Thirty Years of HER3: From Basic Biology to Therapeutic Interventions

The HER proteins are a family of receptor tyrosine kinases that play a role in both normal and tumor cell biology. The family consists of four highly homologous members, EGFR (ERBB1/HER1), HER2 (ERBB2), HER3 (ERBB3), and HER4 (ERBB4), consisting of a ligand-binding extracellular domain, a transmembrane domain, an intracellular kinase domain, and a C-terminal tail (1). The family members (except HER2) are generally activated through extracellular ligand binding, inducing a conformational change, followed by homo- or heterodimerization among the family members, eventually leading to the activation of an intracellular signaling cascade (2). The cellular responses include increased cell survival and proliferation, explaining why aberrant EGFR family signaling is strongly connected with oncogenic events (1).

When inactive, the EGFRs exist in a monomeric tethered conformation, but upon ligand binding, the receptor changes into its extended form, exposing a dimerization arm (2, 3). Contact with another open conformation receptor permits formation of a receptor dimer, inducing a further conformational change in the intracellular domain of the receptor complex. This conformational change leads to a transphosphorylation event, where the donor receptor introduces multiple phosphorylations into the C-terminal tail of the acceptor receptor, allowing attachment and activation of downstream signaling cascade (3). The paradigm of EGFR family receptors existing solely in a monomeric form prior to ligand binding has been challenged by suggesting that EGFR can be found in an inactive dimerized form before ligand stimulus (4). The signal inactivation happens via dephosphorylation, receptor internalization, and proteolysis or recycling of the receptor (5).

EGFR binds to at least seven ligands, including EGF, TGFα, heparin-binding EGF-like growth factor, betacellulin, amphiregulin, epiregulin, and epigen (6). Neuregulins (NRGs) 1–4 are the ligands for HER3 and HER4 (7). Unlike the other three EGFR family members, HER2 has no known ligands and is found constitutively in an open conformation with exposed dimerization loop and needs a ligand-bound heterodimerization partner to signal (8). The variety in both ligands and dimerization partners provides diversity in the downstream signaling response (9).

HER3 is a unique EGFR family member with no or little intracellular tyrosine kinase activity. Compared with the other EGFR family members, HER3 diverges at critical residues in the kinase domain, locking it in an inactive-like conformation (10). Although HER3 has been reported to have some kinase activity, it is suggested to be 1,000-fold weaker than the kinase activity of the fully activated EGFR (10, 11). Because HER3 is unable to form homodimers, its activation depends on heterodimerization with another receptor to induce the downstream C-terminal phosphorylation events (12).

The HER3 gene localizes in the long arm of chromosome 12 (12q13.2), encoding a 180 kDa protein (13, 14). The extracellular domain of HER3 is divided into four subdomains (I–IV): subdomains I and III are leucine-rich β-helical areas responsible for the ligand binding, whereas subdomains II and IV are cysteine-rich regions (15). Subdomain II also contains a dimerization arm necessary for the interaction with other receptors. The transmembrane domain is followed by an intracellular domain enclosing a flexible juxta membrane region, kinase domain, and the C-terminal tail. In absence of a ligand, binding between subdomains II and IV keeps HER3 in an inactive state (16). Upon ligand binding, the dimerization partner's kinase domain transphosphorylates the tyrosine residues in the C-terminal tail of HER3 (17).

The preferable dimerization partners for HER3 are EGFR and HER2, followed by lower affinity to HER4 (Fig. 1). HER3 also dimerizes with some non-EGFR family receptors, including mesenchymal–epithelial transition (MET) factor receptor and FGFR2 (18, 19). Six of 11 HER3 tyrosine phosphorylation sites are direct recruiters of PI3K, making HER3 a strong activator for PI3K/protein kinase B (AKT) signaling, important for cancer cell survival (20–22). HER3 also activates MAPK signaling, stimulating cell proliferation. Other suggested effectors of HER3 include JAKs and activators of transcription and proto-oncogene tyrosine-protein kinase SRC signaling pathways involved in signal transduction and increased cell proliferation (23, 24).

