Research ArticleCell biologyOncology Open Access | 10.1172/JCI168146

Ricardo Cruz-Acuña,1 Secunda W. Kariuki,1 Kensuke Sugiura,1 Spyros Karaiskos,1 Eleanor M. Plaster,2 Claudia Loebel,3 Gizem Efe,1 Tatiana Karakasheva,4 Joel T. Gabre,1 Jianhua Hu,1 Jason A. Burdick,5 and Anil K. Rustgi1

1Herbert Irving Comprehensive Cancer Center, Division of Digestive and Liver Diseases, Department of Medicine, Vagelos College of Physicians and Surgeons, Columbia University Irving Medical Center, New York, New York, USA.

2Department of Biomedical Engineering and

3Department of Materials Science and Engineering, University of Michigan, Ann Arbor, Michigan, USA.

4Division of Gastroenterology, Hepatology, and Nutrition, Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA.

5BioFrontiers Institute and Department of Chemical and Biological Engineering, University of Colorado, Boulder, Colorado, USA.

Address correspondence to: Anil K. Rustgi, Herbert Irving Comprehensive Cancer Center, Columbia University Irving Medical Center, 1130 St. Nicholas Ave, New York, New York 10032, USA. Phone: 212.851.4822; Email: akr2164@cumc.columbia.edu.

Find articles by Cruz-Acuña, R. in: JCI | PubMed | Google Scholar

1Herbert Irving Comprehensive Cancer Center, Division of Digestive and Liver Diseases, Department of Medicine, Vagelos College of Physicians and Surgeons, Columbia University Irving Medical Center, New York, New York, USA.

2Department of Biomedical Engineering and

3Department of Materials Science and Engineering, University of Michigan, Ann Arbor, Michigan, USA.

4Division of Gastroenterology, Hepatology, and Nutrition, Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA.

5BioFrontiers Institute and Department of Chemical and Biological Engineering, University of Colorado, Boulder, Colorado, USA.

Address correspondence to: Anil K. Rustgi, Herbert Irving Comprehensive Cancer Center, Columbia University Irving Medical Center, 1130 St. Nicholas Ave, New York, New York 10032, USA. Phone: 212.851.4822; Email: akr2164@cumc.columbia.edu.

Find articles by Kariuki, S. in: JCI | PubMed | Google Scholar

1Herbert Irving Comprehensive Cancer Center, Division of Digestive and Liver Diseases, Department of Medicine, Vagelos College of Physicians and Surgeons, Columbia University Irving Medical Center, New York, New York, USA.

2Department of Biomedical Engineering and

3Department of Materials Science and Engineering, University of Michigan, Ann Arbor, Michigan, USA.

4Division of Gastroenterology, Hepatology, and Nutrition, Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA.

5BioFrontiers Institute and Department of Chemical and Biological Engineering, University of Colorado, Boulder, Colorado, USA.

Address correspondence to: Anil K. Rustgi, Herbert Irving Comprehensive Cancer Center, Columbia University Irving Medical Center, 1130 St. Nicholas Ave, New York, New York 10032, USA. Phone: 212.851.4822; Email: akr2164@cumc.columbia.edu.

Find articles by Sugiura, K. in: JCI | PubMed | Google Scholar |

1Herbert Irving Comprehensive Cancer Center, Division of Digestive and Liver Diseases, Department of Medicine, Vagelos College of Physicians and Surgeons, Columbia University Irving Medical Center, New York, New York, USA.

2Department of Biomedical Engineering and

3Department of Materials Science and Engineering, University of Michigan, Ann Arbor, Michigan, USA.

4Division of Gastroenterology, Hepatology, and Nutrition, Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA.

5BioFrontiers Institute and Department of Chemical and Biological Engineering, University of Colorado, Boulder, Colorado, USA.

Address correspondence to: Anil K. Rustgi, Herbert Irving Comprehensive Cancer Center, Columbia University Irving Medical Center, 1130 St. Nicholas Ave, New York, New York 10032, USA. Phone: 212.851.4822; Email: akr2164@cumc.columbia.edu.

Find articles by Karaiskos, S. in: JCI | PubMed | Google Scholar

1Herbert Irving Comprehensive Cancer Center, Division of Digestive and Liver Diseases, Department of Medicine, Vagelos College of Physicians and Surgeons, Columbia University Irving Medical Center, New York, New York, USA.

2Department of Biomedical Engineering and

3Department of Materials Science and Engineering, University of Michigan, Ann Arbor, Michigan, USA.

4Division of Gastroenterology, Hepatology, and Nutrition, Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA.

5BioFrontiers Institute and Department of Chemical and Biological Engineering, University of Colorado, Boulder, Colorado, USA.

