Targeting carcinoma-associated mesothelial cells with antibody–drug conjugates in ovarian carcinomatosis
No conflicts of interest were declared.
Abstract
Ovarian carcinomatosis is characterized by the accumulation of carcinoma-associated mesothelial cells (CAMs) in the peritoneal stroma and mainly originates through a mesothelial-to-mesenchymal transition (MMT) process. MMT has been proposed as a therapeutic target for peritoneal metastasis. Most ovarian cancer (OC) patients present at diagnosis with peritoneal seeding, which makes tumor progression control difficult by MMT modulation. An alternative approach is to use antibody–drug conjugates (ADCs) targeted directly to attack CAMs. This strategy could represent the cornerstone of precision-based medicine for peritoneal carcinomatosis. Here, we performed complete transcriptome analyses of ascitic fluid-isolated CAMs in advanced OC patients with primary-, high-, and low-grade, serous subtypes and following neoadjuvant chemotherapy. Our findings suggest that both cancer biological aggressiveness and chemotherapy-induced tumor mass reduction reflect the MMT-associated changes that take place in the tumor surrounding microenvironment. Accordingly, MMT-related genes, including fibroblast activation protein (FAP), mannose receptor C type 2 (MRC2), interleukin-11 receptor alpha (IL11RA), myristoylated alanine-rich C-kinase substrate (MARCKS), and sulfatase-1 (SULF1), were identified as specific actionable targets in CAMs of OC patients, which is a crucial step in the de novo design of ADCs. These cell surface target receptors were also validated in peritoneal CAMs of colorectal cancer peritoneal implants, indicating that ADC-based treatment could extend to other abdominal tumors that show peritoneal colonization. As proof of concept, a FAP-targeted ADC reduced tumor growth in an OC xenograft mouse model with peritoneal metastasis-associated fibroblasts. In summary, we propose MMT as a potential source of ADC-based therapeutic targets for peritoneal carcinomatosis. © 2023 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.
Introduction
Epithelial ovarian cancer (OC) represents the vast majority of ovarian tumors. They are classified into low-grade serous (LGSOC), high-grade serous (HGSOC), mucinous, endometrioid, and clear-cell OC. HGSOC accounts for about 70% of OC-related deaths. In contrast, the LGSOC subtype represents 5–7% of serous OCs [1, 2]. Despite sharing the same serous histotype, LGSOC and HGSOC are two separate entities. LGSOC is typically a slow-growing carcinoma, while HGSOC is characterized by its aggressive behavior [3, 4].
A common characteristic of serous OCs is that they mainly metastasize via transcoelomic spread [5]. Peritoneal metastasis frequently evolves very rapidly and without symptoms. This is the main reason why patients are usually diagnosed at advanced-stage disease. Today's standard of care for patients with ovarian carcinomatosis includes debulking surgery combined with taxane-platinum chemotherapy and maintenance with poly(ADP-ribose) polymerase (PARP) inhibitors [6]. Initially, the majority of patients with advanced OC responds to chemotherapy, but up to 85% ultimately have a disease recurrence due to acquired chemoresistance, with a 5-year survival rate of ~30% [7].
Peritoneal carcinomatosis comprises cancer cell detachment from the primary tumor, dissemination through the peritoneal cavity, and attachment to and invasion through the mesothelial cell (MC) monolayer that lines the peritoneal cavity [8]. Mesothelial-to-mesenchymal transition (MMT) plays a fundamental role in the pathogenesis of peritoneal metastasis [9]. The myofibroblast conversion of MCs is a complex cell reprogramming process characterized by the progressive loss of mesothelial markers [cytokeratins, mesothelin (MSLN), calretinin, WT1] and the acquisition of contractile, migratory, and invasive proteins such as alpha-smooth muscle actin (α-SMA) and fibroblast-specific protein (FSP)-1 (S100A4) [10-12]. Our group determined, for the first time, that a sizeable population of carcinoma-associated MCs (CAMs) that originated through MMT accumulates in the peritoneal stroma and promotes tumor adhesion, invasion, vascularization, and growth [12]. In fact, MMT has been widely proposed as a therapeutic target in different peritoneal fibrotic diseases [13-15]. The therapeutic strategies may be designed either to prevent or reverse the MMT itself or to treat its effects such as cellular invasion, fibrosis, and angiogenesis [15]. The problem we face with OC patients is that the majority are diagnosed at an advanced stage of the disease [4], reaching a turning point when it is too late to block or revert the MMT and, therefore, tumor progression control. In this study we propose a new therapeutic approach using monoclonal antibody–drug conjugates (ADCs) for the treatment of ovarian carcinomatosis. ADC-based therapies utilize a monoclonal antibody to specifically deliver the cytotoxic payload into target antigen-expressing cells.
