Establishing a cryopreserved biobank of living tumor tissues for drug sensitivity testing

Abstract

The cryopreservation of cancer tissues to generate frozen libraries is a common practice used worldwide for storing patient samples for later applications. However, frozen samples stored by existing methods cannot be used for initiating living cell cultures, such as patient-derived tumor organoids (PDOs), which offer great potential for personalized treatment. To overcome this challenge, we developed a novel procedure for culturing PDOs using frozen live tumor tissues. We show that tumor specimens stored using this technique maintain their viability and can be successfully used to generate organoids even after long-term freezing, with an impressive success rate of 95.2 %. Importantly, we found that the structural features, tumor marker expression, and drug responses of organoids derived from frozen tissues are similar to those derived from fresh tissues. Moreover, organoids derived from frozen tissues can be routinely passaged and frozen, making them ideal for high-throughput drug screening at any time. Notably, cryopreserved tumor tissues can also be utilized in air-liquid interface (ALI) culture. This method allows for preserving the original tumor microenvironment, making it an invaluable resource for conducting tests on antitumor drug responses, including immune checkpoint inhibitors (ICIs). This innovation has the potential to enable the identification of potentially effective drugs for patients and facilitate the development of novel therapeutic drugs. Thus, we have established protocols for the long-term cryopreservation of cancer tissues to maintain their viability and microenvironment, which are useful for personalized therapy.

Graphical abstract

Highlights

• •We successfully developed a technique for cryopreservation of live tumor tissues for drug sensitivity testing.
• •Organoids derived from frozen tissues retained the main features of the original tumors.
• •Organoids derived from frozen and fresh tissues exhibit similar drug responses.
• •Frozen tumor tissues can be used for ALI cultivation that maintains the tumor microenvironment.

1. Introduction

Biobanking encompasses collecting and storing substantial quantities of biological samples, accompanied by pertinent personal and health data such as health records, family history, lifestyle information, and genetic data [1,2]. These valuable resources primarily facilitate health and medical research endeavors. In recent years, there has been a notable rise in the establishment of large specimen biobanks, including tumor tissues and other samples [3]. These banks typically store specimens directly under frozen conditions, i.e., a deep freezer or liquid nitrogen. Unfortunately, most of these storage methods kill cells, restricting their use only to extracting proteins, DNA, or RNA, rather than for culturing live cells, significantly limiting their overall value and potential applications [4].

Patient-derived organoids (PDOs), three-dimensional tumor models derived from tissue subunits or stem cells under specific culture conditions, are in vitro tumor models developed in recent years [5,6]. The PDO model has fast modeling speed and a high success rate and allows for the recapitulation of key attributes such as tissue structure, heterogeneity and physiological characteristics of the original tumor, providing an ideal model for mimicking the parental tumor in vitro. A broad spectrum of cancer PDO models have been successfully established, allowing a better understanding of cancer development patterns and providing an ideal model for exploring how drugs interact with tumors [[7], [8], [9], [10], [11], [12], [13], [14], [15], [16]]. The results of existing studies have shown that PDO drug susceptibility testing can not only reflect the previous clinical treatment response of patients but also predict effective individualized treatment options for patients, providing a new method for drug research and development [10,[17], [18], [19], [20], [21], [22]].

Currently, most PDO models use fresh tumor specimens as starting materials to ensure the success rate of modeling to some extent [23]. However, a proportion of patients will experience tumor recurrence or metastasis after treatment; these patients are often unable to undergo surgery again, and specimen acquisition is often difficult. In addition, recurrent and/or metastatic tumors typically develop rapidly and exhibit resistance to most treatments, requiring the identification of the optimal treatment regimen within the shortest possible time [24,25]. For these tumor patients, if the living tumor specimens are stored at the time of initial surgery, the frozen specimens could be revived for establishing PDO models and drug susceptibility testing, which would facilitate the timely provision of individualized treatment options for patients.

In our previous study, to ensure long-distance transportation of tumor specimens, we established a cryopreservation method for live tissues. We achieved a 64 % success rate in cultivating PDOs from 55 frozen human breast cancer samples [20]. This method can also be used to preserve patient tumor specimens during initial surgeries to maintain the viability of cells within the tissue for unforeseen future needs. In this study, we improved the cryopreservation method. We established PDOs from 42 frozen tumor samples (PDOs can be successfully established from all these fresh tumor tissues) with a success rate greater than 95 %. We also conducted comparative studies of the growth characteristics, histological structure, expression of tumor markers and drug sensitivity of PDOs generated from fresh and cryopreserved tissues to illustrate the usefulness of PDOs derived from frozen libraries. Furthermore, we demonstrated that cryopreserved tissues suit air-liquid interface (ALI) cultures. This modified organoid culture method aims to preserve the microenvironment of the original tumor and allows for testing of antitumor drug responses, including ICIs.

2. Materials and methods

2.1. Human and mouse specimens

Human breast cancer tissues were collected in our previous study [20]. Human colorectal, lung and kidney cancer specimens were obtained from Kiang Wu Hospital. The ethics committees at the University of Macau and all hospitals involved in this study meticulously assessed and granted their approval (Accreditation number: BSERE16-APP010-FHS). We followed ethical guidelines and obtained informed consent from all donors before collecting any samples. Patient clinical data were retrieved from the medical records system. A total of six mice were used in this study and their strains were as follows: H5690T (Br^co/co, Cre-MMTV), H6164T (Br^co/co, Cre-MMTV), HP6403T (Br^co/+, Cre-MMTV), H6433T (Br^co/co, Cre-MMTV), HP6654T (Neu, Cre-MMTV) and H10166T (Fgfr2^{pLoxpneo-S525w/+}, P53^co/+, Cre-MMTV) [26]. These mice can spontaneously develop breast tumors at approximately 12–24 months of age. All animal experiments were approved by the University of Macau Animal Ethics Committee (Accreditation number: UMARE-015-2019).