Address correspondence to: Anil K. Rustgi, Herbert Irving Comprehensive Cancer Center, Columbia University Irving Medical Center, 1130 St. Nicholas Ave, New York, New York 10032, USA. Phone: 212.851.4822; Email: akr2164@cumc.columbia.edu.

Find articles by Plaster, E. in: JCI | PubMed | Google Scholar

1Herbert Irving Comprehensive Cancer Center, Division of Digestive and Liver Diseases, Department of Medicine, Vagelos College of Physicians and Surgeons, Columbia University Irving Medical Center, New York, New York, USA.

2Department of Biomedical Engineering and

3Department of Materials Science and Engineering, University of Michigan, Ann Arbor, Michigan, USA.

4Division of Gastroenterology, Hepatology, and Nutrition, Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA.

5BioFrontiers Institute and Department of Chemical and Biological Engineering, University of Colorado, Boulder, Colorado, USA.

Address correspondence to: Anil K. Rustgi, Herbert Irving Comprehensive Cancer Center, Columbia University Irving Medical Center, 1130 St. Nicholas Ave, New York, New York 10032, USA. Phone: 212.851.4822; Email: akr2164@cumc.columbia.edu.

Find articles by Loebel, C. in: JCI | PubMed | Google Scholar

1Herbert Irving Comprehensive Cancer Center, Division of Digestive and Liver Diseases, Department of Medicine, Vagelos College of Physicians and Surgeons, Columbia University Irving Medical Center, New York, New York, USA.

2Department of Biomedical Engineering and

3Department of Materials Science and Engineering, University of Michigan, Ann Arbor, Michigan, USA.

4Division of Gastroenterology, Hepatology, and Nutrition, Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA.

5BioFrontiers Institute and Department of Chemical and Biological Engineering, University of Colorado, Boulder, Colorado, USA.

Address correspondence to: Anil K. Rustgi, Herbert Irving Comprehensive Cancer Center, Columbia University Irving Medical Center, 1130 St. Nicholas Ave, New York, New York 10032, USA. Phone: 212.851.4822; Email: akr2164@cumc.columbia.edu.

Find articles by Efe, G. in: JCI | PubMed | Google Scholar

1Herbert Irving Comprehensive Cancer Center, Division of Digestive and Liver Diseases, Department of Medicine, Vagelos College of Physicians and Surgeons, Columbia University Irving Medical Center, New York, New York, USA.

2Department of Biomedical Engineering and

3Department of Materials Science and Engineering, University of Michigan, Ann Arbor, Michigan, USA.

4Division of Gastroenterology, Hepatology, and Nutrition, Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA.

5BioFrontiers Institute and Department of Chemical and Biological Engineering, University of Colorado, Boulder, Colorado, USA.

Address correspondence to: Anil K. Rustgi, Herbert Irving Comprehensive Cancer Center, Columbia University Irving Medical Center, 1130 St. Nicholas Ave, New York, New York 10032, USA. Phone: 212.851.4822; Email: akr2164@cumc.columbia.edu.

Find articles by Karakasheva, T. in: JCI | PubMed | Google Scholar

1Herbert Irving Comprehensive Cancer Center, Division of Digestive and Liver Diseases, Department of Medicine, Vagelos College of Physicians and Surgeons, Columbia University Irving Medical Center, New York, New York, USA.

2Department of Biomedical Engineering and

3Department of Materials Science and Engineering, University of Michigan, Ann Arbor, Michigan, USA.

4Division of Gastroenterology, Hepatology, and Nutrition, Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA.

5BioFrontiers Institute and Department of Chemical and Biological Engineering, University of Colorado, Boulder, Colorado, USA.

Address correspondence to: Anil K. Rustgi, Herbert Irving Comprehensive Cancer Center, Columbia University Irving Medical Center, 1130 St. Nicholas Ave, New York, New York 10032, USA. Phone: 212.851.4822; Email: akr2164@cumc.columbia.edu.

Find articles by Gabre, J. in: JCI | PubMed | Google Scholar

1Herbert Irving Comprehensive Cancer Center, Division of Digestive and Liver Diseases, Department of Medicine, Vagelos College of Physicians and Surgeons, Columbia University Irving Medical Center, New York, New York, USA.

2Department of Biomedical Engineering and

3Department of Materials Science and Engineering, University of Michigan, Ann Arbor, Michigan, USA.

4Division of Gastroenterology, Hepatology, and Nutrition, Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA.

5BioFrontiers Institute and Department of Chemical and Biological Engineering, University of Colorado, Boulder, Colorado, USA.

Address correspondence to: Anil K. Rustgi, Herbert Irving Comprehensive Cancer Center, Columbia University Irving Medical Center, 1130 St. Nicholas Ave, New York, New York 10032, USA. Phone: 212.851.4822; Email: akr2164@cumc.columbia.edu.

