Autophagy inhibition perturbs ERBB2 trafficking and abolishes tumorigenesis in ERBB2-driven breast cancer

ABSTRACT

Macroautophagy/autophagy modulation is increasingly recognized as a potential strategy for cancer therapy. Using a recently developed Rb1cc1 mutant knockin mice model, we have taken a rigorous genetic approach to assess the role of both its autophagy and non-canonical functions in an ERBB2-driven BrCA model. We found that autophagy abrogation virtually abolishes mammary tumorigenesis in the ERBB2-driven model, exhibiting stronger inhibitory effects than in our previous studies using PyMT and brca1-null mouse models. Mechanistically, autophagy inhibition perturbs ERBB2 intracellular trafficking and triggers its release via small extracellular vesicles. Our results demonstrate a new mechanism for autophagy to promote tumorigenesis in ERBB2-driven BrCA and could supplement current strategies for anti-ERBB2 therapy.

Amplification of ERBB2 accounts for approximately 25% of human BrCA. Although anti-ERBB2 therapy has significantly improved outcomes for those patients, resistance to ERBB2-targeted therapies still occurs in some patients as a result of ERBB2 mutation or aberrant activation of downstream pathways of ERBB2. Thus, further understanding of cellular pathways that regulate ERBB2 in mammary tumor cells will potentially yield new targeted therapies for the deadly ERBB2-enriched BrCA.

Autophagy plays both tumor-suppressive and tumor-promoting functions in various cancer models, likely due to its impact on many cellular functions in different contexts. It is also increasingly recognized that many autophagy genes possess both autophagy and non-canonical functions. Therefore, defects observed in the various mouse knockout models could be caused by the loss of their non-canonical functions or the combination of both autophagy and non-canonical functions. We took a rigorous genetic approach to investigate the effect of blocking autophagy in ERBB2-enriched BrCA by either genetically ablating the essential autophagy gene Rb1cc1/Fip200 (i.e., cKO) or genetically disrupted its autophagy function (i.e., cKI) in the MMTV-Neu mouse model [1]. We found that both Rb1cc1 cKO and cKI virtually abolish mammary tumorigenesis in the ERBB2-driven model, formally establishing that this effect is due to the loss of RB1CC1’s autophagy function.

In order to understand the underlying mechanisms for this striking observation (virtually blocking tumor development vs our previous results showing delayed tumorigenesis in PyMT and brca1-null mouse models upon rb1cc1 deletion), we examined the expression of ERBB2 and AKT activity, which are the driving forces for tumorigenesis in this model. Unexpectedly, ERBB2 expression and AKT activity are dramatically decreased in pre-malignant lesions and tumors upon blocking autophagy. Consistently, similar results are also observed in in vitro assays by deleting Rb1cc1 or other autophagy genes such as Atg13 and Atg5, further ruling out the autophagy-independent functions of Rb1cc1 in maintaining ERBB2 levels. Like other membranous tyrosine kinases, ERBB2 is mainly localized in the cell plasma membrane, allowing for its sustained downstream signaling in tumor cells. Surprisingly, ERBB2 is significantly reduced in the plasma membranes and displays diffused localization upon blocking autophagy.

Because endocytosis is a key route for the regulation of cell surface levels of ERBB family members, we initially investigated this aspect of trafficking. Indeed, we found a majority of intracellular ERBB2 accumulated in the intralumenal vesicles (ILVs) of early endosomes and multivesicular bodies (MVBs) upon autophagy blockade. However, blocking autophagy fails to change the endocytosis of ERBB2 in tumor cells. This result prompted us to investigate another route for the perturbed ERBB2 trafficking triggered by autophagy blockade. The Golgi apparatus is an alternative source for the delivery of cargos to endosomes. As such, we investigated ERBB2 localization in the Golgi apparatus and detected the intracellular accumulation of ERBB2 in this organelle in rb1cc1 and atg13 knockout cells, suggesting autophagy blockade diverts ERBB2 trafficking from the Golgi to ILVs rather than to the plasma membrane.

We next tested whether the ERBB2 proteins trapped in ILVs were eventually secreted from cells. Intriguingly, autophagy abrogation substantially increases the enrichment of ERBB2 in sEVs, and the export of sEVs from tumor cells. Most importantly, impeding the release of sEVs by silencing Rab27a nearly recovers the total amount of ERBB2 levels in both rb1cc1 and atg13 knockout cells. These results suggest that increased ERBB2 release through sEVs leads to reduced ERBB2 expression upon autophagy inhibition. Because MVBs can fuse with not only the plasma membrane for secretion but also lysosomes for degradation, these findings also suggest that blocking autophagy can control the trafficking of a subtype of MVBs (i.e., ERBB2-enriched) to preferentially fuse with plasma membrane rather than lysosomes. Future studies will be necessary to elucidate the mechanism by which inhibiting autophagy retains ERBB2 in the Golgi and diverts its trafficking to ILVs. It will also be important to examine the changes in sEV composition upon autophagy inhibition and their impact on metastasis.

The amount of sEV release can be regulated at multiple steps including biogenesis of ILVs, transport of MVBs and membrane fusion. Another finding of our studies is the observation of increased amount of ILVs formed in early endosomes in rb1cc1 and atg13 knockout cells. This indicates that autophagy may also regulate ILVs biogenesis in early endosomes. It is of interest to determine whether autophagy could affect other steps related to sEV release. Thus, our studies raised multiple interesting questions regarding the effect of autophagy flux on biogenesis and heterogeneity of ILVs as well as trafficking of different types of MVBs.

Most documented studies in cancer focus on the classic functions of sEVs, which deliver bioactive molecules to recipient cells in tumor microenvironments and distal organs for improving cancer progression. Our study raises an alternative notion that sEV release could also suppress tumor cells autonomously by expelling oncogenic drivers like ERBB2 to reduce downstream signaling in vivo. Further studies are needed to clarify whether autophagy inhibition could impair the trafficking of other cell surface kinase receptors for release. Aberrant amplification of ERBB2 is also frequently observed in other types of cancer such as lung, gastric, and prostate cancers. Therefore, it is possible that our findings may have far-reaching implications beyond breast cancer and could guide clinical applications to reap the full benefit of targeting autophagy for the treatment of breast and other cancers.

Funding Statement

This work was supported by the National Institutes of Health [CA211066].

Disclosure statement

No potential conflict of interest was reported by the author(s).