In contrast to other EGFR family members, HER3 is not oncogenic when overexpressed alone (25). However, ubiquitous HER3 expression is detected in various cancers, including breast, ovarian, colon, gastric, lung, cutaneous, and pancreatic cancers (26, 27). High HER3 expression is also linked to disease progression and/or poor prognosis in many cancer types (28, 29).

Although HER3 does not cause tumorigenesis on its own, HER2:HER3 heterodimers possess the highest transforming capability among all the possible EGFR family dimers (30, 31). The superior oncogenic capability of the dimer pair makes HER3 critical for HER2-mediated tumorigenesis in multiple tumor types. HER3 overexpression is a frequent event, especially in HER2-positive breast cancers, and mice expressing neu (rodent HER2) transgene exhibit elevated HER3 expression (32, 33). In breast cancer cell lines, HER3 was shown to be critical for maintaining cell viability, while EGFR was dispensable (34). In addition, inhibition of HER3 revokes HER2-dependent tumorigenesis in transgenic mammary tumor models (35). HER3 has also been implicated in the pathogenesis of non–small cell lung cancer (NSCLC), many of which have EGFR mutations (36–38).

Ovarian cancers often express high levels of HER3, and HER3 expression has been associated with poor survival (39, 40). HER3 was deemed essential for ovarian cancer cell proliferation in vitro and in vivo, and the activation of HER3 was mediated by NRG1 autocrine signaling (41). NRG mRNA was detected in 83% of ovarian carcinomas, and addition of ectopic NRG1 stimulated the growth of several ovarian cancer cell lines (42). In addition, NRG-driven HER3 activation or NRG-activating fusions have been reported in multiple other tumor types, including pancreatic, head and neck, colorectal, lung, and prostate cancers (43–51), but NRG is also secreted into the tumor microenvironment by cancer-associated fibroblasts (CAF; refs. 51, 52). Furthermore, NRG inhibition suppressed tumor growth in preclinical models of pancreatic cancer (52), demonstrating the importance of NRG as a HER3-activating oncogenic factor.

Rare oncogenic HER3 mutations have been reported. Recurrent somatic HER3 mutations are found in 11% of colon and gastric cancers and are associated with malignant transformation (53). Although these mutations transform cells in a ligand-independent manner, the oncogenic activity is dependent on coexpression with HER2. Short hairpin RNA–mediated HER3 knockdown delayed tumor growth in HER3-mutant tumors. Also, in breast cancer, several HER3 mutations (F94L, G284R, D297Y, T355I, and E1261A) were shown to have gain-of-function properties (54). HER3 mutations caused increased HER2:HER3 heterodimerization and made the cells resistant to the HER2-targeting drug, lapatinib. Overall, the prognostic value of the mutations and their role in tumor development or progression are still not well understood.

HER3 expression is connected with disease progression and metastatic events in various cancer types. In breast cancer, HER3 expression is linked to increased intravasation and metastasis, and higher HER3 expression was found in metastatic breast cancer (MBC) samples compared with primary tumor samples (55). In another breast cancer study, 30% of the primary tumors were shown to express HER3, whereas the expression was 60% in the matched metastatic samples, suggesting that HER3 expression is linked to metastatic events (56). In a meta-analysis from multiple malignant tumor types, it was confirmed that HER3 expression led to worse overall survival and 1.6-fold higher death risk than in HER3-negative patients (57).

In NSCLC, HER3 mRNA expression was associated with increased metastatic rate and decreased survival (58). In a more recent study, EGFR and HER3 protein levels were analyzed from primary tumors, brain metastases, and circulating tumor cells (CTCs; ref. 27). More than 50% of the primary tumors had EGFR expression and approximately 80% expressed HER3; the numbers for brain metastases were 60% and 90%, again highlighting the importance of HER3 in disease progression. Codetection with EGFR/HER3 was successful in CTCs from the blood of 67% of the patients.

