of androgen receptor function, is commonly prescribed to treat
prostate cancer, but resistant prostate cancer eventually develops which grow aggressively, leading to widespread
metastatic disease and bring demise to prostate cancer patients (1-3). Based on current assessment, the
enzalutamide-resistant type of aggressive prostate cancer is responsible for most of the morbidity and mortality
associated with prostate cancer and ~30,000 lives of American men are lost every year (4). However, lack of
proper understanding about critical molecular targets in hormone refractory disease settings, such as enzalutamide
resistance, largely contributes to the loss of battle against majority of prostate cancer deaths.
Several reports encompassing the involvement of both androgen receptor reactivation or bypass, as well as
androgen-independent signaling, have been forwarded to explain the mechanism of enzalutamide resistance.
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted March 28, 2021. ; https://doi.org/10.1101/2021.03.26.437250doi: bioRxiv preprint
3
However, analysis of multiple cell lines and in vivo models, which were used to explore the molecular basis, have
ended up with identification of subtypes of cells (5-7). Current molecular understanding suggests that in addition
to AR reactivation by mutation or splice variants, manifestation of enzalutamide resistance can be the result of
overgrowth of cells that are developed in the tumor by lineage switching which is triggered by drug-induced
repression or loss of the AR-signaling (8-11). About 10-20% of enzalutamide-resistant prostate cancer show
neuroendocrine (NE) features and no effective treatments are available for this type of aggressive and highly
invasive prostate cancer (12-15). Though the continued growth of the heavily enzalutamide-treated prostate
tumors can be driven by non-androgenic signaling, molecular underpinnings of critical targetable mechanisms in
treatment-emergent NE cells are yet to be fully characterized.
When prostate cancer cells become resistant to strong androgen receptor blocker, such as enzalutamide, their
common characteristics change from slow-growing, non-invasive type to fast-growing, highly invasive cells, but
identity of molecular drivers for their rapid growth and resistance to enzalutamide is still limited. To better
understand the mechanistic basis behind enzalutamide resistance, we developed an in vitro model by chronically
treating human LNCaP and MDA-PCA-2B prostate cancer cells (AR-signaling intact) with gradually increasing
doses of enzalutamide (up to 30µM) for >12 weeks to mimic the clinical conditions in long-term enzalutamide
therapy (16). Resultant cells (LNCaP-ENR and PCA-2B-ENR) are completely resistant to clinically relevant
doses of enzalutamide, the blood level of which may go up to ~34µM in average (17, 18). An unbiased,
comprehensive gene expression analysis revealed that the Trib2 pseudokinase is grossly overexpressed in
enzalutamide-resistant prostate cancer cells, compared to parental enzalutamide-sensitive cells (Fig. 1a-d).
To confirm the array data, we detected Trib2 expression by RT-PCR and Western blot which showed strong
upregulation of Trib2 mRNA as well as protein levels in enzalutamide-resistant (LNCaP-ENR and PCA-2B-
ENR) cells (Fig. 1e, f). Overexpression of Trib2 is correlated with activation of the canonical Akt signaling
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted March 28, 2021. ; https://doi.org/10.1101/2021.03.26.437250doi: bioRxiv preprint
4
module showing increased phosphorylation of Akt (pSer-473) and increased protein level of Bcl-xL which are
standard markers that promote cell survival and decrease apoptosis. To verify whether the in vitro observation of
Trib2 overexpression is valid in vivo, we analyzed prostate tumors in tissue microarrays and found that prostate
PDX tumors overexpress Trib2 when the mice were treated with enzalutamide at 30mg/kg/day for 6 weeks (Fig.
1g, h). Moreover, we found that a vast majority of the clinically advanced metastatic prostate tumors from
enzalutamide-treated patients show a robust increase in the expression of Trib2 proteins (Fig. 1i, j). Altogether,
these findings suggest that triggering of Trib2 overexpression is a fundamental mechanism in prostate cancer cells
both in vitro and in vivo when treated with enzalutamide, a second-generation direct inhibitor of AR activity.
