Regulation of CEACAM5 and Therapeutic Efficacy of an Anti-CEACAM5-SN38 Antibody-Drug Conjugate in Neuroendocrine Prostate Cancer - Beyond the Abstract

Despite the initial success of androgen deprivation therapy (ADT) in treating hormone-sensitive prostate adenocarcinoma, patients inevitably develop tumors that become unresponsive to ADT known as castration-resistant prostate cancer (CRPC). CRPC represents a heterogeneous disease including multiple molecular phenotypes and clinical presentations including small cell neuroendocrine prostate cancer (NEPC) which encompasses approximately 20% of CRPC. NEPC is a poorly differentiated, highly aggressive form of prostate cancer that is usually non-responsive to the standard CRPC treatment modalities such as chemo- and radiation therapies and for which we currently have no effective treatment options.

Neuroendocrine prostate cancer (NEPC) most commonly arises from transdifferentiated prostate adenocarcinoma as a mechanism of treatment escape in response to ADT. However, the mechanistic drivers of this transdifferentiation and the associated molecular and phenotypic differences between prostate adenocarcinoma and NEPC have not been fully elucidated. The overall inter- and intra-tumoral molecular heterogeneity in prostate cancer is quite high and must be a significant factor of consideration while developing novel targeted therapies for prostate cancer.

The identification of tumor-specific cell surface antigens and their targeting has yet to be effectively applied to NEPC. Through the integration of transcriptomics and cell surface protein proteomics, Lee et. al. previously developed a method to construct a cell-specific surface antigen profile, known as a surfaceome, from prostate cancer cell lines to identify high-confidence antigens as effective targets for CAR T cell therapy in PrAD and NEPC.1 Carcinoembryonic antigen-related cell adhesion molecule 5 (CEACAM5) was one of several proteins preferentially enriched in NEPC.1 In the current study, we characterize CEACAM5 expression in a broad set of mCRPC cases and healthy tissue samples, elucidate a potential mechanism regulating the enriched expression of CEACAM5 in NEPC, and assess the targetability of CEACAM5 in prostate cancer using an antibody-drug conjugate.2

Due to the known molecular heterogeneity of prostate cancer, we deemed it important to assess CEACAM5 protein expression across a diverse set of mCRPC tissue samples and in relation to the most common alternative protein targets currently being assessed in the preclinical or clinical setting. Using multiplexed immunofluorescence (IF) we simultaneously quantified protein expression of CEACAM5, Trop2, PSMA, and PSCA at the single-cell level. Consistent with our previous results, CEACAM5 had low systemic expression and was expressed in the majority of NEPC tissues (60%) compared to other molecular subtypes of prostate cancer, such as androgen receptor (AR) positive prostate cancer (ARPC), AR, and neuroendocrine positive or amphicrine prostate cancer, as well as AR and neuroendocrine negative prostate cancer. CEACAM5 expression was also found to be lower in cells expressing Trop2, PSMA, or PSCA, highlighting that CEACAM5+ NEPC may be impervious to developing and current imaging modalities and less responsive to therapeutic strategies targeting these other markers.

Although CEACAM5 expression has been reported in multiple cancers of different origins, it is not clear how CEACAM5 expression is regulated in different cancer types including NEPC. Based on published literature surrounding CEACAM5 expression in other neuroendocrine tumors we hypothesized CEACAM5 expression may be regulated by the pioneer transcription factors, ASCL1 or NeuroD1. We analyzed gene expression data from mCRPC cohorts (SU2C, University of Washington mCRPC TAN rapid autopsies, and the LuCaP PDX series) and found CEACAM5 expression to be generally associated with co-expression of ASCL1 and not NeuroD1. Further examination of inferred transaction factor binding and CEACAM5 chromatin accessibility-supported ASCL1 activity and enhanced accessibility of the core CEACAM5 promoter in the CEACAM5-expressing prostate cancer cell line NCI-H660, but not in the CEACAM5-negative cell line MSKCC-EF1. However, knockdown of ASCL1 in the NCI-H660 line did not alter CEACAM5 expression.