HER2 and HER3 protein levels were evaluated in patients with colorectal cancer with liver metastasis (59). Whereas high HER2 levels were found only from 8% of the primary liver tumors, high HER3 levels were found from 75% of the metastatic samples, again suggesting a role for HER3 in the metastatic process and cancer progression. HER3 is also connected to worse outcomes in pancreatic cancer (60). Furthermore, HER3 is highly expressed in cutaneous tumors and was reported to act as an indicator for poor prognosis in melanoma (61, 62). In contrast to other tumor types, high HER3 expression was associated with better survival in bladder cancers (63). Researchers suggested that this could be explained by increased expression of soluble form of HER3 (sHER3), which is overexpressed in bladder cancer (63). sHER3 is a 85 kDa truncated and secreted form of HER3 (64), which is reported as a negative regulator of HER2, HER3, and HER4 (64, 65).

Although classically HER3 signals from the membrane, occasional nuclear localization of the full-length HER3 has been reported and nuclear HER3 expression has been connected to better overall survival in uveal melanomas (66–68). Functionally, nuclear HER3 increased the mRNA-level expression of cyclin D1, suggesting it might act as a transcriptional activator (66). However, the mechanisms and role of nuclear HER3 in cancer remain to be studied further.

HER3 expression acts as a bypass mechanism for various targeted therapies, and elevated HER3 signaling confers resistance to multiple therapeutic agents.

Because HER3 dimerizes with receptors other than EGFR, including HER2 and MET receptor, HER3 can confer resistance to EGFR-targeting therapies via dimerization with partners other than EGFR (69). Early on, it was shown that HER2:HER3-mediated signaling associates with EGFR tyrosine kinase inhibitor (TKI), gefitinib, resistance in head and neck cancer and in breast cancer (70, 71). Later, it was shown that HER3 ligand, NRG1, and the following increase in HER2:HER3 dimerization confer resistance to EGFR-directed antibody, cetuximab, in colorectal cancer (72). Similarly, an increase in EGFR:HER3 dimerization was found in the majority of the residual cancer burden of patients with cetuximab/panitumumab-resistant breast cancer (73). Other anti-HER TKIs, such as osimertinib, may also induce HER3 upregulation as part of the resistance mechanism (74, 75). This NRG1-driven EGFR inhibitor resistance was reverted by using HER3-selective antibody, patritumab. Interestingly, circulating NRG1 levels were a better indicator for patritumab efficacy than HER3 mRNA expression (37). In another study, MET amplification caused gefitinib resistance via increased HER3/PI3K signaling, and MET amplification was detected in 22% of patients with lung cancer bearing tumors resistant to gefitinib or erlotinib (18).

As with EGFR inhibitors, HER3 is known to confer resistance to HER2-targeted therapies. Trastuzumab (Herceptin) is a monoclonal HER2-directed neutralizing antibody used mainly in HER2-positive breast cancer (76, 77). Although several resistance mechanisms to trastuzumab exist, bypass activation of PI3K/AKT and SRC are the major molecular mechanisms behind therapy escape, and the bypass signaling can be driven by HER3 (78, 79). It was suggested that heterotrimer formation between HER2, HER3, and insulin-like growth factor receptor 1 (IGF1R) is the major inducer of AKT- and SRC-driven trastuzumab resistance in breast cancer cells (24). In the same study, knockdown of HER3 decreased the phosphorylation activity of AKT and SRC signaling and resensitized the cells to trastuzumab, suggesting that dual blocking of HER2 and HER3 is needed to prevent the survival signaling. In addition, stimulation with NRG1 induced trastuzumab resistance in HER2-overexpressing breast cancer cells (80).