2.2. Procedures for cryopreserving live tumor tissues

In our previous studies, the tissue fragments stored were relatively large (with a diameter of approximately 2–5 mm). The cells inside the tissue fragments could not be in sufficient contact with the cryopreservation solution, leading to vulnerability to damage during the freezing and recovery processes. Additionally, the tissues stored in the −80 °C freezer for an extended period were not promptly transferred to liquid nitrogen for storage. Furthermore, some specimens were not processed immediately and frozen promptly after surgery. These factors resulted in a lower success rate of organoid culture from cryopreserved tissues. In this study, we made improvements to these techniques using the following procedures: 1) Tumor tissues were placed on a 500 μm filter in a dish containing culture medium. The tissue blocks were then carefully cut with scissors until they could smoothly pass through the filter to obtain fragments less than 500 μm in diameter. 2) A portion of these fresh tissue fragments were digested for organoid culture. The remaining tissues were suspended at a density of approximately 200–300 pieces per tube in a cryopreservation solution composed of 10 % DMSO (Sigma‒Aldrich) and 90 % FBS (GIBCO). 3) These tissue pieces were then dispensed into cryopreservation tubes and transferred to cryostorage boxes. Initially, they were stored in a −20 °C freezer for 2 h, followed by a −80 °C freezer for 22 h before ultimately being transferred to liquid nitrogen for long-term storage. Careful optimization of each process step is essential to ensure high cell viability and functionality after cryopreservation.

To obtain core needle biopsy tumor tissues from mice, the animals were anesthetized in advance, and then a puncture biopsy was performed using an 18-gauge needle. The approximately 2 cm long biopsy tissues were placed into a cryovial containing 1 mL of cryopreservation solution. Following the gradient cooling method, the samples were transferred to liquid nitrogen for long-term preservation.

2.3. Tissue dissociation

Fresh tissue pieces can be used directly without any prior processing for tissue digestion. However, frozen tissue requires some preparatory steps before digestion can take place. This involves thawing the tissue at 37 °C and centrifuging the tissue at 400×g for 4 min. After this, the resulting tissue pellet was ready for digestion. To ensure optimal tissue dissociation, it is recommended to add an appropriate volume of digestion buffer, typically ranging from 6 to 12 mL depending on the size of the tissues. The composition of the digestion buffer followed the protocol outlined in the published article [20]. Its components included DMEM/F12 medium (GIBCO), supplemented with collagenase type III (Worthington) at a concentration of 300 U/mL, hyaluronidase (Sigma‒Aldrich) at 100 U/mL, insulin (Sigma) at 5 μg/mL, hydrocortisone (Sigma‒Aldrich) at 500 ng/mL, cholera toxin (Sigma‒Aldrich) at 20 ng/mL, EGF (PeproTech) at 10 ng/mL, and FBS (GIBCO) at 5 %. The tissue digestion process was carried out on a shaker at 37 °C for 1–2 h, with intermittent pipetting until all visible pieces disappeared. The resulting cell clusters were spun down at 400×g for 4 min. They were washed once with DMEM supplemented with 5 % FBS and spun down at 400×g for 4 min to obtain a concentrated cell pellet.

2.4. Organoid culture, passage and cryopreservation

The cell clusters were resuspended in 50 % cold Matrigel (Corning) and then seeded into a prewarmed 6-well plate (Corning) using 25 μL drops. The drops were solidified by incubating them for 30 min in a 37 °C and 5 % CO2 incubator. After solidification, each well was supplemented with 2.5 mL of organoid culture medium and refreshed every 3–6 days based on the observed cell growth rate. The L-WRN cell line was used to prepare the conditioned medium. The composition of the organoid culture medium for breast cancer was based on previous formulas with slight modifications using 30 % L-WRN-conditioned medium [20]. The culture media used for the other three types of cancer were Advanced DMEM/F12 supplemented with 30 % L-WRN conditioned medium, 5 μM Y-27632, 1.0 μM SB202190, 0.5 μM A83-01, 50 ng/ml EGF, 500 ng/ml hydrocortisone, 10 mM nicotinamide, 1.25 mM N-acetyl-L-cysteine, 15 mM HEPES, 1 × B27, 1 × GlutaMAX, 1 × insulin-transferrin-selenium-sodium pyruvate, 0.5 μg/ml amphotericin B, 5 μg/ml gentamicin, and 5 μg/ml plasmocin (Table S1).

The organoids were passaged every 1–4 weeks, with the frequency determined by their size and density as reported in previous studies [20]. The culture medium was removed, and the organoids were gently resuspended in cold 0.25 % trypsin using a 1 mL pipette. The resulting mixture was incubated at 37 °C for 4 min; additional pipetting and incubation were performed if necessary. The cells were spun down at 400×g for 4 min, washed and resuspended in 6 mL of DMEM supplemented with 5 % FBS before being spun down again. Finally, the cells were resuspended in Matrigel at 1:3 to 1:6 based on their density and growth rate, seeded into prewarmed 6-well plates and cultured as previously described.

Before cryopreservation, the organoids were first digested into single cells with 0.25 % trypsin according to the above steps. The cells were then washed with DMEM supplemented with 5 % FBS and resuspended in a frozen medium of 10 % DMSO (Sigma‒Aldrich) and 90 % FBS (GIBCO) at 1 × 10⁶ cells/tube density. The cell suspension was then transferred to cryovials and slowly cooled in a −80 °C freezer for 24 h using a cryostorage box before storage in liquid nitrogen. If necessary, the cryopreserved cells can be swiftly thawed at 37 °C, followed by centrifugation. Afterward, the pellet was resuspended in 50 % cold Matrigel and cultured as described previously.

2.5. Whole exome sequencing

The DNeasy Blood & Tissue Kit (QIAGEN) was used to prepare genomic DNA samples in accordance with the manufacturer's instructions. After hydrodynamic shearing of the DNA using a Covaris system, 1 μg of genomic DNA per sample was utilized for library preparation, generating fragments of 180–280 bp. An Agilent SureSelect Human All Exon Kit (Agilent Technologies, CA, USA) was used according to the manufacturer's recommended protocols to prepare whole-exome-captured libraries. Subsequently, the resulting products were subjected to quality and quantity assessments using an Agilent High Sensitivity DNA Kit on an Agilent 2100 Bioanalyzer. A TruSeq PE Cluster Kit v4-cBot-HS (Illumina, San Diego, USA) was used to cluster the libraries on a cBot Cluster Generation System, and the libraries were subsequently sequenced on an Illumina NovaSeq 6000 platform, generating 150 bp paired-end reads.

2.6. Whole-exome sequencing data analysis

Quality control of the DNA-sequencing data was carried out using MultiQC (v1.11) [27]. Adapter trimming was performed using Trim Galore (v0.6.7). Sequence reads were aligned to human genome build 38 (Hg38) using the Burrows-Wheeler Alignment tool with maximal exact matches (v0.7.17) [28], followed by the reduction of duplicates with Sambamba (v0.6.6) [29], the realignment of indels and base recalibration with the Genome Analysis ToolKit (v4.2.2.0) according to the best practice guidelines (Table S2) [30].