Find articles by Hu, J. in: JCI | PubMed | Google Scholar

1Herbert Irving Comprehensive Cancer Center, Division of Digestive and Liver Diseases, Department of Medicine, Vagelos College of Physicians and Surgeons, Columbia University Irving Medical Center, New York, New York, USA.

2Department of Biomedical Engineering and

3Department of Materials Science and Engineering, University of Michigan, Ann Arbor, Michigan, USA.

4Division of Gastroenterology, Hepatology, and Nutrition, Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA.

5BioFrontiers Institute and Department of Chemical and Biological Engineering, University of Colorado, Boulder, Colorado, USA.

Address correspondence to: Anil K. Rustgi, Herbert Irving Comprehensive Cancer Center, Columbia University Irving Medical Center, 1130 St. Nicholas Ave, New York, New York 10032, USA. Phone: 212.851.4822; Email: akr2164@cumc.columbia.edu.

Find articles by Burdick, J. in: JCI | PubMed | Google Scholar

1Herbert Irving Comprehensive Cancer Center, Division of Digestive and Liver Diseases, Department of Medicine, Vagelos College of Physicians and Surgeons, Columbia University Irving Medical Center, New York, New York, USA.

2Department of Biomedical Engineering and

3Department of Materials Science and Engineering, University of Michigan, Ann Arbor, Michigan, USA.

4Division of Gastroenterology, Hepatology, and Nutrition, Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA.

5BioFrontiers Institute and Department of Chemical and Biological Engineering, University of Colorado, Boulder, Colorado, USA.

Address correspondence to: Anil K. Rustgi, Herbert Irving Comprehensive Cancer Center, Columbia University Irving Medical Center, 1130 St. Nicholas Ave, New York, New York 10032, USA. Phone: 212.851.4822; Email: akr2164@cumc.columbia.edu.

Find articles by Rustgi, A. in: JCI | PubMed | Google Scholar

Published October 3, 2023 - More info

Published in Volume 133, Issue 23 on December 1, 2023
J Clin Invest. 2023;133(23):e168146. https://doi.org/10.1172/JCI168146.
© 2023 Cruz-Acuña et al. This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Published October 3, 2023 - Version history
Received: December 16, 2022; Accepted: September 28, 2023 View PDF Abstract

Increased extracellular matrix (ECM) stiffness has been implicated in esophageal adenocarcinoma (EAC) progression, metastasis, and resistance to therapy. However, the underlying protumorigenic pathways are yet to be defined. Additional work is needed to develop physiologically relevant in vitro 3D culture models that better recapitulate the human tumor microenvironment and can be used to dissect the contributions of matrix stiffness to EAC pathogenesis. Here, we describe a modular, tumor ECM–mimetic hydrogel platform with tunable mechanical properties, defined presentation of cell-adhesive ligands, and protease-dependent degradation that supports robust in vitro growth and expansion of patient-derived EAC 3D organoids (EAC PDOs). Hydrogel mechanical properties control EAC PDO formation, growth, proliferation, and activation of tumor-associated pathways that elicit stem-like properties in the cancer cells, as highlighted through in vitro and in vivo environments. We also demonstrate that the engineered hydrogel serves as a platform for identifying potential therapeutic targets to disrupt the contribution of protumorigenic matrix mechanics in EAC. Together, these studies show that an engineered PDO culture platform can be used to elucidate underlying matrix-mediated mechanisms of EAC and inform the development of therapeutics that target ECM stiffness in EAC.

Graphical Abstract Introduction

Over the past 30 years, the incidence of esophageal adenocarcinoma (EAC) has risen dramatically, by 300% to 600% in the US (1). Previous studies have demonstrated that changes in the tumor microenvironment involving a stiffened extracellular matrix (ECM) are associated with EAC progression (2–6). Although clinical observations suggest that increased ECM stiffness drives cell transformation, cancer progression, and metastasis, the underlying pathways of mechanotransduction that lead cancer cells to translate mechanical signals into intracellular protumorigenic pathways are yet to be defined (5). Therefore, additional work is needed to develop physiologically relevant 3D culture models that better recapitulate the human tumor microenvironment and can dissect the contributions of matrix properties to elucidate underlying molecular mechanisms of the disease (7).