Lapatinib is a dual TKI of EGFR and HER2 used against HER2-positive MBC. Lapatinib treatment was shown to induce feedback upregulation of both mRNA and protein levels in breast cancer cell lines, and HER3 knockdown restored the drug sensitivity in lapatinib-resistant cells (81). Another study showed that lapatinib-resistant breast cancer cells were not dependent on HER2:HER3 signaling, but were relying on NRG1-driven HER3:EGFR dimerization (82). In fact, lapatinib can induce the symmetrical HER2:HER3 dimer, which may have unexpected effects on tumor cell proliferation (83).

Because HER3 is a substantial activator of PI3K/AKT survival signaling, it also confers resistance to other targeted therapies. PI3K/AKT inhibitors are known to cause feedback upregulation of HER3 by relieving AKT/FoxO-dependent suppression of HER3 (84), leading to reduced efficacy of PI3K/AKT inhibitors. NRG1 induces resistance to ALK inhibitors and BRAF-V600E inhibitor, vemurafenib (85, 86). Furthermore, transcriptional HER3 activation was connected to resistance to MAPK and RAF kinase inhibitors both in melanomas and thyroid cancer, and most recently, HER3 amplification was described as a clinical bypass mechanism for MET inhibitors in NSCLC (87–89).

HER3 expression is also connected to resistance to hormonal therapies. HER3 plays a critical role in the phosphorylation of HER2 in breast cancer cells, and downregulation of HER3 reversed antiestrogen receptor (ER) tamoxifen resistance in breast cancer cell lines (90). Furthermore, patients with breast cancer bearing tumors coexpressing HER2 and HER3 are more prone to develop tamoxifen resistance measured by disease-free survival (91, 92). Increased activity of EGFR, HER2, and HER3 was also connected to resistance to ER agonist, fulvestrant (93). Fulvestrant treatment enhanced the HER3 expression and phosphorylation in breast cancer cells in an NRG1-dependent manner, and this was suggested as a resistance mechanism for fulvestrant in breast cancer (94). In patients with triple-negative breast cancer (TNBC), high HER3/EGFR protein expression (but not HER3 or EGFR alone) conferred worse 10-year survival after chemotherapy. Interestingly, high HER3/EGFR was associated with worse survival following adjuvant chemotherapy, when compared with patients who did not receive adjuvant chemotherapy (95).

In castration-resistant prostate cancer (CRPC), EGFR:HER3 dimers caused androgen receptor therapy resistance via increased PI3K/AKT signaling. Blocking HER3 with siRNA abolished the growth of these cells, suggesting that HER3 was responsible for the increased PI3K/AKT expression (96). Recently, it was also shown that NRG1 secreted by the stromal cells promotes antiandrogen resistance in CRPC (51). Autocrine NRG1/HER3 signaling, measured by NRG1 qPCR and HER3 phosphorylation status, was shown to induce therapy resistance in mouse models of prostate cancer. Blockade of NRG1/HER3 with mAbs resensitized tumors to hormone deprivation in vitro and in vivo. Androgen deprivation therapy was also shown to increase the amount of NRG1-positive CAFs in patients with prostate cancer, measured by IHC and protein analysis. These studies in breast cancer and CRPC suggest that HER3, as well as NRG1 protein levels could be a useful biomarker for the use and withdrawal of hormonal therapies in cancer.

HER2/HER3 coexpression and PI3K/AKT signaling are connected to increased resistance for several chemotherapeutic agents, including 5-fluorouracil, paclitaxel, camptothecin, and etoposide, in breast cancer cells (97). HER3 expression was also reported to cause paclitaxel resistance in HER2-positive breast cancer cells by enhanced expression of AKT and survivin (98). Another DNA-damaging agent, doxorubicin, induced NRG upregulation and HER3-mediated AKT signaling in ovarian cancer cells and dual use of doxorubicin with HER3 inhibition increased apoptosis in the chemoresistant cells (99).