To gain further insight into the regulation of CEACAM5 expression by these transcription factors in NEPC, we developed a genetically defined system to induce neuroendocrine differentiation of prostate cancer. We introduced expression of ASCL1 and/or NeuroD1 alone or in combination with several other factors involved in neuroendocrine differentiation (dominant-negative TP53 R175H, shRNA targeting RB1, and MYCN expression) in the AR-independent ARPC cell line, C4-2B. In all conditions expressing ASLC1 or NeuroD1, we observed down-regulation of AR and NKX3.1 as well as up-regulation of synaptophysin, INSM1, and CEACAM5. ASCL1 expression in this model significantly increased chromatin accessibility of the CEACAM5 core protomer while NeuroD1 did not. Our data suggest one mechanism by which CEACAM5 expression is induced during neuroendocrine differentiation is through epigenetic regulation by ASCL1. It is possible that both factors affect global epigenetic profiles that silence AR-driven transcriptional programs and provoke neuroendocrine programs (Figure 1).

Diagram of neuroendocrine transdifferentiation model involving regulation by pioneer transcription factors ASCL1 and NeuroD1

Figure 1. Diagram of neuroendocrine transdifferentiation model involving regulation by pioneer transcription factors ASCL1 and NeuroD1.

We lastly assessed the specificity of the CEACAM5-targeting antibody-drug conjugate (ADC) labetuzumab govitecan for CEACAM5-expressing prostate cancer cell lines in vitro as well as the efficacy of labetuzumab govitecan against patient-derived xenografts from the LuCaP series in preclinical studies. Labetuzumab govitecan demonstrated binding specificity and cytotoxicity against the endogenous CEACAM5-expressing line NCI-H660 as well as several other prostate cancer cell lines with lentiviral overexpression of CEACAM5 compared to their CEACAM5 negative counterparts.

For in vivo preclinical studies, all mice bearing CEACAM5-expressing subcutaneous LuCaP49 or LuCaP145.1 tumors treated with 25mg/kg labetuzumab govitecan every 4 days achieved complete response within 14-17 days compared to vehicle or standard cisplatin/etoposide chemotherapy with minimal off-target toxicities based on animal serum chemistries and complete blood count analyses. A half dose of labetuzumab govitecan (12.5mg/kg) given once weekly yielded complete responses in all mice bearing CEACAM5-expressing LuCaP49 within 24 days and significant reduction and 50% of LuCaP176 tumor-bearing mice within the complete 28-day study.

The results of this study have led to planning for a Phase I/II clinical trial assessing labetuzumab govitecan in patients with CEACAM5 positive NEPC. Investigation of whether labetuzumab govitecan is effective for the treatment of other CEACAM5-expressing neuroendocrine carcinomas may also be warranted.

Written by: Diana C. DeLucia, PhD, Research Fellow, and John K. Lee, MD, PhD, Assistant Professor, Human Biology Division, Fred Hutchinson Cancer Research Center, Seattle, Washington


  1. Lee, John K., Nathanael J. Bangayan, Timothy Chai, Bryan A. Smith, Tiffany E. Pariva, Sangwon Yun, Ajay Vashisht et al. "Systemic surfaceome profiling identifies target antigens for immune-based therapy in subtypes of advanced prostate cancer." Proceedings of the National Academy of Sciences 115, no. 19 (2018): E4473-E4482.
  2. DeLucia, Diana C., Thomas M. Cardillo, Lisa S. Ang, Mark P. Labrecque, Ailin Zhang, James E. Hopkins, Navonil De Sarkar et al. "Regulation of CEACAM5 and therapeutic efficacy of an anti-CEACAM5-SN38 antibody-drug conjugate in neuroendocrine prostate cancer." Clinical Cancer Research (2020).
Read the Abstract