The development of ADC-based therapies is complex and requires the discovery of specific surface molecules on the target cell that are not expressed in healthy tissues. In addition, upon binding of the ADC on the cell membrane, the agent needs to be internalized and transported inside the cell, within ‘late endosomes’, where the cytotoxic moiety is released [16-18]. A plethora of ADCs are being investigated in phase I and II clinical trials, where new and promising data are emerging [19]. It is known that MUC16-MSLN interactions mediate the attachment of OC cells to CAMs [20]. On this note, anti-MUC16 monoclonal antibodies conjugated to the microtubule-disrupting agent monomethyl auristatin E (MMAE) show antitumor activity in OC [21, 22]. Additionally, MSLN, frequently overexpressed in tumors such as mesothelioma and OC, has been targeted in xenograft models with a humanized antibody conjugated to DM4 [23]. However, the identification of specific targets in tumor cells mainly fails due to their intrinsic genomic instability. Alternatively, the perineoplastic stroma is currently emerging as an attractive target for ADC-based cancer therapy [17, 24]. In this study, we propose the development of novel immunotherapeutic tools directed against CAMs, which support tumor progression in peritoneal carcinomatosis, by comparing the MMT-related gene expression profiles of ascitic fluid-isolated CAMs in patients with LGSOC or HGSOC, before and after neoadjuvant chemotherapy. We identify the surfaceome of CAMs as an important reservoir of specific targets for the subsequent design of novel ADCs. Our results establish a proof of concept about the feasibility of target CAMs by inhibiting tumor growth with a fibroblast activation protein (FAP)-targeting ADC in a mouse model for peritoneal metastasis of OC.
Materials and methods
Patient samples
This study complies with the Declaration of Helsinki and was approved by the Research Ethics Committee of Fundación Jiménez Díaz – QuirónSalud (Madrid, Spain; ethic approval number: 11/17). Informed written consent to use biological samples was obtained from patients.
Approximately 50 ml of peritoneal ascitic fluid were obtained from patients with serous OC, International Federation of Gynecology and Obstetrics (FIGO) stage IIIc or IV. All patients underwent primary or interval (following neoadjuvant chemotherapy) cytoreductive surgery. Ascitic fluid-isolated MCs were obtained by centrifugation (1,200 rpm, 5 min) of peritoneal effusions, as described [13]. Primary human peritoneal MCs (HPMCs), isolated from omentum samples of nononcological patients, were used as controls [13].
Surgical biopsies from 22 peritoneal implants of LGSOC (n = 5), HGSOC (n = 9), and HGSOC after chemotherapy (HGSOC+ChT) (n = 8) were considered for immunohistochemical evaluation. Additionally, peritoneal biopsies of four colorectal cancer patients and three control peritoneal tissues obtained from nononcological patients were included.
Cell cultures
MCs were grown in Earle's M199 medium (Bilogical Industries, Israel), supplemented with 20% fetal bovine serum (FBS; Cytiva, USA) and 2% Biogro-2 (Biological Industries). Procedures to establish the elliptical factor (EF) of MC cultures [10] are detailed in Supplementary materials and methods.
To induce MMT in vitro, HPMCs were treated with 0.5 ng/ml of transforming growth factor (TGF)-β1 (R&D Systems, USA) for 72 h [25]. Where indicated, HPMCs were incubated with interleukin-11 (IL11; Gibco, USA) at 0.5 ng/ml for 72 h.