Three independent programs identified somatic SNVs with default parameters, namely, MuTect2 (v4.2.2.0)^30, [31], Strelka2 (v2.9.2) [32], and LANCET (v1.1.0) [33]. Paired blood control or adjacent control samples were used as the reference to call somatic mutations of each tumor or PDO independently. The results of each caller were stored in VCF format and further filtered for PASS variants. The vcf2maf tool (https://github.com/mskcc/vcf2maf) was used to convert VCFs into MAFs via annotation by the ENSEMBL Variant Effect Predictor according to reference release 104 [34]. Only mutations detected by at least two callers were kept as true positives and assigned for further analysis. To detect somatic CNAs (copy number alterations), BAM files were analyzed using CNVkit (v0.9.9) by comparing tumors or PDOs to reference blood control or adjacent control samples (Table S2) [35].

2.7. Drug response assay

The drug response assay was conducted in accordance with previous studies [20]. Briefly, organoids were initially dissociated into smaller clusters and then suspended in a culture medium supplemented with 2.5 % Matrigel. The resulting cell suspensions were placed in 384-well plates coated with collagen I gel at 36 μL per well, corresponding to approximately 1500 cells. After allowing for 48 h for seeding, 4 μL of each compound, which had been appropriately diluted, was added to the respective wells. A total of 14 drugs were tested (Table S3), and these compounds were used at their highest concentration of 20 μM, with dilutions made at 3-fold intervals for seven dilutions. Three technical replicates of each drug were examined across three plates. When testing drug panels containing 430 kinase inhibitors (Table S3), the highest drug concentration used was 10 μM, and dilutions were performed at 3-fold intervals for eight dilutions. No duplicate samples were analyzed with these drug panels. A positive control was achieved using bortezomib (1 μM), whereas a negative control was established utilizing 0.2 % DMSO. Cell viability was assessed after four days of drug incubation using the CellTiter-Glo 2.0 assay (Promega) following the manufacturer's instructions. The results were normalized to those of the controls and are expressed as a percentage of cell viability. The IC₅₀ values were determined by employing GraphPad Prism 7. In cases where certain organoid lines resisted specific drugs, their IC₅₀ values surpassed the highest detectable concentration and were assigned the maximum tested drug concentration. Conversely, some organoid lines exhibited high sensitivity to specific drugs, with IC₅₀ values exceeding the lowest detectable concentration, and thus were assigned the minimum tested drug concentration.

2.8. Live-cell imaging of organoids after drug treatment

Clustered cells were resuspended in 50 % cold Matrigel (Corning) and then seeded onto the chambered glass of a preheated 35 mm glass bottom dish with 4 chambers (Cellvis, D35C4-20-0-N) using 25 μL drops. The drops were solidified by incubation at 37 °C and 5 % CO₂ in a cell culture incubator for 30 min. After solidification, 0.6 mL of organoid culture medium was added to each chamber. Depending on the cell growth rate, the cells were cultured for 2–3 days, followed by adding drugs and continued culture for four days. NucRed Dead 647 (one drop per mL; Thermo Fisher) and TO-PRO-3 (1:1500; Thermo Fisher) were added for fluorescent labeling of living and dead cells (referred to as 'Imaging medium'), respectively. The cells were further incubated for 2 h to allow sufficient staining with the fluorescent dyes. The plate was placed in a Nikon A1R confocal system with an incubation chamber (37 °C, 5 % CO2). The emission spectra of fluorescent real-time dead cell dye (NucRed Dead 647, 638 nm) and living cell dye (TO-PRO-3, 562 nm) were individually imaged using the λ mode of the laser in multispectral imaging. Using the Plan Apochromat × 20/0.8 numerical aperture objective lens for flat imaging, the following settings were used: online fingerprint mode, bidirectional scanning, optimal Z-stack step size, total Z-stack thickness of 200 μm, resolution of 1024 × 1024, and capturing of an image every 2.5 μm [36].

2.9. Air-liquid interface (ALI) culture and drug testing

ALI cultures were conducted following previously described methods with minor adjustments [37,38]. To prepare the ice-cold collagen solution, solution A (rat collagen I, R&D Systems, 3447-020-01), solution B (10 × Ham's F-12), and solution C (sterile reconstitution buffer containing 2.2 g NaHCO₃ in 100 mL 0.05 M NaOH and 200 mM HEPES) were mixed at a volume ratio of 8:1:1. Subsequently, 150 μL of collagen solution was added to the inserts (SPLInsert Standing, 37524) and incubated in a 37 °C incubator for 20–30 min until the collagen had completely coagulated. The tumor tissues (<500 μm) were then resuspended with the ice-cold collagen solution, and 50 μL of the tissue suspension (∼200 pieces) was added to the respective inserts. After the top tissue-containing gel solidified, 400 μL of ALI basal medium or medium supplemented with drugs was added to the outer chambers of the culture inserts. The chambers were subsequently incubated in a humidified incubator at 5 % CO₂ and 37 °C for 4 days. At the end of the culture, 200 μL of 50 μg/mL MTT solution (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT)) was added to each well and incubated at 37 °C for 2 h. The upper layer containing the tissues was then removed and placed in a 96-well plate, and 200 μL of DMSO was added to dissolve the formazan in each well. The absorbance value of each well was read at a wavelength of 490 nm. The ALI basal medium was F-medium, and 500 mL of medium consisted of 373 mL of Dulbecco's modified Eagle's medium (DMEM), 125 mL of Ham's F-12 Nutrient Mix, 5 μg/mL insulin, 250 ng/mL amphotericin B, 10 μg/mL gentamicin, 0.1 nM cholera toxin, 0.125 ng/mL EGF, 25 ng/mL hydrocortisone, and 10 μM Y-27632. To assess the antitumor activity of mouse PD-L1 antibody (BioLegend, 124301) or human PD-1 antibody (BeiGene, SFDA S20190045) (Table S3), 500 IU/mL of recombinant mouse IL-2 (PeproTech, 212-12) or 500 IU/mL of recombinant human IL-2 (PeproTech, 200-02) was added to the culture medium to maintain immune cell activity, respectively.