Patient-derived tumor organoids have become attractive preclinical models for studying cancer biology, as they retain the biological characteristics of the primary tumor (7–9). Indeed, our lab has shown that patient-derived EAC 3D organoids (EAC PDOs) can serve as avatars for studying cellular responses to anticancer drugs, as they recapitulate patients’ drug responses in the clinic (10, 11). Patient-derived organoids (PDOs) are traditionally grown in Matrigel, a heterogeneous, complex mixture of ECM proteins, proteoglycans, and growth factors secreted by Engelbreth-Holm-Swarm mouse sarcoma cells (12). However, Matrigel suffers from lot-to-lot compositional and structural variability and cannot recapitulate the independent role of matrix properties in disease progression due to the inability to uncouple its physicochemical properties (12–14). For instance, making changes to the bulk concentration (e.g., decrease in matrix density) of Matrigel is a common approach for varying its mechanical properties; however, these changes unavoidably alter other matrix properties, such as adhesive ligand density and fiber density/structure (14) (Supplemental Figure 1A; supplemental material available online with this article; https://doi.org/10.1172/JCI168146DS1). Therefore, although modulation of the bulk concentration of Matrigel results in changes in EAC PDO formation (density), growth (area), and transcriptional expression of EAC-associated genes (Supplemental Figure 1, B–D), it is unclear whether this effect is mediated by differences in mechanical or biochemical matrix properties. To address this important gap, well-defined engineered hydrogels are an evolving and important component of tumor PDO culture systems as alternatives to Matrigel, particularly for introducing user-defined microenvironment signals for studying human epithelial tumors (15–19).

Here, we describe a modular, tumor ECM–mimetic hydrogel platform with defined physicochemical properties that support EAC PDO culture and growth. Hydrogel mechanical properties, adhesive ligand presentation, and protease-dependent degradation were key parameters in engineering a hydrogel that supported EAC PDO viability and growth. Particularly, hydrogel mechanical properties controlled EAC PDO formation and growth and activation of tumor-associated pathways. For instance, we showed that increased matrix mechanics enable upregulation of the Yes-associated protein 1 (Yap)/SRY-box transcription factor 9 (Sox9) axis, eliciting stem-like properties in the EAC PDOs and further elucidating an underlying molecular mechanism of the disease. Additionally, the engineered hydrogel served as a platform for identifying potential therapeutic targets for disrupting the contribution of protumorigenic increased matrix mechanics in EAC. Whereas previous work has established engineered hydrogels as tumor ECMs for investigating multicellular assembly and tumor invasion using cancer cell lines (20–22) or studying tumor PDO resistance to therapy (19, 23), we are the first, to our knowledge, to analyze the contributions of ECM mechanical properties to EAC PDO growth, proliferation, and identification of matrix mechanics–mediated drivers of stem-like properties as therapeutic targets through in vitro and in vivo models. Finally, the modular nature of the engineered hydrogel platform allows for potential adaptation to the culture of 3D organoid models of other human cancers. Thus, we provide mechanistic and translational insights with broad applicability.

Results

Engineered hydrogel supports EAC PDO development. We selected a hydrogel platform based on hyaluronic acid (HA), specifically through the crosslinking of norbornene-functionalized HA (NorHA) macromer (Supplemental Figure 2A), which exhibits native biofunctionality and has been extensively developed for in vitro cell activation by stiffening events and several preclinical in vivo applications (24). HA has inherent biological importance due to its binding to cell receptors (e.g., CD44; refs. 25, 26) and is a major component of the tumor niche (Figure 1, A and B), creating a microenvironment that is favorable for tumor angiogenesis, invasion, and metastasis (27, 28). Certainly, EAC patient biopsies showed increased expression of HA in the tumor microenvironment (Figure 1A), whereas RNA-Seq analysis of 286 esophageal carcinoma (ESCA) tissues collected from The Cancer Genome Atlas (TCGA) and The Genome-Tissue Expression Project (GTEx) (29) confirmed increased expression of HA synthesis genes (HAS1, HAS2, UDGH) as compared with 283 normal esophageal tissues (Figure 1B). Moreover, our hydrogel system offers marked advantages due to its well-defined structure, covalent incorporation of peptide sequences for enhanced cell/matrix interactions, and user-defined hydrogel stiffness, which is controlled by varying the crosslinking peptide concentration, which mediates crosslinking via a thiol-norbornene reaction to form a NorHA hydrogel (Figure 1C and Supplemental Figure 2A) (30). Indeed, HA-based hydrogels have been used by other groups to study tumor progression and resistance to therapy of other cancer types (e.g., colorectal and pancreatic adenocarcinomas) (31, 32). We added to this model system tunable matrix properties that enable an advancement in studying ECM-mediated tumor progression using PDOs.