RNA sequencing
MCs of four omentums, six LGSOC, three HGSOC, and five HGSOC + ChT patients were used for RNA sequencing (RNA-seq) studies. Cell cultures were lysed with QIAzol Reagent (QIAGEN, USA), and total RNA was extracted with the miRNeasy Mini Kit (QIAGEN), following the manufacturer's instructions. Total RNA integrity was checked using RiboGreen-based fluorescent determination and analyzed using a Bioanalyzer (Agilent Technologies, USA). RNA-seq was performed by the Genomics Facility at Parque Científico de Madrid (Spain) using Illumina platform. Raw data obtained in FASTQ format were preanalyzed by the Genomics and Next Generation Sequencing Core Facility at the CBMSO (Madrid, Spain) [26]. Further details can be found in Supplementary materials and methods.
Mouse model of peritoneal metastasis
A total of 22 Swiss nu/nu female mice, 6–7 weeks old, were used (Charles River Laboratories, Barcelona, Spain). The experimental protocol followed the National Institutes of Health Guide for Care and Use of Laboratory Animals and was favorably approved by the Community of Madrid (Spain; PROEX number 273/19). See Supplementary materials and methods for details.
Procedures for cell sorting (FACS), qRT-qPCR, western blotting, enzyme-linked immunoassay (ELISA), immunohistochemistry and statistics are described in Supplementary materials and methods. Primers for RT-qPCR are shown in supplementary material, Table S1.
Results
RNA-seq of CAMs stratifies patients with metastatic OC
CAMs identified by the co-expression of podoplanin (PDPN) and FAP (Figure 1A and supplementary material, Figure S1) were isolated from the peritoneal ascitic fluid of LGSOC, HGSOC, and HGSOC+ChT patients, and subjected to RNA-seq. Heatmaps for cluster classification (Figure 1B) and minus average (MA) plots (Figure 1C) revealed a clear separation between each condition and PDPN+/FAP− HPMCs (supplementary material, Figure S1). Changes in gene expression were obtained when compared LGSOC versus HGSOC, HGSOC versus HGSOC+ChT, and LGSOC versus HGSOC+ChT (supplementary material, Figure S2). The number of differentially expressed genes (DEGs) (q value < 0.05) for each comparative is shown in supplementary material, Table S2. Accordingly, principal component analysis (PCA) (Figure 1D) and heatmaps (supplementary material, Figure S3) show a clear variation between all groups. Only a total of 29 common DEGs (q value < 0.05) of the three groups compared to the control group were obtained, suggesting that there is a unique gene expression profile for each condition (Figure 1E). Therefore, ascitic fluid-isolated CAMs stratify serous OC patients according to the histological subtype. Additionally, HGSOC+ChT patients differed from those untreated by showing a distinct gene expression signature.
Ascitic fluid-isolated CAMs reflect MMT-related changes that take place in the peritoneal metastatic niche
Ensembl ID of DEGs in ascitic fluid-isolated MCs of LGSOC, HGSOC, and HGSOC+ChT datasets (all of them referred to HPMCs) were submitted to Ingenuity Pathway Analysis (IPA) software for mapping to canonical pathways and identification of upstream regulators. A significant enrichment in several canonical signaling pathways related to cancer mechanisms, fibrosis, and MMT was observed for the three datasets (Figure 2A and supplementary material, Table S3). TGF-β1, a key inducer of MMT [25], was identified among the principal upstream regulators (supplementary material, Table S4). A MMT-related gene signature was identified by the downregulation of mesothelial markers [CALB2, PDPN, cytokeratins (KRTs), CDH1 and KDR] and the upregulation of MMT-associated genes [TGF-β1, SNAI1, VEGFA, collagens, matrix metalloproteinases (MMPs) and FAP]. This signature was significantly more robust in HGSOC compared to LGSOC patients. Moreover, the group of HGSOC+ChT showed a smaller MMT-related gene signature compared to untreated HGSOC patients (Figure 2B). Ex vivo cultures of CAMs drained from a total of 35 serous OC patients were classified into epithelioid and nonepithelioid phenotype according to EF measurements. Remarkably, 50% of LGSOC patients drained CAMs with an epithelioid morphology, while 16 of 17 (94.1%) HGSOC patients drained nonepithelioid cells. Additionally, the percentage of nonepithelioid ascitic fluid-isolated CAMs in HGSOC+ChT patients was reduced to 75% compared with untreated HGSOC patients. The comparison between these three conditions was statistically significant (χ2 test p value = 0.01; Figure 2C and Welch-ANOVA p value = 0.01; supplementary material, Figure S4).