2.10. Histology and immunohistochemistry

The tumor tissue and ALI culture samples were fixed in 4 % paraformaldehyde and subsequently embedded in paraffin using standard protocols. The organoids were resuspended in cold PBS through pipetting, spun down, and fixed in 4 % paraformaldehyde for 24 h. Subsequently, the organoids underwent dehydration and paraffin embedding using methods described in previous studies [20]. For all analyses, paraffin sections 5 μm in thickness were used. Histopathological evaluation was conducted using hematoxylin-eosin (H&E) staining, while immunohistochemical analysis was employed to detect the expression of tumor markers. The primary antibodies utilized in this study were anti-ER (Abcam, ab16660, 1:200), anti-PR (Abcam, ab101688, 1:400), anti-HER2 (CST, 2165, 1:200), anti-CK20 (CST, 13063, 1:200), anti-CDX2 (Abcam, ab76541, 1:800), anti-MUC2 (CST, 88686, 1:200), anti-Ki-67 (CST, 9449S, 1:500), anti-CD4 (Abcam, ab183685, 1:300), anti-CD8 (CST, 98941, 1:300), anti-CD45 (CST, 70257, 1:400), anti-F4/80 (CST, 70076, 1:300), anti-SMA (Abcam, ab124964, 1:300) and anti-PD-L1 (CST, 13684, 1:200) (Table S4). The images were taken using a Nikon microscope (Nikon, ECLIPSE, Ci).

2.11. Quantification and statistical analysis

All analyses were conducted in GraphPad Prism 7 (GraphPad Software, La Jolla, CA), and the values are presented as the means ± SEMs of individual samples. The P values reported in the Figures were obtained using unpaired two-tailed t-tests. P values < 0.05 were considered to indicate statistical significance. The sequence analysis details are described in the relevant methods section.

3. Results

3.1. Establishment of organoids from a cryopreserved biobank of living tumor tissues

To investigate the impact of freezing on cellular viability within tissues, we compared the viability of three fresh and frozen colorectal cancer specimens. The results showed that the cellular viability of the fresh tissues from the three specimens (CPT17, CPT18, and CPT19) was 85.65 %, 87.9 %, and 89.55 %, respectively, whereas the cellular viability within the frozen tissues was 83 %, 84.73 %, and 85.58 %, respectively. This indicates that while freezing reduces cellular viability within tissues, the reduction is within 4 %. To investigate the success rate of culturing PDOs from frozen tumor tissues, we used 42 surgically resected human cancer tissue samples, including two kidney cancer tissue samples, 11 breast cancer tissue samples, 13 lung cancer tissue samples, and 16 colon cancer tissue samples. PDOs can be successfully established from all these fresh tumor tissues. These specimens were frozen in liquid nitrogen for 4–46 months. Using a protocol established in the laboratory (Fig. 1A), our results showed that PDOs could be successfully established in 40 cryopreserved tissues, except for one lung cancer tissue and one colorectal cancer tissue, with a success rate of 95.2 % (40/42) (Fig. S1, A-C; data file S1). Importantly, all nine specimens that had undergone freezing for more than 40 months were successfully cultured to obtain PDOs (Fig. S1C). This achievement is compelling evidence that our frozen technique is highly proficient in sustaining viable cells within tumor tissue over an extended period.

Fig. 1.

Establishing organoids using cryopreserved human tumor tissues. (A) Schematic workflow for collecting, processing, cryopreserving surgically resected tumor tissues, and culturing organoids. (B) Bright-field microscope images displaying the morphology of five pairs of organoids derived from fresh (top row) and frozen (bottom row) human tumor tissues. Scale bar, 100 μm. (C) Bright-field microscopy images showing the morphology of a pair of fresh (CPDO8, top row) and frozen (F-CPDO8, bottom row) colorectal cancer tissue-derived organoids at different time points during culture. Scale bar, 100 μm. (D) Histological and immunohistochemical images showing the structural organization and the expression of the proliferation marker (Ki-67) and colorectal cancer-related markers (CK20, CDX2, and MUC2) in primary colorectal cancer (CPT16) and its fresh tissue-derived (CPDO16) and frozen tissue-derived (F-CPDO16) organoids. Scale bar, 20 μm.

We also investigated the establishment of cryopreservation methods for small specimens, such as biopsy tissues from mouse tumors, and studied the effect of cryopreservation time on their activity (Fig. S2A). We examined biopsy tumor tissues from four different mice. We found that after one week and six months of cryopreservation, tumor tissues could be used for organoid culture and maintained a growth rate similar to organoids derived from fresh tissues (Fig. S2, B-C). These results indicate that small tissues preserved by this method, which do not negatively impact cell viability, can also be used for organoid culture.

Next, we conducted a morphological analysis of these organoids. We found that most of the breast cancer, lung cancer and kidney cancer organoids exhibited cystic or solid phenotypes, and one breast cancer organoid exhibited a grape-like morphology (Fig. 1B and Fig. S3A). While most colorectal cancer organoids had hollow torus structures, a small number had solid morphology (Fig. 1B and Fig. S3A). The morphology of organoids derived from cryopreserved tissues did not differ from those derived from fresh tissues. In addition, we compared the growth patterns of fresh and cryopreserved tissue-derived organoids and found that they had very similar growth rates (Fig. 1C).

Overall, we established protocols for collecting and long-term preservation of tumor tissues with key features for maintaining their viability for later application, which is usually not considered by currently frozen cancer libraries. Our results showed that cryopreserved tissues, when properly frozen even after a long period, had a high success rate of organoid culture, and the established organoids were highly consistent with fresh tissue-derived organoids in terms of morphological structure and growth rate.

3.2. Both cryopreserved and fresh tissue-derived PDOs recapitulate the histopathological and genetic characteristics of parental tumors

Although previous studies have demonstrated that organoids derived from fresh tissue well recapitulate the histological features of parental tumors, whether organoids derived from cryopreserved tissue also have this property remains to be determined. We compared the histopathological features of breast and colorectal organoids with those of parental tumors. The results of hematoxylin and eosin (H&E) staining showed that the organoids derived from cryopreserved tissues could retain the histological characteristics of their parental tumor tissues as those derived from fresh tissues (Fig. 1D and Fig. S3, B-C). We subsequently assessed the expression of the cell proliferation marker Ki-67, breast cancer-related markers [estrogen receptor (ER), progesterone receptor (PR), and human growth factor receptor 2 (HER2)] and colorectal cancer-related markers (CK20, CDX2 and MUC-2) in the corresponding organoids and tumor tissues. The immunohistochemical (IHC) staining showed that the expression patterns of the markers were highly consistent in the organoids derived from the corresponding fresh/cryopreserved tissues and tumor tissues. These results indicate that cryopreserved tissue-derived organoids, such as fresh tissue-derived organoids, can preserve the histopathological features of the original tumor (Fig. 1D and Fig. S3, B-C).