Additional studies link HER3 with resistance to radiotherapy. Ionizing radiation (IR) increases the phosphorylation of EGFR, HER2, HER3, and HER4, and silencing of HER3 reduced the cancer cell viability after treatment with IR in in vitro and in vivo mouse models (100, 101).

mAbs and small-molecule TKIs have been essential in targeting EGFR and HER2 in various tumor types; however, because of the impaired kinase activity, HER3 was long ignored as a therapeutic target. Recently, HER3 has come more into focus as the importance of HER3 in tumor progression and drug resistance has emerged (Fig. 2).

Because HER3 has only minimal kinase activity, HER3-directed antibodies have been the most pursued strategy to target HER3 so far. Various HER3-directed mAbs have been under preclinical and clinical development (Table 1). Most of these agents have been tested in solid tumor types for safety, tolerability, and preliminary efficacy in phase I studies. Seribantumab, lumretuzumab, and patritumab showed the most promising activity in clinical trials so far, progressing up to phase II and phase III studies. Although preclinical mouse studies demonstrate the importance of HER3 in cardiovascular development (102, 103), no significant cardiovascular effects have been observed when anti-HER3 antibodies have been evaluated as single agents or in combination with erlotinib or trastuzumab (104–107). In the phase I trial of patritumab, no dose-limiting toxicities were observed and no MTD was reached (104). The most common treatment-related toxicities were mild and included fatigue, diarrhea, and nausea. Although most of the mAbs demonstrated favorable toxicity profiles, the single-agent activity of the HER3 mAbs has been limited, and development for most of the HER3 antibodies for clinical use has been discontinued. However, recent advances in the development of bispecific (EGFR/HER3, HER3/IGF1, and HER2/HER3) antibodies and antibody–drug conjugates (ADC) have created new hope for HER3 targeting. Bispecific antibodies generally inhibit the kinase by blocking the ligand-binding site (108, 109). ADCs are mAbs conjugated to cytotoxic agents via synthetic linkers and have been shown to induce receptor endocytosis and degradation, as well as cancer cell death (110, 111). Allosteric HER3 antibodies, which do not block NRG1 binding, and can bind HER3 even in the presence of NRG1, maybe more effective than antibodies that block NRG1 binding, but have yet to enter clinical development (112).

Seribantumab (MM-121) is a fully human IgG2 mAb binding to HER3 while blocking the NRG ligand binding and the ligand-dependent downstream activity of HER3 (113). In preclinical studies, seribantumab reduced HER3 activity and growth of xenograft tumors (41, 113). Seribantumab reached several phase I and phase II studies and it was tried either as a single agent or in combination with EGFR-inhibiting antibodies, chemotherapies, or PI3K inhibitors (107, 114). NRG ligand levels were shown to correlate with seribantumab response (107). The furthest phase II studies in combination with paclitaxel or exemestane (aromatase inhibitor) in ovarian cancer and breast cancer did not reach the clinical endpoint of progression-free survival (PFS); however, retrospective analysis showed that there was survival benefit in the NRG-high patient group (115, 116). In EGFR-dependent tumors, there was a limited activity for combination of seribantumab and cetuximab with or without irinotecan chemotherapy (117). Similarly, HER3-specific mAb, lumretuzumab, was combined in a phase Ib study with EGFR inhibiting cetuximab or erlotinib, and even though the toxicity was acceptable, clinical activity was modest in HER3-positive solid tumors (118). Lumretuzumab was evaluated in MBC together with paclitaxel and pertuzumab, but the combination was associated with high incidence of diarrhea and narrow therapeutic window, and the clinical trial was discontinued (119). Recently, a phase II clinical trial with single-agent seribantumab was initiated in solid tumors with NRG1 fusions (NCT04383210).