Immunohistochemical detection of PDPN and α-SMA was performed in parietal peritoneal biopsies obtained from LGSOC patients and compared to HGSOC patients, regardless of whether or not they had received chemotherapy. Supplementary material, Figure S5 shows a control peritoneum where PDPN+/α-SMA staining depicts a conserved mesothelial monolayer. Supplementary material, Figure S6 shows staining of calretinin and pan-cytokeratin as alternative CAM markers. Large α-SMA+ areas surrounding submesothelial metastatic nodules were observed independently of histological subtype. These α-SMA+ zones overlapped with PDPN+ submesothelial CAMs in HGSOC patients. However, PDPN staining was limited to a few α-SMA+ CAMs in LGSOC and HGSOC+ChT peritoneal implants (Figure 3). These data are in accordance with our RNA-seq data where PDPN expression was maintained in HGSOCpatients and downregulated in LGSOC and HGSOC+ChT patients (Figure 2B).
Identification of potential targets on surface of CAMs for ADC design
We extracted a five-gene signature from our RNA-seq study integrated by FAP, mannose receptor C type 2 (MRC2), interleukin-11 receptor alpha (IL11RA), myristoylated alanine-rich C-kinase substrate (MARCKS), and sulfatase-1 (SULF1). All of them were genes encoding proteins predicted to be (1) significantly upregulated in at least CAMs of HGSOC patients compared to HPMCs (Figure 4A), (2) localized in the cell plasma membrane [27] (supplementary material, Figure S7), and (3) associated with poor prognosis in serous OC [28] (supplementary material, Figure S8).
We found that HPMCs significantly upregulated FAP upon TGF-β1 treatment. Accordingly, CAMs drained by HGSOC patients showed a significant increased expression of FAP compared with control MCs. While FAP staining was not detected in the mesothelial layer of nonpathological peritoneal tissues, it overlapped with PDPN+/α-SMA+ submesothelial CAMs in serial sections of HGSOC peritoneal biopsies (Figure 4B).
Independently of the OC histological subtype, CAMs significantly overexpressed MRC2 compared to control MCs (Figure 4C and supplementary material, Figure S9). Additionally, MRC2 was detected in submesothelial CAMs expressing PDPN and α-SMA (Figure 4C).
We have also found that, upon TGF-β1 treatment, MCs produced high levels of IL11. Accordingly, at least those CAMs that drained into the peritoneal cavity of HGSOC patients secreted high levels of IL11. Stimulation of HPMCs with IL11 led to the downregulation of E-cadherin, as well as the upregulation of MMP2 and Collagen I and its cognate receptor IL11RA (Figure 4D, supplementary material, Figures S9 and S10). In this regard, control peritoneal tissues did not show IL11RA expression. However, peritoneal implants of HGSOC patients showed IL11RA staining in CAMs and in the metastatic tumor parenchyma (Figure 4D).
MARCKS and SULF1 were specifically detected in CAMs that were accumulated in superficial or deeper submesothelial compact zones of OC peritoneal implants (Figure 5A).
Detection of potential targets for ADC treatment in peritoneal CAMs of colorectal cancer peritoneal implants
A single-cell sequencing analysis of primary colorectal tumors compared to adjacent nonmalignant colon tissues [29] identified a MC-derived fibroblast cluster co-expressing PDPN and ACTA2, which in turns overlapped with our five-gene signature predicted in peritoneal MCs drained by ovarian carcinomatosis patients (supplementary material, Figure S11). Consistent with these data, Figure 5B shows representative images for the immunohistochemical detection of FAP, MRC2, MARCKS, and SULF1, mainly limited to CAMs surrounding colorectal tumor peritoneal implants. IL11RA staining was detected in peritoneal CAMs, but also in the tumor parenchyma, as observed for OC patients.