Whole exome sequencing (WES) analysis was performed on 12 pairs of cryopreserved/fresh tissue-derived organoids and parental tumors (from six breast cancer patients and six colorectal cancer patients) to determine whether cryopreserved tissue-derived organoids could maintain the genomic characteristics of their parental tumors similar to those of fresh tissue-derived organoids. Genome-wide exome copy number alteration (CNA) analysis revealed that in these two types of tumors, organoids derived from cryopreserved tissues retained the DNA copy number characteristics of the parental tumor to some extent, similar to organoids derived from fresh tissues (Fig. 2A and Fig. S4A). We observed that the number of CNA genes in cryopreserved and fresh tissue-derived organoids was usually similar but greater than that in the corresponding primary tumors. We next compared the CNA profiles of cancer-related genes in organoids and corresponding tumors. Consistent with the genome-wide results, the CNA patterns of cancer-related genes in the organoids derived from fresh tissue-derived organoids were well recapitulated in the organoids derived from cryopreserved samples but were generally greater than those in the primary tumors (Fig. 2B). We further compared single nucleotide variant (SNV) genes in parental tumors and organoids. In both cryopreserved tissue-derived organoids and fresh tissue-derived organoids, mutations in tumor-related genes, such as TP53, PIK3CA, COL11A1, AKT1 and SMAD3/4, in the original tissues were well preserved (Fig. 2C). CNA scatterplots further confirmed that the CNA genes were more highly enriched in the organoids derived from cryopreserved and fresh tissues than in primary tumor tissues (Fig. 2D and Fig. S4, B-C).

Fig. 2.

Organoids derived from fresh and cryopreserved tumor tissues recapitulate the genetic characteristics of the parental tumors. (A) Comparison of copy number alteration (CNA) landscapes between 6 pairs of organoids derived from fresh (PDOs) and cryopreserved tumor tissues (F-PDOs) and their parental tumors. Copy number amplifications are shaded in red, while copy number deletions are shaded in blue. (B) Heatmap comparing the copy number alterations (CNAs) of cancer genes in fresh (PDOs) and cryopreserved tumor tissue-derived organoids (F-PDOs) with those in the parental tumors. (C) Heatmap comparing the somatic mutated cancer genes in fresh (PDOs) and cryopreserved tumor tissue-derived organoids (F-PDOs) with those in the parental tumors. (D) CNA scatterplot of primary tumor tissues (green), fresh tumor tissue-derived organoids (dark blue), frozen tumor tissue-derived organoids (red), and germline (gray). The copy number is displayed on the y-axis. The chromosomes are arranged from left to right.

In summary, cryopreserved tissue-derived organoids, similar to fresh tissue-derived organoids, preserve the histological characteristics of the original tumors, the expression of tumor markers and the genomic characteristics well. Similar to fresh tissue-derived organoids, organoids derived from cryopreserved tissues typically exhibit more CNA and SNV genes than the parental tumors. This is mainly because cancer organoids are composed entirely of tumor cells, while primary tissues contain other types of cells, such as stromal cells and immune cells, leading to lower tumor cell purity than organoids.

3.3. Comparing the drug responses of organoids derived from cryopreserved and fresh tumor tissues

To assess the feasibility of using cryopreserved tumor tissue-derived organoids to evaluate tumor drug response in patients, we compared the drug sensitivity of PDOs derived from cryopreserved and fresh tumor tissues. We designed a drug panel containing 14 commonly used drugs for clinical treatment (seven chemotherapeutics and seven targeted therapeutics) to detect drug susceptibility in three colorectal cancer organoids and two breast cancer organoids. Drug sensitivity was expressed as the concentration inhibiting 50 % of the cells (IC₅₀). There were significant differences in the response of PDOs to different drugs. Among the seven tested chemotherapeutic agents, PDOs were more sensitive to epirubicin HCl and doxorubicin HCl, more resistant to gemcitabine HCl, bleomycin sulfate, clofarabine, and cytarabine, and had greater differences in sensitivity to topotecan HCl (Fig. 3A). Organoids derived from different individuals were sensitive to neratinib and afatinib, which target HER2 and EGFR, but resistant to gefitinib and erlotinib HCl, which target EGFR (Fig. 3A). Importantly, we detected high consistency in the drug responses of organoids derived from cryopreserved and fresh tumor tissues from the same individual (Fig. 3A–B and Fig. S5A). This indicates that the drug responses of the derived organoids are not affected by the preservation of tumor tissue freezing. Like organoids derived from fresh tumor tissues, they should be able to represent patient-specific drug responses.

Fig. 3.

The drug responses of organoids derived from cryopreserved and fresh tumor tissues are consistent. (A) Heatmap showing the IC₅₀ values of 14 cancer therapy drugs in five pairs of fresh tumor tissue-derived organoids (PDOs) and frozen tumor tissue-derived organoids (F-PDOs). The right side lists the tested drugs and their targets. (B) Comparison of the representative drug response curves of a pair of colorectal cancer organoids derived from fresh tissue (CPDO5) and frozen tissue (F-CPDO5). The results are expressed as the mean ± SEM. (C) 3D dual-spectral images showing the drug responses of a pair of colorectal cancer organoids derived from fresh (CPDO5) and frozen tissues (F-CPDO5) after 96 h of treatment with the sensitive drug (neratinib, 1.0 μM) and the insensitive drug (clofarabine, 1.0 μM). The purple color represents viable organoids, while the yellow color represents dead organoids. (D) Quantification of the cytotoxic effects of the drugs on the organoids in Panel (D) through 3D live-cell imaging. The results are expressed as the mean ± SEM (n = 3). ∗∗ indicates P < 0.01; ns indicates P > 0.05.

To confirm the similarity of drug responses between organoids derived from cryopreserved and fresh tumor tissues, we treated two pairs of organoids (CPDO5/F-CPDO5 and BPDO7/F-BPDO7) with drugs that have different inhibitory effects. After 96 h of drug treatment, we observed that CPDO5/F-CPDO5 exhibited relatively high sensitivity to neratinib (1.0 μM) (Fig. 3C and D; Videos S1, S2, S4 and S5). Following neratinib treatment, only a small number of surviving and smaller-sized organoids were observed, and the proportion of dead organoids was significantly greater than that in the control group (Fig. 3C and D; Videos S1, S2, S4 and S5). However, neither CPDO5 nor F-CPDO5 showed sensitivity to clofarabine (1.0 μM) treatment (Fig. 3C and D; Videos S1, S3, S4 and S6). Following treatment with afatinib (0.1 μM), both BPDO7 and F-BPDO7 showed a significant decrease in the proportion of viable organoids and a significant increase in the proportion of dead organoids (Fig. S5, B-C; Videos S7, S8, S10 and S11). However, BPDO7 and F-BPDO7 exhibited limited responses to gemcitabine HCl (2.0 μM) treatment (Fig. S5, B-C; Videos S7, S9, S10 and S12). These results further confirm that organoids derived from cryopreserved tissues exhibit similar drug responses to those derived from fresh tissues.