Patritumab (U3-1287) is a fully humanized HER3-targeting antibody targeted toward the extracellular domain of HER3, which blocks HER3 ligand binding (104). Patritumab suppressed proliferation and survival of cancer cells in in vitro and in vivo xenograft models (120). It was also effective as a combination with anti-EGFR–targeting antibodies in both wild-type EGFR tumor models and models resistant to the first-generation EGFR inhibitors. Circulating NRG ligand was a predictive biomarker for patritumab efficacy in patients with NSCLC (121, 122). In HER2-positive breast cancer, patritumab together with trastuzumab and paclitaxel was reported to have an overall response rate (ORR) of 39% (105). In the further clinical studies assessing the efficacy of patritumab (including a phase III study in NSCLC together with erlotinib), the drug failed to meet the efficacy criteria (NCT02134015). However, as a continuation, the novel HER3 ADC, U3-1402, was constructed using patritumab as the antibody component (110).

Patritumab deruxtecan (HER3-DXd; U3-1402) is a HER3-directed ADC composed of patritumab, a cleavable tetrapeptide-based linker, and a topoisomerase 1 inhibitor (exatecan derivative) payload (123). Patritumab deruxtecan was shown to have preclinical efficacy in NSCLC cells resistant to EGFR inhibitors, as well as in colorectal cancer xenografts, and it was shown to have superior efficacy compared with patritumab alone (110, 123). Patritumab deruxtecan efficacy is associated with high baseline HER3 expression (110). It was recently reported that patritumab deruxtecan induces antitumor immune response through DXd-induced cell damage and immune activation. Patritumab deruxtecan sensitizes HER3-expressing tumors for anti–PD-1 checkpoint blockade in vitro and in vivo, suggesting that it could be beneficial to combine patritumab deruxtecan with immunotherapy agents (124). Preliminary results from a phase I/II study in MBC showed that patritumab deruxtecan has a manageable safety profile and there was an ORR of 42.9% in heavily pretreated patients (125, 126). In another phase I study in metastatic EGFR-mutant and EGFR TKI-resistant NSCLC, patritumab deruxtecan led to a response rate of 25% (127). Importantly, clinical efficacy was observed in cancers with diverse EGFR TKI resistance mechanisms as HER3 is not a known resistance mechanism to EGFR TKIS. In addition, two EV20-derived HER3-specific ADCs, EV20-Sap and EV20/MMAF, have been recently reported (128, 129). EV20-Sap was shown to be effective in preclinical models of melanoma, whereas EV20/MMAF showed preclinical activity in melanoma and breast cancer (130). Also, 9F7-F11–derived HER3 ADC was recently reported (131). So far, no clinical studies with these agents have been reported.

Because tumors often express more than one EGFR family member and the family members are known to induce resistance to single-HER strategies, pan-HER therapies have been developed to overcome compensation mechanisms. Pan-HER therapies are either antibody mixtures, bispecific antibodies directed to multiple antigens, or TKIs targeting more than one EGFR family member. Although most of the pan-HER strategies have focused on cotargeting EGFR and HER2, some strategies attempting to cotarget HER3 have recently emerged.

Pan-HER (Sym013) is a mixture of six antibodies targeting EGFR, HER2, and HER3, and it was shown to reduce cancer cell growth in vitro and in vivo (132). Interestingly, pan-HER was effective even in cells with acquired resistance to cetuximab, trastuzumab, or pertuzumab, or in cells additionally stimulated with EGFR family ligands (132). In a mechanistic study, pan-HER prevented the EGFR family dimer formation and blocked the switch in HER dependencies (133). A phase I/II study with Sym013 was initially launched, but clinical development was subsequently discontinued so the toxicity profile remains unknown (NCT02906670).

