Background
In the past few years, there has been an increased interest in patient-derived organoids (PDOs) derived from diseased human tissues. In contrast to the in vivo models of patient-derived mouse xenograft (PDX) that require large amounts of specimens and normally 4–8 months for development [
1,
2], PDOs can be cultured from patient materials and expanded with high efficiency within 1 month. Although immortalized cell lines have proven valuable in the study of tumor pathogenesis, these in vitro models have the obvious drawbacks of bearing little resemblance to the parental tumors [
3]. PDOs have been applied to model various cancers, e.g, colorectal [
4], prostate [
5], lung [
6], liver [
7], breast [
8], pancreatic [
9], endometrial [
10], bladder [
11], ovarian [
12], esophagus [
13], and gastrointestinal [
14] cancers. PDOs have been used as in vitro disease models that recapitulate the pathological characteristics, genetic alterations, and heterogeneity of their corresponding primary tumors and can potentially serve as “avatars” for selecting clinical therapeutic regimen. Several studies have demonstrated that PDOs are able to precisely predict patient responses to drug and radiation treatments in the clinic [
14‐
20].
Thyroid cancer has become the most common endocrine malignancy with an increasing incidence in recent years [
21]. Histopathologically, thyroid cancer can be stratified into four categories: papillary thyroid cancer (PTC, 80–85%), follicular thyroid cancer (FTC, 10–15%), anaplastic thyroid cancer (ATC, 2–3%), and medullary thyroid cancer (MTC, 2–3%) [
22]. PTC and FTC are collectively referred to as differentiated thyroid cancer (DTC). Among all types of thyroid cancer, PTC is the most common histologic type, and is characterized by the most frequent mutation of BRAF
V600E [
22,
23]. This mutation leads to phosphorylation of the downstream mitogen-activated protein/extracellular signal-regulated kinase kinase (MEK) and extracellular signal-regulated kinase (ERK), resulting in malignant transformation and potential loss of differentiated functions [
24]. Despite an excellent prognosis in the majority patients with PTC, about 20% of the cases develop regional recurrence or distant metastasis, and more than half will not respond to conventional therapy such as postoperative thyroid-stimulating hormone suppression and radioactive iodine (RAI) treatments [
25]. Poor prognosis has been reported in these patients [
26].
Targeted therapy, especially inhibition of BRAF
V600E, may be an effective strategy for the treatment of metastatic PTC harboring this mutation. In fact, BRAF inhibitors have been already employed in the treatment of multiple cancers harboring the BRAF
V600E mutation. However, the therapeutic efficacies of these inhibitors vary from excellent responses in some cancers to drug resistance/tumor recurrence in others [
27]. For instance, the therapeutic responses of BRAF-mutant cancers to these inhibitors ranged from a response rate of 48% in melanoma to 5% in colorectal cancer [
28,
29]. In a phase I trial, treatment with the selective BRAF inhibitor vemurafenib in 3 patients with metastatic BRAF
V600E-mutant PTC yielded a partial response in one and prolonged stabilization of disease in the others [
30]. Several other Phase I and II studies using BRAF-inhibitors (dabrafenib and vemurafenib) have shown anti-tumor activity in a portion of patients with progressive BRAF
V600E-mutant PTC [
31‐
33]. However, small sample size and limited length of follow-up have made it difficult to predict which therapies are best suited for each patient, and which patients would likely best respond to the treatment.
Combination use of BRAF and MEK inhibitors has become a standard therapeutic approach in patients carrying a BRAF
V600E activating mutation. As for the observed effect of the combination therapy, BRAF and MEK inhibitors have shown significant improvements in clinical outcomes in BRAF
V600E-mutant melanoma [
34‐
40], non-small-cell lung cancer [
41], and ATC [
42]. However, little is known about the therapeutic effects of BRAF and MEK inhibitor combinations in advanced and metastatic BRAF
V600E mutation-harboring PTC. Also, there is as yet no way to clinically predict the therapeutic efficacies of BRAF-mutated cancers to BRAF and MEK inhibitors. Patient-derived organoids may provide a potential platform to test combination therapies aimed at finding drug synergisms and predicting therapeutic effect.