3.4. Characteristics of organoids derived from cryopreserved tumor tissues after long-term culture

Next, we investigated the effects of long-term culture on organoids derived from cryopreserved tumor tissues. We passed the colorectal cancer organoid F-CPDO3 and breast cancer organoid F-BPDO5 for 23 and 19 generations, respectively, with a culture duration of more than one year. After long-term culture, we found that organoids derived from cryopreserved tumor tissues still exhibited growth rates similar to those of early-passage cryopreserved and fresh tumor tissue-derived organoids. The morphology of these organoids remained unchanged, with both retaining the solid structure observed in early-passage cryopreserved and fresh tumor tissue-derived organoids (Fig. 4A and Fig. S6A). Further histological examination and IHC staining revealed no significant alterations in tissue structure or the expression of Ki-67, breast cancer-related markers (ER, PR and HER2) or colorectal cancer-related markers (CK20, CDX2 and MUC-2) in the two organoid lines after long-term culture (Fig. 4B and Fig. S6B).

Fig. 4.

After long-term culture, organoids derived from cryopreserved tumor tissues retain their morphological structures, histological features, and genomic characteristics. (A) Bright-field microscopy images showing the morphology of a 5th passage fresh colorectal cancer tissue-derived organoid (CPDO3-P5), as well as the morphology of corresponding frozen tumor tissue-derived organoids of the 5th passage (F-CPDO3-P5), 13th passage (F-CPDO3-P13), and 23rd passage (F-CPDO3-P23). Scale bar, 100 μm. (B) Histological and immunohistochemical images showing the structural organization and the expression of the proliferation marker (Ki-67) and colorectal cancer-related markers (CK20, CDX2, and MUC2) in primary colorectal tumor (CPT3), as well as in 5th passage fresh tumor tissue-derived organoids (CPDO3-P5) and frozen tumor tissue-derived organoids of the 5th passage (F-CPDO3-P5), 13th passage (F-CPDO3-P13), and 23rd passage (F-CPDO3-P23). Scale bar, 20 μm. (C) Comparison of CNA landscapes between two pairs of organoids derived from fresh tumor tissues (PDOs) and frozen tumor tissues of different passages (F-PDOs) and their parental tumors. Copy number amplifications are shaded in red, while copy number deletions are shaded in blue. (D) Heatmap comparing the CNAs of cancer genes in two pairs of organoids derived from fresh tumor tissues (PDOs) and frozen tumor tissues of different passages (F-PDOs) and their parental tumors. (E) Heatmap comparing the somatic mutated cancer genes in two pairs of organoids derived from fresh tumor tissues (PDOs) and frozen tumor tissues of different passages (F-PDOs) and their parental tumors.

We subsequently performed WES analysis on different generations of F-CPDO3 and F-BPDO5. The results of the genome-wide CNA analysis indicated that after continuous passaging, F-CPDO3 retained the characteristic of having fewer CNA genes, while F-BPDO5 retained the characteristic of having more CNA genes in early generations of organoids (Fig. 4C–D andFig. S6, C-D). Further analysis revealed that the mutation status of cancer-related genes was well preserved in both organoid lines after continuous passaging. F-CPDO3 exhibited a greater number of SNV gene mutations, while F-BPDO5 exhibited a lower number of SNV gene mutations (Fig. 4E). Our results indicate that organoids derived from cryopreserved tumor tissues retain the genomic characteristics of early-generation organoids after long-term culture.

To assess whether long-term culture affects the drug response of cryopreserved tumor tissue-derived organoids, we conducted drug susceptibility tests of F-CPDO3 and F-BPDO5 at different passages. Our results indicated that even after continuous passage up to passage 19 (F-BPDO5) and passage 23 (F-CPDO3), frozen-preserved tissue-derived organoids still exhibited similar drug responses to those of early-passage cryopreserved tissue-derived organoids and fresh tissue-derived organoids (Fig. S7, A-B).

In summary, organoids derived from cryopreserved tumor tissues can undergo long-term passaging in culture while maintaining morphological structures, histological features, and genomic characteristics similar to those derived from fresh tissues. Additionally, their response to drugs does not significantly change during passaging.

3.5. Using organoids derived from cryopreserved tumor tissues of drug-resistant patients for high-throughput drug screening

Drug resistance, including intrinsic resistance and acquired resistance, is one of the main obstacles in cancer treatment, greatly limiting the choice and use of cancer drugs. To identify drugs that effectively reverse tumor resistance, we established a drug panel containing 430 kinase inhibitors. We conducted high-throughput drug sensitivity screening on two organoids derived from cryopreserved tissues of drug-resistant tumors (Fig. 5A). F-BPDO5 was derived from tumor tissue that was resistant to neoadjuvant chemotherapy. Our drug screening results showed that F-BPDO5 exhibited resistance to 200 kinase inhibitors (IC₅₀ ≥ 10 μM), moderate response to 200 kinase inhibitors (2 μM < IC₅₀ < 10 μM), and sensitivity to 30 kinase inhibitors (IC₅₀ ≤ 2 μM). Among these 30 most sensitive kinase inhibitors, 13 targeted the PI3K/Akt/mTOR signaling pathway, and 10 targeted the cell cycle signaling pathway (Fig. S8, A-B). The IC₅₀ values of the three most sensitive drugs, GSK2126458 (GSK458), dinaciclib (SCH727965), and BGT226 (NVP-BGT226), were 0.004673 μM, 0.011 μM, and 0.06567 μM, respectively (Fig. S8, A and C). F-BPDO1 was derived from a HER2+ breast cancer patient who developed multidrug resistance and multisite systemic metastases after undergoing seven rounds of treatment with more than 10 different drugs. High-throughput drug screening revealed that F-BPDO1 exhibited resistance to most of the tested kinase inhibitors (IC₅₀ ≥ 10 μM), moderate response to 30 kinase inhibitors (2 μM < IC₅₀ < 10 μM), and sensitivity to only 10 kinase inhibitors (IC₅₀ ≤ 2 μM). Among these 30 most sensitive kinase inhibitors, 10 targeted the PI3K/Akt/mTOR signaling pathway, 7 targeted the cell cycle signaling pathway, and 6 targeted the protein tyrosine kinase signaling pathway (Fig. 5B and C). The IC₅₀ values of the three most sensitive drugs, dinaciclib (SCH727965), CUDC-907 and flavopiridol (alvocidib), were 0.02773 μM, 0.3043 μM, and 0.3204 μM, respectively (Fig. 5B and D). Interestingly, both F-BPDO1 and F-BPDO5 were highly sensitive to dinaciclib (SCH727965) (Fig. 5B–D, E and F; Figs. S8, A, C, D and E; Videos S13, S14, S15, S16, S17 and S18). Dinaciclib is a novel and potent small molecule inhibitor of cyclin-dependent kinase (CDK) 1, CDK2, CDK5, and CDK9 involved in the cell cycle. It has been demonstrated to have clinical activity in refractory chronic lymphocytic leukemia [[39], [40], [41], [42], [43]]. Our results suggest that dinaciclib may have promising applications in treating multidrug-resistant breast cancer. In conclusion, precryopreserved viable tumor tissues can be revived when needed and used for organoid culture, enabling the screening of effective drugs for patients with refractory tumors such as metastasis/recurrence or for application in developing novel drugs. This greatly enhances the utility of existing biobanks.

