Integration of ERα-PELP1-HER2 signaling by LSD1 (KDM1A/AOF2) offers combinatorial therapeutic opportunities to circumventing hormone resistance in breast cancer

Abstract

LSD1, an epigenetic modifier, and PELP1, an estrogen receptor co-activator, integrate estrogen receptor ERα and HER2 receptor tyrosine kinase signaling to promote aromatase expression and hormone resistance in a preclinical model of post-menopausal breast cancer. In the previous issue of Breast Cancer Research, Cortez et al. show, for the first time, that knockdown or drug-mediated inhibition of PELP1 or LSD1 suppresses LSD1-mediated transcriptionally activating histone marks at ERα target genes, inhibits aromatase gene expression, and sensitizes hormone refractory breast cancer cells to tamoxifen or letrozole treatments. The relevance of PELP1-LSD1 signaling to other nuclear hormone receptor-dependent cancers and structural considerations for the selective drug targeting of LSD1 are further discussed in this editorial.

The fact that the majority of breast cancers express estrogen receptor-alpha (ERα) and thus are estrogen-dependent is both a curse and a blessing. It is a blessing because ERα is a well-defined molecular target that can be efficiently inhibited by drugs in adjuvant therapy to avert relapse as well as in the palliative treatment of advanced disease. It is a curse because a significant percentage of patients fail to respond to treatment and end up relapsing. Thus, for instance, although the ERα antagonist tamoxifen has proven to be one of the most successful drugs to be developed for the targeted therapy of cancer, more than half of patients with ERα-positive breast cancer show intrinsic or acquired tamoxifen resistance. Therefore, the mechanisms of ERα pathway drug resistance and the means of circumventing them represent high-priority fields in breast cancer research.

A predominant mechanism by which ERα drives breast cancer pathogenesis is the so-called nuclear genomic pathway. In this process, ERα recruits co-activator complexes to gene targets to potentiate their transcription. Co-activator complexes facilitate transcriptional activation in part by interacting with chromatin remodeling and histone-modifying enzymes which render the target chromatin template permissive to transcriptional activation. One such protein is LSD1 [1], a flavin adenine dinucleotide-dependent amine oxidase that catalyzes methyl group removal from methylated histone H3 lysine 4 (H3K4) and H3 lysine 9 (H3K9) to effect transcriptional repression or activation, respectively (to avoid gene symbol and species ambiguity [2], all genes discussed in this editorial are accompanied by their unique National Center for Biotechnology Information (NCBI) GeneID: LSD1, also known as KDM1 or AOF2; GeneID 23028; encodes lysine-specific histone demethylase 1). LSD1 demethylates H3K9 of ERα chromatin targets in an estrogen-dependent manner, leading to hydrogen peroxide production and recruitment of 8-oxoguanine DNA glycosylase 1 and topoisomerase IIb, which promote DNA bending and the looping out of enhancer-promoter intervening DNA sequences [3].

In the previous issue of Breast Cancer Research, Vadlamudi and colleagues [1], from the University of Texas and from MD Anderson Cancer Center, implicate PELP1 (also known as HMX3 or MNAR; GeneID 27043; encodes proline-, glutamic acid-, and leucine-rich protein 1), a transcriptional co-activator that harbors neither DNA binding nor activation domains [4] and that couples LSD1 to ERα [5], in an ERα-PELP1-LSD1 axis of hormone resistance in breast cancer [1].