Previously, we have developed PTC organoids from human tumors and found that these models can accurately recapitulate the histological and genetic features of disease in vitro [
43]. The robust organoid models derived from PTC tissues have potential to be used to aid in the selection of optimal anticancer drugs for individual patients. In our study, we established patient-derived PTC organoid models that recapitulate respective tumor characteristics. BRAF inhibitor monotherapy showed moderate sensitivity in treating BRAF
V600E-mutant PTC organoids. The combination therapy of BRAF inhibitors with MEK inhibitors, receptor tyrosine kinases (RTK) inhibitors, or chemotherapeutic drugs significantly suppressed the growth of PTC organoids. Our study suggests that PTC organoids may be a potential preclinical tool used for the selection of drug treatment regimens for thyroid cancer patients.
Methods
Patient and sample collection
Papillary thyroid cancer tissues were gathered between August 2021 and June 2022 at Peking University Shenzhen Hospital, China. This study was approved by the Human Ethical Committee of the Hospital (Approval No. 2019-024), and written informed consent was signed prior to acquisition of samples from all patients involved. Patients’ characteristics are described in Table
1. Each tissue sample underwent histological assessment by two senior pathologists. The resected tumor tissues were placed into ice-cold Advanced DMEM/F12 medium (Cat. No. 12634-010, Gibco) and shipped to the laboratory on ice within 2 h of the surgery, for immediate further processing.
Table 1
Clinical characteristics of patients with papillary thyroid cancer
PTC-1 | Female | 60 | 1.0 × 0.7 × 0.3 | + | T1 | N0 | I |
PTC-2 | Female | 33 | 2.2 × 1.4 × 1.4 | + | T2 | N1a | I |
PTC-3 | Male | 34 | 1.9 × 1.5 × 1.2 | + | T1b | N1a | I |
PTC-4 | Male | 33 | 3.0 × 2.4 × 1.6 | + | T3 | N0 | I |
PTC-5 | Male | 20 | 1.9 × 1.6 × 1.0 | + | T1b | N1b | I |
PTC-6 | Female | 28 | 3.5 × 2.3 × 1.8 | − | T2 | N1b | I |
PTC-7 | Female | 31 | 2.5 × 2.0 × 2.0 | − | T2 | N0 | I |
PTC-8 | Female | 27 | 2.1 × 1.6 × 1.5 | − | T1b | N1a | I |
PTC-9 | Male | 32 | 3.5 × 3.0 × 3.0 | − | T2 | N1 | I |
Organoid culture
Detailed procedures for PTC organoid derivation have been described previously by our group [
43]. Briefly, the minced tumor tissues were digested with 5 mg/mL collagenase type II (Cat. No. 17101-015, Gibco) in the presence of Y-27632 dihydrochloride (10 µM, Cat. No. M1817, Abmole) for 40 min in a 37 ℃ shaking water bath. The digested tissues were further digested with 5 mL TrypLE Express (Cat. No. 12605-010, Gibco) for 5 min at 37 °C, strained over a 70 µm filter (Cat. No. 258368, NEST Biotechnology), and embedded in Matrigel (Cat. No. 356231, Corning). Tumor organoids were cultured in complete medium consisting of Advanced DMEM/F12 medium supplemented with 1% HEPES (Cat. No. 15630-080, Gibco), 1% GlutaMAX (Cat. No. 3505-0061, Gibco), 1% antibiotic–antimycotic (Cat. No. 15240-062, Gibco), 1 × B27 (Cat. No. 17504-044, Gibco), 1.25 mM N-acetyl-L-cysteine (Cat. No. A9165, Sigma-Aldrich), 10 mM Nicotinamide (Cat. No. N0636, Sigma-Aldrich), 500 ng/mL R-spondin-1 (Cat. No. 120-38, Peprotech), 100 ng/mL Noggin (Cat. No. 120-10C-1000, Peprotech), 5 ng/mL FGF-7 (Cat. No. 100-19, Peprotech), 10 ng/mL FGF-10 (Cat. No. 100-26, Peprotech), 50 ng/mL EGF (Cat. No. AF-100-15, Peprotech), 500 nM A83-01 (cat. no. SML0788, Sigma-Aldrich), and 10 μM SB202190 (Cat. No. S7067, Sigma-Aldrich). The medium was changed every 3–4 days. The organoids were visualized under a Carl Zeiss microscope (AXIO OBSERVER 3, Germany).