Fig. 5.

The identification of effective drugs against drug-resistant tumors by utilizing organoids derived from cryopreserved tumor tissues of drug-resistant patients. (A) Schematic workflow of the collection, processing, cryopreservation of tumor tissues, organoid culture, and high-throughput drug screening for effective drugs targeting drug-resistant tumors. (B) High-throughput drug screening (using 430 kinase inhibitors) identified 30 drugs as the most sensitive to organoids (F-BPDO1) derived from frozen breast cancer tissue with multidrug resistance. The bar plots display the IC₅₀ values of the 30 drugs. (C) Pie chart displaying the signaling pathways targeted by the 30 drugs most sensitive to F-BPDO1. (D) Drug response curves of the three most sensitive drugs for F-BPDO1 (dinaciclib (SCH727965), CUDC-907, and flavopiridol (alvocidib)). (E) 3D dual-spectral images showing the drug responses of F-BPDO1 after treatment with dinaciclib at concentrations of 0.02 μM and 0.05 μM for 96 h. The purple color represents viable organoids, while the yellow color represents dead organoids. (F) Quantification of the cytotoxic effects of the drugs on the organoids in Panel (E) through 3D live-cell imaging. The results are expressed as the mean ± SEM (n = 3). ∗∗ indicates P < 0.01; ∗∗∗ indicates P < 0.001.

3.6. Cryopreserved tumor tissues can be utilized for ALI culture and evaluating antitumor agents, including ICIs

Organoids can effectively preserve the genomic and transcriptomic features of the original tumor, offering promising prospects for precise cancer treatment. However, organoid culture results in the loss of tumor-associated fibroblasts and immune cells. In contrast, air-liquid interface (ALI) models overcome this limitation by successfully maintaining the tumor microenvironment, making them suitable for detecting antitumor drugs, including ICIs [37,44]. We successfully revived cryopreserved tumor tissues and generated ALI cultures (Fig. 6A). SMA-positive fibroblasts and CD45⁺, CD4⁺, CD8⁺, and F4/80⁺ immune cells were well preserved in the ALI model cultured with frozen and fresh tumor tissues (Fig. 6B and Fig. S9A). Furthermore, the proliferative activity and expression of PD-L1 in the cells were effectively restored (Fig. S9A). To further investigate the feasibility of utilizing cryopreserved tumor tissues for ALI culture and screening effective antitumor drugs, we first compared the drug responses of two pairs of cryopreserved and fresh mouse tumor tissue ALI models. After 96 h of drug treatment, the frozen and fresh HP6403T tumor tissue ALI models exhibited highly similar drug responses to the same concentration of BKM120 (Fig. S9, B-D). Importantly, the cryopreserved and fresh tumor ALI models generated with HP6654T not only displayed comparable drug responses to the targeted agent afatinib but also exhibited similar sensitivity to the monoclonal antibodies targeting the immune checkpoint PD-L1 (Fig. 6C–E). We further studied a pair of human breast cancer tissues and two pairs of human colorectal cancer tissues. Similar to the mouse tumor samples, we observed in all three pairs of human tumor tissue ALI models that cryopreserved samples exhibited similar drug responses to chemotherapy drugs and monoclonal antibodies targeting PD-1 (Fig. 6F–H and Fig. S9, E-J). In summary, our cryopreserved tumor tissue samples can be effectively utilized for ALI culture, enabling the preservation of the cellular constituents and tumor microenvironment of the original tumors, including fibroblasts, immune cells, and tumor cells. This methodology is valuable for screening chemotherapeutic and targeted therapy drugs and pharmacodynamic assessment of ICIs.

Fig. 6.

Cryopreserved tumor tissues can be utilized for air-liquid interface (ALI) culture and screening for antineoplastic agents. (A) Schematic workflow demonstrating the collection, processing, and cryopreservation of tumor tissues, ALI culture, and their application in screening antitumor drugs, including immune checkpoint inhibitors (ICIs). (B) Histological and immunohistochemical images demonstrating that similar to fresh tissues, cryopreserved tissues preserve the original mouse tumor tissue (HP6654T) structure and the infiltration of SMA-positive fibroblasts, CD45-positive leukocytes and CD8-positive T lymphocytes after ALI culture. (C) Light microscopy images of fresh and frozen mouse tumor tissues (HP6654T) cultured with ALI, treated with different drugs and stained with MTT. (D) The cytotoxic effects of drugs on fresh tumor tissues were analyzed using MTT in Panel (C). The results are expressed as the mean ± SEM (n = 3). ∗∗ indicates P < 0.01. ns indicates P > 0.05. (E) The cytotoxic effects of drugs on frozen tumor tissues were analyzed using MTT in Panel (C). The results are expressed as the mean ± SEM (n = 3). ∗∗ indicates P < 0.01. ns indicates P > 0.05. (F) Light microscopy images of fresh and frozen patient breast cancer tissues (BPT12) cultured with ALI and treated with different drugs and stained with MTT. (G) The cytotoxic effects of drugs on fresh tumor tissues were analyzed using MTT in Panel (F). The results are expressed as the mean ± SEM (n = 3). ∗∗ indicates P < 0.01. ∗∗∗ indicates P < 0.001. (H) The cytotoxic effects of drugs on frozen tumor tissues were analyzed using MTT in Panel (F). The results are expressed as the mean ± SEM (n = 3). ∗∗ indicates P < 0.01. ∗∗∗ indicates P < 0.001.