Figure 1

Engineered hydrogel supports EAC PDO development. (A) Images of patient tissue sections from normal and EAC biopsies stained for HA (HABP), CD44, or fibronectin (FN1). Scale bar: 100 μm. Original magnification, ×5 (insets). (B) Bulk RNA-Seq analysis of ESCA and normal pancreatic tissue samples for fibronectin and hyaluronan-associated genes (HAS1, HAS2, UGDH). n = 286 for ESCA; n = 283 for normal. (C) Relationship between crosslinker density (mg/mL) and storage modulus, G′ (mean ± SEM; n = 3 independently prepared hydrogels per condition). (D) Quantification of PDO viability as assessed by calcein-AM labeling at 7 days after encapsulation. Viability is quantified as the percentage of PDOs that stained positive for calcein-AM (mean ± SEM; n = average number of calcein-AM+ organoids per hydrogel; at least 20 organoids per hydrogel were analyzed). Welch’s t test with 2-tailed comparison showed no significant differences between groups. NS = P > 0.05. (E) Representative transmitted light and fluorescence microscopy images of EAC PDOs cultured in NorHA hydrogels or Matrigel. Scale bar: 200 μm. (F) Representative fluorescence microcopy images of EAC PDOs within NorHA hydrogels stained for MUC5ac, E-cad, and CK8 at 14 days after encapsulation. Scale bar: 50 μm. Three independent experiments were performed, and data are presented for 1 of the experiments. Every independent experiment was performed with 4 gel samples per experimental group.

In our study, we explored a NorHA hydrogel formulation that supports the viability of EAC PDOs generated in Matrigel. After EAC PDOs were grown in Matrigel, they were retrieved, dissociated into single cells, and encapsulated in NorHA hydrogels with mechanical properties similar to those of Matrigel (G′ = 100 Pa; Figure 1C and Supplemental Figure 2B). NorHA hydrogels were engineered to present a constant 2.0 mM RGD adhesive peptide (GCGYGRGDSPG) density and crosslinked with the protease-degradable peptide VPM (0.5 mg/mL; GCNSVPMSMRGGSNCG). Incorporation of 2.0 mM RGD adhesive peptide and a protease-degradable crosslinker to engineered hydrogels has previously promoted epithelial organoid development (13, 33, 34). Moreover, the RGD adhesive ligand (α5β1 and ανβ3 integrin–binding peptide) is found in many adhesive proteins, including fibronectin, a major ECM protein component in EAC (2, 35). Indeed, EAC patient biopsies showed increased expression of fibronectin as compared with normal tissue (Figure 1A), and RNA-Seq analysis from TCGA and GTEx (29) confirmed increased expression of fibronectin (FN1) in ESCA as compared with normal esophageal tissues (Figure 1B). EAC PDOs grown in NorHA hydrogels functionalized with RGD and crosslinked with VPM demonstrated viability comparable to those grown in Matrigel (Figure 1, D and E). However, when EAC PDOs were grown in NorHA hydrogels presenting an inactive scrambled peptide (RDG) or functionalized with RGD and crosslinked with nondegradable agent 1,4-dithiothreitol (DTT), EAC PDOs showed reduced viability at 7 days after encapsulation, as compared with hydrogels functionalized with RGD and crosslinked with VPM (Supplemental Figure 3, A–C). Moreover, PDOs cultured in the engineered hydrogel formulation showed expression of the epithelial marker E-cadherin (E-cad) and of EAC-specific markers mucin 5AC (MUC5ac) and cytokeratin 8 (CK8) at levels similar to those of PDOs cultured in Matrigel (Figure 1F). Finally, to determine whether NorHA hydrogels are suitable for the culture of other human PDOs, we embedded Barrett’s esophagus (a precursor or premalignant condition that predisposes to EAC) PDOs (BE PDOs) (36) in the engineered NorHA hydrogel (Supplemental Figure 3, D–F). BE PDOs cultured in the engineered matrix maintained high viability and comparable growth and formation (density) compared with BE PDOs cultured in Matrigel (Supplemental Figure 3, D–F). Taken together, these data suggest the requirement of specific matrix properties that are essential for organoid viability and formation, establishing the engineered NorHA hydrogel as a culture system for EAC PDOs that has the potential to be adapted for the generation of different human tumor organoids.