Proof of concept for use of ADCs in a mouse model of OC peritoneal metastasis
Fabre et al. previously showed that FAP targeting with OMTX705, a novel antibody linked to TAM470, represents a potent strategy as a single agent and in combination with chemotherapy for the treatment of different solid tumors [24]. However, the activity of OMTX705 has not been demonstrated in peritoneal metastases of abdominal cancers. For this purpose, SKOV3-luc-D3 cells, an OC cell line expressing luciferase, were i.p. injected into nude mice. The tumor was allowed to grow, and OMTX705 was administered i.v. once a week over 4 weeks. As control groups, mice received PBS (vehicle) or anti-FAP unconjugated antibody (OMTX005) (Figure 6A). Mouse weights were maintained during the experiment (Figure 6B). In vivo bioluminescence imaging showed 53.1–67.9% tumor growth inhibition compared to control groups (Figure 6C). Accordingly, metastatic implants in parietal peritoneal tissues were quantified, showing that OC peritoneal carcinomatosis was significantly reduced in mice receiving OMTX705 treatment compared to control groups (Figure 6D). Additionally, a large accumulation of submesothelial CAMs (positive for FAP) were detected immunohistochemically in control mice bearing SKOV3 cells (Figure 6E). In contrast, the mesothelial monolayer was predominantly conserved upon OMTX705 treatment (Figure 6E). To support the FAP binding, peritoneal tissue sections were stained with an anti-human IgG antibody. As shown in Figure 6E, OMTX005 bound to intratumoral carcinoma-associated fibroblasts (CAFs), and no staining was observed in OC cells.
Discussion
It has been suggested that peritoneal colonization is preceded by the liberation of molecular signals by a tumor in the niche that is susceptible to metastasis [9]. Epithelial OC is a highly heterogeneous disease with distinct intrinsic subtypes and different clinical behavior [30]. Therefore, it is tempting to speculate that the response of the mesothelium upon peritoneal ascitic fluid exposition will differ according to the specific tumor secretome profile. On this note, Pakula et al showed that the MMT-related changes evoked by aggressive OC cell lines are more pronounced than those triggered by their less harmful counterparts [31]. However, the experimental use of cell lines does not entirely represent the biological responsiveness to the ascites observed in patients [32]. In this study we obtained the whole transcriptomic gene expression pattern of primary ascitic fluid-isolated CAMs of LGSOC and HGSOC patients. Our data revealed that CAMs stratified OC patients according to their tumoral histotype. The enrichment of several canonical pathways related to MMT indicated that the mesenchymal conversion of MCs is a common event in OC patients with peritoneal progression. However, a significantly stronger MMT-associated gene signature was detected in HGSOC patients compared to LGSOC patients, suggesting that the aggressiveness of the tumor is reflected in the MMT-related changes that are taking place in the peritoneal cavity. Interestingly, we found that LGSOC and HGSOC histotypes were linked with specific stages of the MMT process, earlier in the former and advanced in the latter. Taken together, these data suggest that the mesothelial compartment of the ascitic fluid could represent a reliable reservoir of biomarkers with powerful diagnostic value for patients with peritoneal metastasis.
The main challenge in the clinical management of patients with aggressive OC histological subtype is the development of tumor resistance and relapses following chemotherapeutic treatment [33]. Standard chemotherapy is mainly focused on exerting a reductive effect in the tumor compartment. However, signals secreted by CAFs after chemotherapy may enhance chemoresistance and tumor regrowth from residual cancer cells [8, 34, 35]. In this context, we found that ascitic fluid-isolated CAMs from chemosensitive patients recovered the expression of some mesothelial markers and downregulated mesenchymal genes, suggesting a partial reversion from an advanced MMT toward earlier stages. Therefore, the consideration in further ‘omic’ studies of CAMs from chemoresistant OC patients could shed light on the discovery of novel theragnostic biomarkers in the ascitic fluid of individual patients [36].