Bispecific antibodies can target two distinct tumor-associated antigens and could thus overcome some of the problems of redundant kinase activity. Bispecific antibodies have been developed to simultaneously block either EGFR/HER3, HER2/HER3, or HER3/IGF1R signaling. Duligotuzumab (MEDH7945A) is an EGFR/HER3-directed bispecific antibody that has two identical binding sites binding to the extracellular domain of either EGFR or HER3 (134). Duligotuzumab exhibited antitumor activity in both in vitro and in vivo models and overcame EGFR inhibitors and radiation resistance in preclinical models (135, 136). Although duligotuzumab demonstrated an acceptable safety profile and showed some clinical activity together with cisplatin/5-fluorouracil in a phase IIb study in head and neck cancers, it failed to provide clinical benefit in another study in metastatic colorectal cancer (137, 138). No further clinical activity with the compound has been reported. Recently, a pharmacokinetics predictive study with another EGFR/HER3 bispecific antibody, SI-B001, was reported (139). Bispecific antibody, IgG3-43, was shown to prevent the growth and cancer stem cell expansion in TNBC cells, demonstrating that some preclinical activities in the field of EGFR/HER3 bispecifics are still ongoing (109).

Pertuzumab is an HER2-targeting mAb that blocks HER2/HER3 heterodimerization (34, 140). In a phase I study of 21 patients with solid tumors, pertuzumab was well-tolerated with a pharmacokinetics profile supporting 3-week dosing (141). Patients with HER2-positive breast cancer treated with combination therapy of pertuzumab with chemotherapy and trastuzumab had improved disease-free survival (142). In first-line patients with HER3-positive, HER2-low MBC, ORR was 56% after administration of lumretuzumab (500 mg) every 3 weeks, in combination with pertuzumab [840 mg loading dose (LD) followed by 420 mg] every 3 weeks, and paclitaxel (80 mg/m²) weekly (119). When the patients were given the same combination therapy without the LD of pertuzumab, ORR was 39% (119). However, this therapy demonstrated a small therapeutic range, with a high proportion of patients in the study experiencing grade 3 diarrhea (119).

MM-111 is a bispecific antibody for HER2/HER3 that inhibits heregulin-induced HER3 activation and slows down the tumor growth in HER2-dependent preclinical models (108). The combination of MM-111 and trastuzumab or lapatinib further inhibits growth of HER2-overexpressing cells. MM-111 was tested in clinical trials as a single agent, together with chemotherapies, or together with other HER2-targeting therapies (143). MM-141 (istiratumab) was designed to target HER3 and IGF1R, but it failed to show clinical benefit (144). Zenocutuzumab (MCLA-128) is an HER2/HER3-targeting antibody that showed the most promise in the clinical setting. It is currently being evaluated in solid tumors with NRG1 fusions in phase I studies, and promising interim results were recently published (NCT02912949).

Sapitinib (AZD8931) is a pan-EGFR–targeting inhibitor that inhibits the activation of EGFR, HER2, and HER3 (145). A phase I/II study in advanced breast cancer failed to reach the clinical endpoint (146). Sapitinib was tested in breast cancer together with hormonal therapy, but there was increased toxicity with no additional benefit (147). Sapitinib led to a worse PFS compared with placebo in metastatic colorectal cancer, and no further clinical activities with sapitinib have been reported (148).

It is important to mention that also indirect targeting of HER3 through its dimerization partners can be considered as HER3-targeting strategies, but those approaches are outside the scope of this review.

In addition to direct therapeutic strategies to target HER3, some approaches inhibit HER3 indirectly.

Proteasomal degradation has emerged as a new therapeutic modality, including proteolysis-targeting chimera (PROTAC) inhibitors linked to a warhead that directs the drug target into cellular degradation (149). Although PROTAC for HER3 has not been reported, at least partial HER3 degradation can be achieved with mAbs. The HER3-directed antibody, NG33, was shown to induce HER3 degradation and inhibit the growth of HER2-driven cancer cells (150). HER3 was also degraded by a cross-linked form of trastuzumab binding to HER2. Only HER2 and HER3, but not EGFR, were pulled into degradation (111). Another unique approach to inhibit and degrade HER3 is TX2-121-1, a covalent ligand that binds to HER3 receptor and induces partial HER3 degradation by interfering with HER3 dimerization with HER2 and MET (151). HER3 mRNA degradation can be induced by antisense oligonucleotides or miRNAs, but this concept has not been tested in clinical trials to date (152).