Matrix mechanics control EAC PDO development. As ECM mechanical properties influence epithelial cell behavior (33, 37), we investigated the influence of crosslinker density, which controls hydrogel mechanical properties (Figure 1C), on EAC PDO size, formation, and proliferation (Figure 2). Hydrogels were engineered to present constant NorHA macromer and adhesive ligand densities, but with varying crosslinker densities. EAC PDOs were embedded in NorHA hydrogels with mechanical properties that ranged from a “soft” hydrogel (0.5 mg/mL VPM; G′ = 100 Pa, similar to Matrigel) to a “stiff” hydrogel (1.2 mg/mL; G′ = 1000 Pa, similar to tumor ECM) and cultured for 14 days. Importantly, the mechanical properties of our stiff hydrogel have been shown to promote protumorigenic behavior in normal cells in previous in vitro studies (38, 39) and compare favorably with measurements of human tumor stiffness (G′ ≥ 1,000 Pa) (38, 40–42). EAC PDOs embedded in stiff (G′: 1,000 Pa) NorHA hydrogels showed significant increases in the size (area) and formation (density) of organoids formed per hydrogel as a function of matrix stiffness (Figure 2, A–C). Similarly, PDOs embedded in the stiff hydrogel condition showed increased cell proliferation, as compared with organoids embedded in the soft hydrogel condition (Figure 2, C and D). Interestingly, when we embedded cells in “stiffer” NorHA hydrogels, namely of G′ = 1,800 Pa and 2,800 Pa, the organoids showed significant reduction in formation, growth, and cell proliferation or no instances of organoid formation, respectively, as compared with the stiff (G′ = 1,000 Pa) NorHA hydrogel (Supplemental Figure 4). Moreover, the engineered hydrogel was able to support culture and expansion of 3 different EAC PDO lines for at least 3 passages (~1.5 months), and after the long-term culture, all EAC PDO lines maintained a significant increase in PDO formation (density), size (area), and cell proliferation as a function of matrix stiffness (Supplemental Figure 5). Together, these data demonstrate that the NorHA hydrogel supports robust long-term in vitro culture and expansion of PDOs and that a restricted range of matrix stiffness (G′ = 1,000 Pa) nurtures EAC PDO growth, formation, and proliferation. These observations establish the engineered hydrogel system as an innovative platform for investigating the independent contributions of matrix mechanics in EAC PDO development.

Figure 2

Engineered hydrogel stiffness modulates EAC PDO development. (A) Quantification of PDO (A) size (area) and (B) density as a function of matrix stiffness at 14 days after encapsulation. Data are represented as mean ± SEM. (A) n = at least 300 organoids analyzed across 4 hydrogels per group; (B) n = 4 hydrogels per group. (A and B) Kruskal-Wallis test with Dunn’s multiple-comparisons test showed significant differences between 100 Pa and 350 Pa or 1000 Pa. (C) Representative transmitted light and fluorescence microscopy images of EAC PDOs and (D) quantification of proliferating cells (%Ki67+) in EAC PDOs cultured in NorHA hydrogels of different stiffnesses at 14 days after encapsulation. Data are represented as mean ± SEM. n = at least 15 organoids analyzed per group. Kruskal-Wallis test with Dunn’s multiple-comparisons test showed significant differences between 100 Pa and 1000 Pa. Scale bar: 100 μm. (A–D) Three independent experiments were performed, and data are presented for 1 of the experiments. Every independent experiment was performed with 4 gel samples per experimental group. *P < 0.05; **P < 0.01.

Matrix mechanics modulates YAP activation in EAC PDOs. Recent work has demonstrated that dysregulated Yap activation is essential for the growth of most solid tumors, acting by inducing cancer stem cell features, proliferation, and metastasis (43–45). Dysregulated Yap activation is a major determinant of stem cell properties by direct upregulation of SOX9 (46, 47) in EAC. However, the pathophysiological event or events that elicit upregulation of the Yap/Sox9 axis in EAC remain elusive. Therefore, as Yap functions as a sensor of the structural and mechanical features of the cellular microenvironment, we investigated whether changes in matrix biomechanics played a role in the expression of Yap and Sox9 in EAC PDOs. EAC PDOs embedded in the stiff hydrogel showed a significant increase in the expression and nuclear localization of Yap as well as SOX9 expression as compared with the 3D organoids within the softer hydrogels or Matrigel (Figure 3, A and B, Supplemental Figure 4B, and Supplemental Figure 6, A and B). This phenotype was also observed in our 3 EAC PDO lines after long-term culture (Supplemental Figure 5, D–F). Moreover, the protein expression of the esophageal cancer putative stem cell marker CD44 (48, 49) was significantly higher in EAC PDOs within the stiff hydrogel as compared with 3D organoids within the softer hydrogels (Figure 3C). These data suggest that matrix mechanics induce aberrant activation of the Yap/Sox9 axis, endowing stem-like properties to the EAC PDOs, as evidenced by increased PDO growth, formation, cell proliferation, and CD44 expression in 3D organoids embedded in the stiff (G′ = 1,000 Pa) NorHA hydrogel.

