A key observation of our article was that treatment with the ErbB3 ligand heregulin 1β (HRG) promoted ErbB3 nuclear translocation, whereas activation of the AR reversed this translocation in HSPC but not in CRPC cells. at HRG markedly influenced ErbB3 nuclear localization in the absence but not presence of AR protein, indicating the AR’s inhibitory effect. Conversely, in HSPC LNCaP cells, in the presence of HRG, AR knockdown (k/d) significantly upregulated nuclear ErbB3 levels, an effect that was not observed in the presence of the AR ligand dihydrotestosterone (DHT) – suggesting that while ErbB3 nuclear import is dependent on ligand-dependent AR transcriptional activity, it is likely that ErbB3 export from the nucleus is dependent on AR expression or ligand-independent AR function. Mutations in the tyrosine kinase domain of ErbB3 resulted in a small increase in cytoplasmic ErbB3 but otherwise did not affect ErbB3 localization, indicating that the kinase domain is not necessary for localization.
Although the kinase domain was not involved in ErbB3 localization, it appeared to be involved in cell viability. In CRPC 22Rv1 cells, ErbB3 overexpression increased viability, whereas mutations Y1222A, Y1289A, Y1328A in the kinase domain to annul three major sites of phosphorylation suppressed it, and treatment with HRG failed to reverse the suppression. Surprisingly, this was not true of HSPC LNCaP cells. ErbB3 is thought to be kinase dead; however, its “kinase” domain is rich in phosphorylation sites – of these, six (Y1054, Y1197, Y1222, Y1260, Y1276, and Y1289) serve as binding sites for the p85 subunit of phosphatidylinositol 3-kinase (PI3K), an important enzyme that regulates various cellular functions including proliferation, survival, differentiation, migration, and trafficking. In contrast, Y1328 is a binding site for Shc, which activates Ras, which in turn triggers the phosphorylation of the mitogen activated protein kinase (MAPK) pathway. ErbB3 has other sites (Y1199, Y1262) that also trigger the MAPK cascade via binding the adaptor protein Grb2. Hence it is likely that in LNCaP cells, the neutralization of just three sites may not have been enough to kill the kinase domain altogether, but in 22Rv1 cells it certainly affected cell viability significantly. Akt is an important downstream target of PI3K. Phosphorylation of Akt was triggered in LNCaP cells by the expression of the mutant protein in the absence of the ErbB3 ligand, which indicates the presence of a feedback loop allowing PI3K activation when these three phosphorylation sites are mutated, likely by hyperphosphorylation of the other p85PI3K binding sites. However, once the ligand to ErbB3 was introduced, the triple mutation failed to have an effect. Thus, the results did not as such show that the kinase domain had no effect on LNCaP cell viability, just that the three sites alone may control viability in 22Rv1 cells, but not in LNCaP.
One reason for this observation may be that ErbB3 can form inactive homodimers (since it is kinase dead) or propagate downstream signals via heterodimerization with other receptor tyrosine kinases (RTK), such as EGFR or ErbB2, which is an ErbB3-preferred dimerization partner. While EGFR is activated by several specific ligands, including EGF, ErbB2 does not have a known ligand. Binding of ErbB3 to EGFR or HER2 activates downstream signaling, whereas ErbB3 homodimer activity may be weaker. Indeed, the primary role of ErbB3 is considered to be a platform upon which to expand the ErbB intracellular signaling repertoire. The inability of HRG to increase viability (over control conditions) in wild-type ErbB3 transfectants in LNCaP cells may indicate an excess of ErbB3 homodimerization which precludes its heterodimerization with other activating kinases, whereas, in 22Rv1 cells, ErbB3 may be more prone to heterodimerization, leading to a higher effect on cell viability.
Although not noted in the article, several figures showed that nuclear ErbB3 was consistently observed as multiple lower molecular weight bands when compared to cytoplasmic ErbB3. It is known that newly formed (nascent) ErbB3 in the endoplasmic reticulum undergoes post-translational modification (PTMs), that includes glycosylation, resulting in receptor maturation. Mature ErbB3 is therefore of a higher molecular weight compared to the nascent form. We showed that the lower molecular form resided exclusively in the nucleus, consistent with reports of nascent ErbB3 homodimers being found in nuclei where they are unlikely to have undergone glycosylation. In contrast, cytoplasmic ErbB3 went on to maturation and glycosylation.
While several humanized monoclonal anti-ErbB3 antibodies have been developed for therapeutic use, they would target plasma-membrane bound ErbB3. Nuclear ErbB3 would be unsuited to traditional targeting because its ligand-binding and kinase domains would be sequestered within or around the nucleus, rendering antibody- and kinase-domain-based inhibitory methods ineffective. An emerging method called Hydrophobic Tagging (HyT, essentially a method of small-molecule-induced intracellular protein degradation which consists of a bifunctional molecule, with one end binding to a protein of interest and the other end engaging quality control mechanisms through a hydrophobic adamantyl moiety) may be particularly suitable for ErbB3, widely-considered an ‘undruggable’ target. In fact, this method is also being investigated in the context of AR inhibition. A second method is called PROTACS, which utilizes bifunctional, small-molecule E3 ligase ligands to induce proximity-induced ubiquitination of a target protein, overcomes some of the limitations of the HyT technology, and is closer to clinical translation. This method may be suitable for ErbB3 because it is a target of the E3 ubiquitin ligases Nrdp1 and EBP1.
Overall, we proposed that the presence of nuclear ErbB3 indicated the start of a global increase in ErbB3 protein, initiated by AR inhibition. Future studies will attempt to elucidate whether physical binding of AR and ErbB3 occurs during their inverse regulation, what the nature of ErbB3 interactions with AR splice variants entail, and whether changes in HRG-mediated PSA production can be used as a therapeutic marker for AR activity and tumor cell growth and proliferation, given the observed differences of this ligand upon PSA expression in LNCaP cells (one possibility is the influence of passage number dependence of LNCaP cells). If ErbB3 itself is to be therapeutically targeted, then its inhibition would be most effective once cytoplasmic ErbB3 levels have begun to rise - most likely in response to AR inhibitory therapy including enzalutamide and abiraterone acetate - rather than waiting until its levels have peaked at which point the tumor may have evolved multiple molecular mechanisms of resistance.
Written by: Maitreyee K. Jathal,1,2 Maria Mudryj,1,2 Paramita M Ghosh1,3,4*
- Research Service, VA Northern California Health Care System, Mather, CA,
- Department of Medical Microbiology and Immunology, University of California Davis, CA,
- Department of Urologic Surgery, University of California Davis, Sacramento, CA,
- Department of Biochemistry and Molecular Medicine, University of California Davis, Sacramento, CA.
Read the Abstract