Vadlamudi and colleagues [4] have shown that PELP1 is a dimethylated H3K4 (H3K4me2) and H3K9me2 reader and tilts LSD1 specificity toward H3K9me2, thus gearing LSD1 activity into transcriptional activation of ERα targets. PELP1 also induces the expression of the CYP19A1 gene (also known as cytochrome P450; GeneID 1588), which encodes aromatase, an estrogen synthase, thus promoting in situ estrogen synthesis and the ensuing autocrine or paracrine activation of ERα [6]. Aromatase induction was previously shown to be under positive control of the HER2 receptor tyrosine kinase proto-oncogene (also known as ERBB2 or CD340; GeneID 2064; encodes human epidermal growth factor receptor 2) [7], and Vadlamudi and colleagues show that the HER2-aromatase pathway is also under positive regulation by PELP1-LSD1 signaling [1], further implicat-ing PELP1 and LSD1 in ERα-driven breast cancer patho-genesis and hormone resistance. LSD1, therefore, may offer a molecular targeted therapy to counteract endo-crine resistance and autocrine/paracrine resistance that often take place in breast cancer in pre- and post-menopausal women, respectively. To that effect, Vadlamudi and colleagues show, in preclinical models of post-menopausal breast cancer, that blocking LSD1 activity with a drug approved by the US Food and Drug Administration (FDA) inhibits aromatase activity and reduces tumor growth in vivo. Importantly, the authors also show that inhibiting LSD1 sensitizes tamoxifen-resistant cells to tamoxifen, thus making the case for the use of LSD1-targeting drugs in combination therapies to circumvent tamoxifen resistance in breast cancer [1].

LSD1, which was first reported to be associated with cancer only three years ago, has since been found to possess oncogenic properties in several unrelated cancers ranging from bladder and lung cancers to sarcomas, and its inhibition reduces or blocks cell growth in many of these cancers (reviewed in [8, 9]). Currently, two FDA-approved drugs can block LSD1 activity: pargyline (brand names: Eudatin, Eutonyl, Eutonyl-ten, Lopac-P-8013, and Supirdyl), an anti-hypertensive drug, and tranylcypro-mine (brand name: Parnate), which is used to treat major depressive episodes in adults. Both drugs are monoamine oxidase (MAO) inhibitors with relatively broad target spectra and are associated with a large number of side effets and drug interactions, thus severely limiting their clinical use. Vadlamudi and colleagues judiciously show that N-[(1S)-3-[3-(trans-2-aminocyclopropyl)phenoxy]-1-(benzylcarbamoyl)propyl]benzamide, an experimental drug that was synthesized by Miyata and colleagues at Nagoya City University in Japan and that inhibits LSD1 at concentrations that are two orders of magnitude lower than those of pargyline and tranylcypromine [10], also efficiently inhibits aromatase production and breast cancer cell growth in vitro [1]. To date, the data show that targeting the ERα-PELP1-LSD1 axis is likely to be a viable strategy in breast cancer treatment.

In addition to being associated with breast cancer, PELP1 is associated with other hormone-dependent cancers such as endometrial, ovarian, and prostate cancers [11–13] as well as with astrocytic brain tumors [14]. Although the latter are not traditionally thought of as hormone-dependent neoplasms, astrocytomas express various levels of several nuclear hormone receptors and co-activators, and normal astrocytes are responsive to estradiol modulation [14–16]. It would be of interest to investigate whether PELP1-LSD1 crosstalk also takes place in these cancers and, in the affirmative, whether drug targeting of the PELP1-LSD1 axis also makes therapeutic sense in these tumors. The rationale of such investigations is further supported by recent reports of the inhibition of growth, or induction of apoptosis, of prostate and astrocytic tumors following drug inhibition of LSD1 in in vitro and in vivo preclinical models [17, 18].

Together, the central role played by LSD1 in the ERα-, PELP1-, and HER2-dependent biology of breast cancer and its hormone-dependent or -independent roles in other cancers warrant that new selective LSD1-targeting drugs approved for clinical use be urgently developed. In that regard, drugs that target the N-terminal LSD1 SWIRM domain, a six-α-helical structural module fre-quently found in chromatin-associated proteins and important to LSD1 protein stability, might oer signifi-cantly improved selectivity, and thus fewer side effects, in comparison with current FDA-approved MAO inhibi-tors. In particular, the first and second α helices of the LSD1 SWIRM domain spanning residues 172 to 196 show insignificant sequence and structural homology to the closest human SWIRM domain-containing proteins, providing an ideal platform for the selective targeting of LSD1.