Figure 3

Engineered hydrogel stiffness modulates YAP activation in EAC PDOs. (A) Transcriptional expression and (B) representative fluorescence images of YAP and SOX9 in organoids within NorHA hydrogels of different stiffness at 14 days after encapsulation. Data are represented as mean ± SEM. n = 3 technical replicates, representative of 3 independent experiments. (A) One-way ANOVA with Tukey’s multiple comparisons test showed significant differences between 100 Pa and 1000 Pa, 350 Pa and 1000 Pa, and 1000 Pa and Matrigel. *P < 0.05; **P < 0.01; ***P < 0.001. RNA levels normalized to 100 Pa. Scale bars: 100 μm. Original magnification, ×5 (insets, 100 Pa and 350 Pa); ×3 (insets, 1,000 Pa). (C) Quantification and representative fluorescence microscopy images of CD44 expression in EAC PDOs cultured in NorHA hydrogels of different stiffness at 14 days after encapsulation. Data are represented as mean ± SEM. n = at least 20 organoids analyzed per group. Kruskal-Wallis test with Dunn’s multiple-comparisons test showed significant differences between 100 Pa and 1000 Pa, and 350 Pa and 1000 Pa. *P < 0.05; ****P < 0.0001. Scale bar: 100 μm. (D) Schematic of in vitro experiment of EAC PDOs within NorHA hydrogels being treated with YAP inhibitor verteporfin. Created with BioRender.com. (E) Representative immunohistochemistry microscopy images and quantification of Yap expression in EAC PDOs cultured in NorHA hydrogels of different stiffnesses at 7 days after encapsulation and treated with 5 nM verteporfin or DMSO. Data are represented as mean ± SEM. n = 4 hydrogels per group. One-way ANOVA with Tukey’s multiple-comparisons test showed significant differences between 1000 Pa+DMSO and every other group (**P < 0.01), and no significant differences among other groups (P > 0.05). Scale bar: 100 μm. (A–E) Three independent experiments were performed, and data are presented for 1 of the experiments. Every independent experiment was performed with 4 gel samples per experimental group. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

The expression of other EAC-associated genes, TP53 and STAT3, significantly increased in EAC PDOs within the stiff hydrogel as compared with the 3D organoids within the soft hydrogel (Supplemental Figure 6, C–E). Interestingly, whole-exome sequencing (WES) of the EAC PDOs cultured in NorHA hydrogels of different stiffness (100 or 1000 Pa) or Matrigel revealed the presence of a TP53 gene Pro72Arg (rs1042522) single-nucleotide polymorphism, which is located in the p53 proline-rich domain and has been associated with increased tumor metastasis (Supplemental Figure 6F) (50). Further analysis revealed that the overall mutational profile of the PDOs, as well as the type and number of TP53 variations or mutations, did not change as a function of matrix type (Matrigel versus NorHA) or matrix stiffness (100 Pa versus 1000 Pa NorHA) within a 14-day time period (Supplemental Figure 6, F and G). These data demonstrate that the engineered hydrogel can serve as a platform for studying the influence of matrix mechanics in the activation of regulatory mechanisms in EAC PDOs. Surprisingly, bulk RNA-Seq of EAC PDOs did not reveal differentially expressed genes in the 3D organoids within the stiff hydrogel as compared with 3D organoids within the soft hydrogel. However, when comparing EAC PDOs within the stiff hydrogel versus Matrigel, analyses revealed 424 differentially expressed genes (224 upregulated and 200 downregulated genes). While YAP was upregulated, this was not statistically significant (Supplemental Figure 7A). Among the top 10 significantly upregulated genes, 5 have been associated with tumor progression, metastasis, or recurrence (CALB2, ref. 51, ECM1, refs. 52, 53; TNIK, ref. 54; IGFBP4, ref. 55; and TCN1, ref. 56); Supplemental Figure 7B), including TNIK (57) and IGFBP4 (58), which are reported downstream targets of Yap. Finally, gene ontology of up- and downregulated genes showed enrichment of biological processes that have been associated with tumor pathobiology (Supplemental Figure 7C).

Another advantage of HA-based hydrogels is their permissive diffusional properties, which allow diffusion of small molecules, including drugs and inhibitors, to cells (Figure 3D) (24, 59). Therefore, we investigated to determine whether the effect of increased matrix stiffness on EAC PDOs was repressed via inhibition of the nuclear translocation of Yap using verteporfin, a commercially available small molecule inhibitor whose efficacy and potential as therapy has been described previously (60, 61). Addition of verteporfin to the cell culture media of EAC PDOs grown in NorHA hydrogels (Figure 3D) resulted in significant reduction in Yap expression and the formation (density) and size of organoids within the stiff hydrogel, as compared with the vehicle control (DMSO) (Figure 3E and Supplemental Figure 8A). However, EAC PDOs embedded in the soft hydrogel and exposed to verteporfin showed no significant differences in Yap expression, organoid formation, and size as compared with vehicle control (Figure 3E and Supplemental Figure 8A). Similarly, introduction of YAP siRNA via lipofection to the EAC PDOs grown in the stiff NorHA hydrogel resulted in significant reduction in 3D organoid formation (density), size (area), Yap and Sox9 expression, and Yap nuclear localization as compared with the control siRNA (Supplemental Figure 8, B–E). However, EAC PDOs embedded in the soft hydrogel with the addition of Yap siRNA showed no significant differences in 3D organoid formation, size, Yap and Sox9 expression, and Yap nuclear localization as compared with control siRNA (Supplemental Figure 8, B–E). These complementary data further underscore that matrix mechanics modulate the activation of the Yap/Sox9 axis. Together, these data elucidate a mechanism showing how matrix mechanics influence EAC pathogenesis and nominate YAP as a potential therapeutic target in this context.