Previous studies from our group suggested that ascitic fluid-isolated CAMs reflected the changes taking place in the peritoneal metastatic niche of OC patients [13]. Here, we reaffirm this assertion by means of the immunohistochemical detection of PDPN in OC peritoneal implant biopsies. PDPN is physiologically expressed in the mesothelium lining serous cavities and is frequently used for tracking cells of a mesothelial origin in peritoneal fibrotic diseases [37]. Under pathological conditions, PDPN promotes an epithelial-to-mesenchymal transition (EMT), leading to cell motility/migration, invasiveness, and ECM degradation [38-40]. In agreement with our RNA-seq data, we found that PDPN staining was mainly maintained in CAMs of HGSOC peritoneal biopsies. In this regard, it was shown that PDPN+ CAFs predicted poor cancer prognosis [41]. In addition, we found that PDPN expression was limited in LGSOC biopsies, suggesting that intertumoral differences in terms of aggressiveness are represented in the surrounding peritoneal stroma. We determined that the expression of PDPN was significantly reduced in ascitic fluid-isolated MCs, as well as in the peritoneal biopsies of patients exposed to chemotherapy, supporting the idea that OC chemosensitivity is also reflected in the tumor peritoneal microenvironment.
The delineation of molecular signatures from CAFs is paving the way toward OC biomarker discovery. In addition, CAF-associated gene profiles are an important source for the design of novel therapeutic approaches. On this note, interfering with key signaling pathways that regulate the MMT process itself could represent a feasible therapeutic target in the context of OC peritoneal metastasis [13]. However, most OC patients experience nonspecific symptoms, and, usually at diagnosis, the tumor presents with peritoneal extension, which means that it is too late to control tumor progression by interfering with the MMT process. Therefore, an alternative therapeutic strategy could specifically deplete the population of CAMs, which support OC peritoneal progression, with engineered ADCs. However, the design of novel ADC-based therapies requires the successful identification of specific targets on the cell surface to allow an efficient antibody binding. In this regard, a five-gene cell surfaceome signature composed of FAP, MRC2, IL11RA, MARCKS, and SULF1 was extracted from our RNA-seq study. All of them were significantly upregulated in ascetic fluid-isolated CAMs from at least HGSOC patients. Accordingly, the expression of all five proteins was successfully validated in HGSOC peritoneal biopsies by overlapping with PDPN+/α-SMA+ submesothelial CAMs, and basal expression was not detected in the mesothelial monolayer of control peritoneal biopsies. Some molecular aspects of these CAM-specific targets are discussed in what follows.
FAP protein is an integral membrane serine protease [42] expressed during embryonic development [43]. In adults, FAP is restricted to reactive fibroblasts at sites of wound healing, fibrosis, and cancer [44-46]. Its expression was previously observed by immunofluorescence on the membrane of HPMCs treated with TGF-β1 [47]. Interestingly, we found by cytometry of nonpermeabilized cells that a high percentage of ascitic fluid-isolated CAMs from serous OC patients were FAP+.
MRC2, also known as ENDO180, plays a role in ECM remodeling by mediating the internalization and lysosomal degradation of collagen ligands [48]. Important pathological functions for MRC2 have been identified in various mesenchymal malignancies such as sarcoma, glioblastoma, and leukemia [49], as well as in some epithelial tumors [50-52] and fibrotic diseases [53]. Interestingly, the genetic deletion of MRC2, which is predominantly expressed by a subpopulation of matrix-remodeling CAFs, profoundly limited tumor growth and metastasis [54]. Here, we identified MRC2 as a potential molecular target for ADC-mediated drug delivery in peritoneal carcinomatosis since a significant upregulation of the receptor in ex vivo CAM cultures compared to control MCs was observed. In fact, a MRC2-targeting ADC has been successfully tested in a xenograft mouse model of leukemia [49].
Our group previously revealed that CAMs showed activation of Smad3-dependent TGF-β signaling, which is disrupted in OC cells, despite their elevated TGF-β production [13]. In this regard, the secretion of IL11 by TGF-β-stimulated CAFs triggers GP130/STAT3 signaling in cancer cells, conferring on them a survival advantage and chemoresistance [55-57]. In the context of OC peritoneal metastasis, we found that both TGF-β-stimulated MCs and CAMs secreted high levels of IL11. Upon IL11 treatment, MCs decreased e-cadherin expression, upregulated relevant MMT-related genes, and led to the expression of its cognate receptor IL11RA. These data suggest that TGF-β1, a major inducer of MMT [25], can initiate an autocrine loop of the IL11/IL11RA signaling pathway in MCs, which contributes to the translation of MMT-related genes. Regarding IL11RA, its expression has been closely associated with fibroblasts in fibrosis and epithelial cells in cancer [58-60]. Concerning these data, we found prominent IL11RA staining in both CAMs and metastatic cells of HGSOC peritoneal implants. This bispecific expression opens a new line of research focused on the development of ADCs triggered to attack both tumor cells and their niche, especially in chemoresistant peritoneal micrometastases.