For passaging, growth medium was removed, and 3 mL of TrypLE Express was added to the Matrigel-cell suspension droplets and incubated at 37 °C for 3–5 min. Following mechanical blowing, AdDMEM/F12 containing 10% FBS would be added. Then, the suspension was centrifuged at 300 ×g for 5 min. PTC organoids were passaged at a 1:2–1:4 dilution every 1–4 weeks. For freezing, PTC organoids were resuspended in Recovery Cell Culture Freezing Medium (Cat. No. 12648-010, Gibco), cooled, and stored in liquid nitrogen. When required, organoids were thawed using standard thawing procedures, embedded in Matrigel and cultured as described above.
To compare the PTC organoid-forming efficiency, 1000 cells were seeded into a 24-well plate, and overlaid with organoid culture medium. Organoid numbers were counted with a light microscope after 10 days in culture.
Histology and immunostaining
Tissues and organoids were fixed in 4% paraformaldehyde for 24 h, followed by dehydration, paraffin embedding, and serial sectioning at 4 μm. Tissues and organoids sections were subjected to hematoxylin and eosin (H&E) and immunofluorescence analysis using standard procedures. The histological diagnosis was made according to the standard classification. For immunofluorescence analysis, sections were boiled for 30 min in EDTA solution (pH 8.0) for antigen-retrieval, and blocked in 5% BSA blocking buffer for 30 min to reduce nonspecific staining. Primary antibodies against CK19 (1:200, Cat. No. Kit-0030, Maixin Biotech, China), galectin-3 (1:500, Cat. No. ab76245, Abcam), Ki-67 (1:200, cat. no. ab16667, Abcam), and BRAF (V600E mutant) (1:200, Cat. No. 29002, Cell Signaling Technology) were applied to the sections and incubated overnight at 4 °C. After being washed with PBS (3 × 10 min), the sections were subsequently incubated with secondary antibody, Cy3-labeled goat anti-mouse IgG (H + L) (1:300, Cat. No. A0521, Beyotime), Cy3-labeled goat anti-rabbit IgG (H + L) (1:300, cat. no. A0516, Beyotime, China), or Alexa Flour 488-labeled goat anti-mouse IgG (H + L) (1:300, Cat. No. A0428, Beyotime) for 1 h at room temperature. Nuclei were counterstained with DAPI (Cat. No. C1002, Beyotime) for 10 min and cover-slipped with an antifade polyvinylpyrrolidone mounting medium (Cat. No. 10981, Sigma-Aldrich). Finally, the images were captured on a Carl Zeiss microscope (Axio Imager.M2p, Germany).