Prostate cancer cells variably respond to androgenic signaling for proliferation, and the AR-signaling status varies
widely among established prostate cancer cell lines. Additionally, some prostate cancer cell lines are devoid of
androgen receptor expression and are naturally independent of androgenic stimulus. This provided us with an
opportunity to find whether there is any correlation between AR function and the expression levels of Trib2. To
address this, we analyzed a range of prostate cancer cell lines, and found that the Trib2 protein level is low in
prostate cancer cells with high AR activity, while the level of Trib2 protein is high in prostate cancer cells where
the AR signaling function is lowered by drug treatment or is naturally low or lost due to mutation or deletion of
the AR gene (Fig. 2a, b). These findings raised a fundamental question whether a functionally negative correlation
exists between AR-mediated signaling and the expression of the Trib2 gene.
If Trib2 is negatively regulated by AR signaling, then it is expected that re-introduction and activation of AR in
AR-null cells will interfere with Trib2 expression under normal cell culture conditions. Thus, we transfected the
high Trib2-expressing, AR-negative/low cells (PC3, DU145, 42D
ENZR
) with full length human AR gene
constructs and found that the Trib2 protein level was significantly decreased when the cells were transfected to
overexpress AR (Fig. 2c). Thus, reactivation of AR signaling in AR-negative cells (which are naturally high in
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted March 28, 2021. ; https://doi.org/10.1101/2021.03.26.437250doi: bioRxiv preprint
5
Trib2 protein levels) inhibits the expression level of Trib2. This finding, together with our previous observation
of the upregulation of Trib2 by enzalutamide, suggest that the expression of Trib2 is negatively regulated by AR-
signaling.
To address this, we sought to examine the effect of AR transcriptional activity on Trib2 gene expression. We
suspected that the promoter of the Trib2 gene may have androgen-response elements (AREs) which presumably
bind with AR and in turn inhibit the promoter activity to suppress Trib2 expression. Bioinformatic analysis and
ChIP-seq data indicated presence of several AREs within ~5kb promoter region upstream of transcription start
site (TSS) of Trib2 gene (Fig. 2d, e), supporting the concept that ligand-bound AR may bind with Trib2 promoter
to modulate its expression. To further confirm the AR signaling-mediated transcriptional repression of TRIB2, we
transfected LNCaP and MDA-PCA-2B cells with Trib2 promoter-luciferase constructs and found that activation
of AR by R1881 decreased luciferase activity, while inhibition of AR-signaling by enzalutamide increased the
luciferase activity in a concentration-dependent manner (Fig. 2f-h). These findings suggest that the active AR
may interact with ARE(s) and in turn inhibits the expression of TRIB2 via a negative regulation (Fig. 2i), as
observed in some other genes, such as BRN2 and SPINK1 (19,20).
Because overexpression of Trib2 was observed both in enzalutamide-treated cells and tumors, we asked the
question whether Trib2 plays any role in these cells. No specific, target-validated inhibitor of Trib2 is
commercially available for molecular and in vivo preclinical studies. Thus, we used shRNA-mediated knockdown
and found that downregulation of Trib2 strongly inhibits the viability and colony growth of enzalutamide-resistant
cells (Fig. 3a-c). Interestingly, normal, non-cancer cells, such as human foreskin fibroblasts (HFF), which do not
express detectable Trib2 proteins, remained unaffected with Trib2 shRNA, suggesting that Trib2 plays a critical
but selective role in enzalutamide-resistant cells. Recently, it was demonstrated that the EGFR kinase inhibitor,
Afatinib (AFA), destabilizes Trib2 protein by covalent modification and primes Trib2 for degradation (21). Thus,
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprintthis version posted March 28, 2021. ; https://doi.org/10.1101/2021.03.26.437250doi: bioRxiv preprint