MARCKS is an actin filament crosslinking protein with pleiotropic functions, such as cell motility, integrin activation, membrane trafficking, phagocytosis, and exocytosis [61-63]. In agreement with our data, MARCKS has been identified as a novel stroma-oriented therapy in OC [64], and its accumulation in lysosomes was previously determined in fibroblasts [65, 66].
Finally, SULF1 has been described as a TGF-β1-induced extracellular heparan sulfate endosulfatase implicated in fibrogenesis [67].
Besides serous OC, other abdominal malignancies, including colorectal [9, 12], gastric [68], and pancreatic tumors [12], trigger MMT-mediated peritoneal carcinomatosis. In this regard, we detected FAP, MRC2, IL11RA, MARCKS, and SULF1 immunostaining in the stroma of colon cancer peritoneal implants, indicating that the application of ADC-based therapies could be extended to other tumors presenting advanced peritoneal disease. Moreover, ADCs targeted to CAMs could prevent the locoregional extension of primary abdominal tumors with no peritoneal disease at the moment of diagnosis, since MMT contributes to the generation of CAMs in locally advanced primary colorectal carcinomas [11].
This work attempted to summarize the potential of some specific markers to be targeted with ADCs in the context of peritoneal carcinomatosis. However, we are aware that further studies should delve into the internalization routes of surface targets in CAMs and rule out their expression in healthy adult tissues outside the peritoneum.
The generation of ADC-engineered therapeutic tools is currently booming. The first ADC was approved for the treatment of OC in November 2022 [69], and several ADCs are under investigation in this patient population. Clinical response rates with ADCs in OC range from 5% to 45%, with improved response in tumors with high expression of the target antigen and markedly improved response rates when utilized in combination with chemotherapy (52–80%) [19, 22, 23]. While more ADCs are being designed to target tumor cells, there are still few advances in the development of immunotherapeutic tools targeted at the stroma, and currently, no in vivo assays exist in the context of MMT-mediated peritoneal carcinomatosis. Here we provide, for the first time, a feasible proof of concept by demonstrating the antitumoral activity of an i.v. FAP-targeted ADC therapy (OMTX705), without significant adverse effects, in a mouse model of OC peritoneal metastasis. Importantly, none of the ADCs developed up till now has targeted the FAP+ CAMs present in the peritoneal metastatic niche of OC, and none is using a tubulysin payload such as TAM470, which presents an interesting advantage over MMAE or DM4 for not being a substrate of P-glycoprotein transporters. As a mechanism of action, OMTX705 has been found to bind selectively to FAP+ CAFs in the tumor stroma of patient-derived xenograft murine models, be internalized, and be cleaved by specific proteases from the late endosome to release its cytotoxic payload moiety. Once released, TAM470 is able to bind alpha-tubulin and prevents its polymerization into microtubules, thereby inducing apoptosis of FAP+ target cells [24].
Acknowledgements
This work was supported by the following grants to MLC: Spanish Ministry of Science and Innovation/Fondo Europeo de Desarrollo Regional (MICINN/FEDER) Grant PID2019-110132RB-I00/AEI/10.13039/501100011033 and Marie Sklodowska-Curie Innovative Training Networks-European Training Networks Grant 812699.
Author contributions statement
LPA, PS, LGC, RSC and MLC designed the research. PS and MLC supervised the study. LPA, PS, GTGM, VK and HTS developed the methodology. LPA, PS and EMAP acquired and analyzed the experimental data. JAJH, LGC and RSC provided patient samples. MF, IE, CF and LS provided expertise for the use of OMTX705 in animal experiments. LPA, PS and MLC wrote the manuscript and prepared the figures. All authors reviewed the manuscript.
Open Research
Data availability statement
Data files from transcriptome profiling analysis were deposited in the European Genome-Phenome Archive (EGA) and assigned the EGA accession number EGAS00001003747.