Whole-exome sequencing (WES) and data analysis
Genomic DNA from tumor tissues, tumor-derived organoids, and the paired tumor-adjacent normal tissues were isolated with an AllPrep DNA Mini Kit (Qiagen). The corresponding patients’ tumor-adjacent normal tissues were sequenced and used as a reference. Sequencing libraries were prepared using the Agilent Sure-Select Human All Exon kit (Agilent Technologies, CA, USA) according to the manufacturer’s protocol. The libraries were clustered using TruSeq PE Cluster Kit v4-cBot-HS (Illumina, San Diego, USA) on a cBot Cluster Generation System and sequenced using Illumina NovaSeq with 150 bp paired-end reads. The mean sequencing depth for paired tumor-adjacent normal tissues were approximately 100 ×(~ 10 Gb per sample), and for tumors and organoids were approximately 200 ×(~ 20 Gb per sample). Low-quality reads and adaptors were filtered using Fastp (v0.12.6) [
44]. Single-nucleotide variants (SNVs) were detected using the Genome Analysis Toolkit (GATK, v4.1.9) [
45]. Reads were aligned to the human reference genome GRCh37 with the Burrows-Wheeler Aligner (BWA, v0.7.17) [
46]. SNVs and indels were analyzed by providing the reference (tumor-adjacent normal tissues) and tumors or organoids sequencing data to MuTect 2 and Strelka 2, respectively [
47]. Effect predictions and annotations were added using ANNOVAR (version Feb 2016) [
48]. To detect high quality somatic copy number variations (CNVs), BAM files were performed for read-depth variations using Control-FREEC v11.4 [
49]. Mutational signatures were clustered with deconstructSigs [
50], based on the set of 30 known mutation features.
Drug treatment and organoid viability assay
PTC organoids were dissociated into single cells and small clusters, strained over a 70-µm filter to eliminate large organoids. Then organoids were resuspended in 2% Matrigel/growth medium (10,000 organoids/mL) and plated in Ultra Low Attachment Round Bottom 96-well plates (Cat. No. 7007, Corning) in triplicate. On the following day, drugs were diluted in organoid medium and added into each well with a six-point fivefold dilution series from 3.2 × 10
–3 to 10 µM. For drug combination testing, PTC organoids were seeded into 96-well plates by following the same protocol as described above, and cultured with various doses of targeted agents, individually or in combination. The detailed information and maximum concentration of each drug is listed in Table
2.
Table 2
List of drugs used in this study
BRAFV600E | Vemurafenib | TargetMol | T2382 | 10 |
BRAFV600E | Dabrafenib | MedChemExpress | HY-14660 | 10 |
MEK1 | Selumetinib | MedChemExpress | HY-50706 | 10 |
MEK1/2 | Trametinib | MedChemExpress | HY-10999 | 10 |
VEGFR, Raf-1 | Sorafenib | MedChemExpress | HY-10201 | 10 |
VEGFR, PDGFRβ | Lenvatinib | MedChemExpress | HY-10981 | 10 |
VEGFR, c-Met | Cabozantinib | MedChemExpress | HY-13016 | 10 |
VEGFR, EGFR | Vandetanib | MedChemExpress | HY-10260 | 10 |
VEGFR, PDGFRβ | Sunitinib | MedChemExpress | HY-10255A | 10 |
DNA topoisomerase II | Doxorubicin | MedChemExpress | HY-15142A | 10 |
Microtubule | Vincristine | MedChemExpress | HY-N0488A | 10 |
Microtubule | Paclitaxel | TargetMol | T0968 | 10 |
DNA synthesis | Cisplatin | TargetMol | T1564 | 10 |
Organoid viability was analyzed using the CellTiter-Glo® 3D Reagent (Cat. No. G9683, Promega) according to manufacturer’s specifications following 5 days of drug incubation, and results were normalized to dimethyl sulfoxide (DMSO)-treated control organoids. For testing the combined effect of two drugs, organoids were treated with each drug alone or in combination before undergoing a viability assay. Luminescence reading was performed in a Synergy H1 Hybrid Multi-Mode Microplate Reader (BioTek, USA). All experiments were performed in duplicate in three biological replicates (different passages of PTC organoids). Luminescence readings from drug-treated wells were normalized against that of DMSO-treated control wells, and drug sensitivity was shown by the half-maximal inhibitory concentration (IC50), the slope of the dose–response curve, and the area under the dose–response curve (AUC). For PTC organoids that were completely resistant to a drug, the values of the IC50 and AUC could not be generated, but for analysis purposes, they were given the maximum IC50 of the drug or AUC = 1 within the organoid panel.