Matrix mechanics control EAC PDO development and YAP activation in vivo. Engineered hydrogels have been utilized previously as organoid delivery vehicles (24, 34). Therefore, we embedded organoids in our engineered tumor ECM-mimetic hydrogels and transplanted into dorsal subcutaneous spaces of immunocompromised mice to study the effects of matrix mechanics (Figure 4A). After 4 weeks, PDOs showed expression of the epithelial marker E-cad and of EAC-specific markers CK8 and MUC5ac (Figure 4B). However, implanted stiff hydrogels contained significantly larger (area) EAC PDOs as compared with organoids within soft hydrogels (Figure 4C). Additionally, PDOs embedded in the stiff hydrogel showed increased cell proliferation, Sox9 expression, and nuclear localization of YAP, as compared with organoids embedded within the soft hydrogel (Figure 4, D and E). These data suggest that matrix mechanics control PDO growth and proliferation via dysregulated activation of the Yap/Sox9 axis in an in vivo environment. Finally, as studies suggest that increased ECM stiffness stabilizes mutant p53 (62) and we have previously shown that mutant p53-Yap interactions promote esophageal cancer progression (63), we investigated p53 expression in EAC PDOs within NorHA hydrogels. We observed that EAC PDOs within the stiff hydrogel show increased nuclear p53 localization as compared with organoids within the soft hydrogel (Supplemental Figure 6H). Interestingly, we did not observe a significant change in Stat3 expression as a function of matrix stiffness in vivo (Supplemental Figure 6I). Together, these data further underscore that matrix mechanics modulate the dysregulated activation of Yap/Sox9 and potentially the expression of other EAC-associated proteins, elucidating underlying mechanisms of EAC in the context of matrix mechanics using an engineered in vivo model.

Figure 4

Engineered hydrogel stiffness-dependent growth of EAC PDOs in in vivo xenograft model. (A) Schematic of in vivo transplantation experiment of EAC PDOs within NorHA hydrogels into mouse subcutaneous pockets. Created with BioRender.com. (B) Representative fluorescence microcopy images of EAC PDOs within NorHA hydrogels stained for MUC5ac, E-cad, and CK8 at 28 days after encapsulation and in vivo transplantation. Scale bar: 100 μm. (C) Histological (H&E) microcopy images and quantification of PDO size (area) as a function of matrix stiffness at 28 days after encapsulation and in vivo transplantation. Data are represented as mean ± SEM. n = at least 5 organoids analyzed per group. Welch’s t test with 2-tailed comparison showed significant differences between 100 Pa and 1000 Pa. **P < 0.01. Scale bar: 100 μm. (D) Quantification and representative fluorescence microscopy images of percentages of proliferating cells (%Ki67+) per EAC PDOs as a function of matrix stiffness at 28 days after encapsulation and in vivo transplantation. Data are represented as mean ± SEM. n = at least 6 organoids analyzed per group. Mann-Whitney U test showed significant differences between 100 Pa and 1000 Pa. ***P < 0.001. Scale bar: 100 μm. (E) Representative fluorescence microcopy images of EAC PDOs within NorHA hydrogels stained for Sox9 and Yap at 28 days after encapsulation and in vivo transplantation. Quantification of percentage of nuclear Yap+ cells (%Yap+) per EAC PDO as a function of matrix stiffness. Data are represented as mean ± SEM. n = 6 organoids analyzed per group. Welch’s t test with 2-tailed comparison showed significant differences between 100 Pa and 1000 Pa. ***P < 0.001. Scale bar: 100 μm. Original magnification, ×5 (insets). (A–E) Two independent experiments were performed, and data are presented for 1 of the experiments. Every independent experiment was performed with 2 gels per mouse and 5 mice per experimental group.

Yap inhibition represses the effect of matrix mechanics in EAC PDOs in vivo. Patient-derived xenograft (PDX) models have proven to be highly effective in predicting the efficacy of both conventional and novel anticancer therapeutics