Recently, an HER3-targeting vaccine (Ad-HER3-FL) was created. Ad-HER3-FL stimulates the production of HER3-specific T cells and antibodies in mouse models, suggesting that HER3 might be a good target for antitumor vaccines (153). In the same study, Ad-HER3-FL was also combined with anti–PD-1, showing enhanced response compared with the vaccine alone. Along the same lines, an earlier study described an immunoreacting HER3 epitope (HER-3872-868), and this peptide was used to provoke antitumor immune responses in preclinical models of lung cancer and head and neck cancer (154). Although these new therapy modalities to target HER3 are promising, further preclinical and clinical evaluation is required.

HER3 is an exceptional EGFR family member that is not oncogenic alone, but can cooperate with other receptors to induce tumorigenesis, metastatic events, and drug resistance. HER3 is a compelling cancer therapeutic target, but so far, no HER3-directed therapies have been approved for clinical use. In contrast to EGFR and HER2, which have been broadly targeted with TKIs, HER3 has been mainly targeted with monoclonal or bispecific antibodies due to the minimal kinase activity, either by blocking the ligand binding or heterodimerization with other receptors. In clinical trials, the safety profile of these antibodies has been acceptable, but the efficacy has been disappointing. It is unclear whether the failure of the antibody therapies has been due to the use of wrong antibody epitopes, pharmacokinetics problems, or lack of biomarkers, but with some exceptions, the clinical development of most of these agents has been discontinued. However, retrospective studies and a meta-analysis from HER3 mAbs in various tumor types suggest that NRG expression could be used as a predictor for HER3 mAb response in the future (37, 122, 155). It has also been suggested, that HER3 antibodies that do not compete with the NRG binding site could be more effective in tumors with high NRG1 expression before antibody treatment.

The interest in targeting HER3 has persisted and new mechanisms to target HER3 have constantly emerged. New therapeutic strategies, including HER3-directed ADCs, are being investigated in a broad range of cancers. The potential advantage of ADCs over mAbs could be that the cancer cells need to express HER3, but do not have to be fully dependent on HER3 to induce cell death. Other novel strategies to target HER3 include proteasomal degradation, antisense oligos, and most recently an HER3-targeted peptide vaccine. So far, these agents have shown efficacy in preclinical models, but the clinical safety and efficacy remain to be determined. The winning strategy to therapeutically target HER3 remains to be seen, but HER3 is as a promising drug target and the era of drugging the “undruggables” has already started.

P.A. Jänne reports grants and personal fees from AstraZeneca and Daiichi Sankyo during the conduct of the study, grants and personal fees from Boehringer Ingelheim, Eli Lilly, and from Takeda Oncology; personal fees from Pfizer, Roche/Genentech, Chugai Pharmaceuticals, Loxo Oncology, SFJ Pharmaceuticals, Voronoi, Biocartis, Novartis, Sanofi, Mirati Therapeutics, Transcenta, Silicon Therapeutics, Syndax, Nuvalent, Bayer, and Esai; and grants from Astellas Pharmaceuticals, Revolution Medicines, and PUMA outside the submitted work; and is an inventor on a DFCI-owned patent on EGFR mutations issued and licensed to Lab Corp. No disclosures were reported by the other author.

This study was supported by the American Cancer Society (CRP-17-111-01-CDD to P.A. Jänne) and the NCI (R35 CA220497 to P.A. Jänne). H.M. Haikala was supported by Sigrid Jusélius Foundation and The Finnish Cultural Foundation. Medical writing and editorial support were provided by Jennifer Meyering, MS, RN, CMPP of AlphaBioCom, LLC and funded by Daiichi Sankyo, Inc.