Statistical analysis
Results were expressed as mean ± SEM and differences among groups were performed by one-way ANOVA followed by Tukey multiple comparison test using the SPSS 19.0 software (SPSS, Inc., IL, USA). A p-value < 0.05 was considered significant. All statistical analyses were performed using the GraphPad Prism software version 7.0 (GraphPad Software, Inc., CA, USA) or SPSS 19.0. Drug interactions were statistical analyzed with ComboSyn Software (ComboSyn Inc. Paramus, NJ) according to the Chou-Talalay method [
51,
52].
Discussion
Precision medicine is a promising strategy for providing the right therapy for the right patient, which is much required for cancer patients. The lack of a reliable method to predict treatment response is one of the major limitations in clinical oncology. In the past few years, organoid methodology has gained a lot of attention for modeling healthy and malignant tissues. Organoids hold great promise for precision medicine, mirroring its value in basic, translational, and clinical cancer research [
55]. Tumor organoids have been proven to more faithfully recapitulate the histological architectures, molecular characteristics, genomic profiles, mutational signatures, and expression features of primary tumors. In our study, we established BRAF
V600E-mutant and wild-type PTC organoid models in vitro. Under the optimized culture condition, PTC organoids were cultured and passaged for more than 3 months without showing any decrease in growth rate and significant change in morphology. The histopathological characteristics, molecular hallmarks, genetic profiles, and mutational signatures of PTC organoids were concordant with those detected in the matched patient tumors. More importantly, the BRAF
V600E mutation status present in parent tumors were well conserved in the derived organoids.
During the past decades, a great number of anticancer drugs developed from screening 2D cell lines have failed in in vivo studies and clinical trials. PDOs open a door to fill the gap between drug testing on cell lines and clinical trials. Researchers have presented a range of examples for personalized medicine applying organoids in various cancer types [
7,
8,
56,
57]. In the current study, anticancer drug screening using PTC organoid cultures indicated remarkable differential responses of different patients to the treatment. The fact that a variable response to targeted agents and chemotherapeutic drugs was observed in vitro indicates that this method records intra- and interindividual differences. Therefore, by predicting patients’ responses to drugs using their tumor organoids as ‘‘proxies’’, the best treatment strategy for the individual patients may be selected.
BRAF
V600E mutation occurs in most PTCs and some ATCs deriving from PTCs, and it is recognized as the most frequently genetic alteration occurring in thyroid cancers [
22,
58,
59]. This make BRAF
V600E an underlying prognostic biomarker and therapeutic target for thyroid cancer. MEK inhibitor, targeting the RAS/RAF/MEK/ERK signaling pathway, has also been shown separately to enhance survival in patients with metastatic melanoma [
60]. Treatment of BRAF-mutant PTC cell lines with BRAF inhibitor decreased phosphorylation of MEK and subsequently ERK1/2, blocked cell cycle progression, and inhibited tumor xenograft growth [
61]. In our study, we identified BRAF and MEK inhibitions were efficacious only in BRAF
V600E-mutant organoids, but not in BRAF wild-type organoids, suggesting that PTC organoids may be helpful for predicting patients’ response to targeted therapy. With the introduction of BRAF and MEK molecular targeted therapies for patients with advanced or metastatic PTC, determination of BRAF mutation status is crucial to selecting patients who will most likely benefit from this therapy.
Although BRAF
V600E is considered a promising therapeutic target for several cancers, especially for BRAF mutation-harboring melanoma [
28,
62], the BRAF inhibitor monotherapy was less effective than anticipated in clinical trials of patients with PTC harboring the BRAF
V600E mutation [
32,
33]. In our study, BRAF
V600E-mutant PTC organoids were only moderately sensitive to the BRAF
V600E inhibitor (vemurafenib or dabrafenib) treatment. To prevent drug resistance and/or improve response rate, combination therapy targeting BRAF and MEK was assessed and demonstrated synergistic benefit. Many studies have verified that the addition of a MEK inhibitor to a BRAF inhibitor improved progression-free survival and overall survival over BRAF inhibitor monotherapy in patients with BRAF
V600E-mutant melanoma [
34‐
40]. In an open-label Phase II trial, dabrafenib plus trametinib therapy for 16 patients with BRAF
V600E-mutant ATC also generated a significantly higher response rate [
42]. Dual treatment with dabrafenib and trametinib effectively killed BCPAP cells (BRAF
V600E-harboring thyroid cells) after 2 days, indicating that these cells were sensitive to the drugs [
63]. Given the success of the combined treatment in melanoma, continuing efforts are recommended in PTC, which can now be explored with the organoid models in our present study. The combinations of BRAF inhibitor with MEK inhibitor resulted in enhanced treatment response compared with BRAF inhibitor monotherapy. The promising results from PDOs may provide the basis for clinical treatment of locally advanced or metastatic BRAF
V600E-mutant PTC patients.
Our study identified a dozen BRAF inhibitor-based drug combination strategies with good synergistic effects against PTC organoids. The TKIs are a class of small molecules or peptides that have been developed and clinically tested that inhibit either cytosolic or receptor tyrosine kinases [
64]. These enzymes can phosphorylate many regulatory proteins in the cell, and can trigger signal transduction cascades that regulate many cellular functions such as proliferation, differentiation, and metabolism [
65]. We tested the PTC organoids’ sensitivities to several commonly used TKIs, and these assays revealed differential drug responses of individual PTC organoid lines. Interestingly, vandetanib and sunitinib showed more inhibitory effect on PTC organoid lines than sorafenib, lenvatinib and cabozantinib, indicating the two TKIs may be especially beneficial for PTC patients.
BRAF inhibition introduced by V600E mutation causes a rapid feedback activation of human epidermal growth factor receptor (HER)/epidermal growth factor receptor (EGFR) in colorectal cancer cell lines [
66‐
68]. Actually, inhibition of the BRAFV600E/MEK/ERK axis in thyroid cancer cells also results in reactivation of a variety of RTKs such as HER2/HER3 [
69,
70], platelet-derived growth factor receptor-beta (PDGFRβ) [
69], and EGFR [
71]. Treating thyroid cancer cells with BRAF inhibitors will set free C-terminal binding proteins (CTBPs), which was revealed as important transcription factors to promoting expression of HER3 [
69]. The reactivation of HER family members (HER2/HER3) induced the relieving of MAPK/ERK pathways inhibition, leading to the resistance towards BRAF inhibitors in thyroid cancer cells [
69,
70]. Further study demonstrated that combination of HER inhibitor to BRAF/MEK inhibitor overcomes resistance to vemurafenib in BRAF mutated thyroid cancer cells [
69]. In addition to the activation of HER family members, there was upregulation of PDGFRβ in response to vemurafenib in thyroid cancer cell lines [
69]. This is of interest because activation of PDGFRβ has been proposed as a mechanism of acquired resistance to vemurafenib in patients with metastatic melanoma [
72]. Furthermore, Notarangelo and colleagues reported that the exposure of thyroid cancer cells to vemurafenib resulted in a rapid feedback activation of EGFR pathway, and dual EGFR and BRAF blockade induces suppression of ERK signaling, inhibition of cell proliferation, and synthetic lethality [
71]. In our study, BRAF
V600E-mutant PTC organoids derived from some patients were sensitive to vandetanib (inhibits EGFR, vascular endothelial growth factor receptor (VEGFR), and rearranged during transfection (RET)) and/or sunitinib (inhibits PDGFRβ, VEGFR, and fibroblast growth factor receptor (FGFR)). The combination of vandetanib/sunitinib with BRAF inhibitor exhibited more inhibitory effect on PTC organoids than BRAF inhibitor alone. Feedback mechanism that upregulation of EGFR/PDGFRβ which in turn results in reactivation of MAPK pathway may be a probable explanation for drug resistance in BRAF inhibitor monotherapy. Our study provides basis for in vivo and clinical studies using combination of BRAF inhibitor and vandetanib/sunitinib may become a potential therapeutic regimen for BRAF
V600E-positive patients.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.