4. Discussion

Establishing a biobank is very important for translational medical research. It provides valuable resources for scientific research by collecting and utilizing biological samples, biological information, and data on a large scale and efficiently [2,45,46]. There are two main preservation methods for cancer tissues in biobanks: deep freezing and ambient temperature. Samples preserved by conventional methods can be used for the extraction of DNA, RNA, proteins, histological analysis, etc., but they are generally not suitable for culturing. They are considered "dead banks" incapable of generating living cells, which greatly limits the application value of biobanks. Establishing patient-derived xenograft (PDX) models, 2D cell lines, PDO models, or other models derived from patients can retain the activity of tumor cells within the tumor tissue. However, due to constraints such as time and cost, the scale of establishing these model libraries is generally limited. To overcome these issues, we propose establishing a biobank of live tissues from primary tumors during surgery or biopsy. Frozen samples can be revived for drug sensitivity testing based on PDO or ALI cultivation and/or other purposes.

Relapse and metastasis occur at high frequencies and present significant obstacles in tumor treatment, posing challenges to effectively controlling tumor growth with initial treatment methods and leading to unfavorable outcomes [47,48]. Recent research has made progress in unraveling potential mechanisms contributing to relapse and metastasis, including genomic mutations, activation of signaling pathways, and alterations in the tumor microenvironment [[49], [50], [51]]. Despite these advancements, limited information is available for patient stratification and predicting treatment outcomes [52]. Fulfilling an unmet critical clinical need requires an accurate model that faithfully replicates the characteristics of late-stage cancer and its corresponding response to treatment. PDOs have emerged as promising personalized models to address this need. They have played a pivotal role in preclinical research and investigations of individualized precision treatment for various solid tumors [20,22,53,54]. However, establishing PDO models in most current studies typically relies on fresh tumor specimens. For patients with recurrent and metastatic tumors, who often face rapid disease progression and formidable treatment challenges, there is an urgent need for personalized and precise treatments. Unfortunately, detecting or obtaining specimens from recurrent and metastatic tumors can be very difficult, which hampers the modeling and application of organoids.

In our previous study, we developed a methodology to freeze live tissue samples to address the challenge of transporting specimens over long distances [20]. However, technical limitations in the early stages, such as freezing large tissue blocks and delaying transfer to liquid nitrogen storage, resulted in a low success rate of organoid culture from cryopreserved tissues. Subsequently, we refined our tissue-freezing technique, resulting in a notable increase in the success rate of organoid culture. In this study, we standardized the technology by filtering small tissue blocks through a 500 μm mesh to obtain more uniform samples. The tissues were then gradually cooled and transferred to liquid nitrogen on schedule. This approach facilitates the long-distance transportation of specimens and enables us to freeze tumor tissue specimens obtained during early surgical procedures. In cases where tumor samples are unavailable during recurrence or metastasis, frozen tissues can be revitalized for organoid culture, providing potential personalized treatment options for patients.

This study explored the success rate of generating organoids from frozen tissues. We conducted tests on 42 frozen human tumor specimens from which organoid models were successfully established using fresh tissue samples from four different cancer types: breast, colorectal, lung, and kidney cancer. Our findings revealed that, except for two instances where organoid cultures failed, the remaining 40 cases were successful, resulting in a success rate of 95.2 %. Furthermore, it is worth noting that some tumor samples remained viable for organoid establishment even after being frozen for nearly four years. Moreover, our findings unequivocally demonstrate the applicability of this technique in effectively cryopreserving diminutive specimens, including biopsy tissues. In conclusion, our technology can sustain diverse tumor specimens for long-term freezing while preserving the viability of tumor cells within tissues.

Cellular interactions within the tumor microenvironment (TME) are crucial in cancer progression and closely influence drug response. ICIs block the suppressive effect of tumor cells on immune cells, effectively restoring or augmenting immune cell-mediated killing of tumor cells [55]. This approach introduces a fresh perspective and methodology for cancer treatment. However, clinical data indicate that the overall efficacy of ICI therapy is approximately 20 %, which falls short of expectations. Additionally, ICI therapy is accompanied by certain side effects and, in some cases, can lead to an excessive immune response [56]. Thus, accurate population prediction that would benefit from ICI treatment has become pivotal for its clinical application. In recent years, ALI cultivation has emerged as an individualized tumor model. It retains the essential components of the original tumor, including tumor cells, fibroblasts, and immune cells [37,44]. ALI culture not only facilitates the testing of conventional chemotherapy and targeted therapy drugs but also enables the evaluation of ICI efficacy. Importantly, our research highlights that cryopreserved tumor tissues can be utilized for ALI cultivation, yielding characteristics and drug responses comparable to those of ALI models cultured with fresh tissues.

In this study, we used tissues from a total of 42 human tumors, including breast cancer, colorectal cancer, lung cancer, and kidney cancer. We compared the success rates of culturing organoids using frozen and fresh tissues. We also studied the similarities in histological features, genomic characteristics, and drug responses of organoids derived from frozen tissues and fresh tissues. However, the sample size was small, with only two specimens available for kidney cancer patients. In the next step, it will be necessary to expand the sample size and investigate the success rate of applying this method to freeze and culture tissues from other types of tumors. Furthermore, this study's longest tissue freezing time was 46 months, and the impact of longer cryopreservation on cell viability in the tissue requires further investigation.

5. Conclusion

In summary, we successfully devised a simple and easy-to-operate technique for the cryopreservation of living tissues. Tumor tissues cryopreserved using this technique can maintain cell viability and serve as a reliable source for culturing organoids. Our results demonstrate that culturing organoids from frozen tumor tissues results in comparable success rates, histological and genetic characteristics, and drug responsiveness to those obtained from fresh tissues. Organoids derived from cryopreserved tumor tissues can also be continuously cryopreserved and passaged, enabling high-throughput drug sensitivity screening and the identification of potentially effective drugs to combat drug-resistant tumors. Cryopreserved tumor tissues are also applicable for ALI culture, preserving the original tumor microenvironment and enabling the evaluation of antitumor drug responses, including ICIs. Our research has expanded the biobank concept, laying the foundation for establishing a library of live tissue samples, which will provide immense support for precise cancer treatment and translational medical research.

Ethics approval and consent to participate

The ethics committees at the University of Macau and all hospitals involved in this study meticulously assessed and approved the study. We followed ethical guidelines and obtained informed consent from all subjects involved in the study.

CRediT authorship contribution statement

Ping Chen: Writing – review & editing, Writing – original draft, Visualization, Validation, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Jing-Bo Zhou: Writing – review & editing, Visualization, Software, Methodology, Investigation, Formal analysis, Data curation. Xiang-Peng Chu: Writing – review & editing, Methodology, Investigation, Formal analysis, Data curation. Yang-Yang Feng: Writing – review & editing, Visualization, Methodology, Investigation, Formal analysis, Data curation. Qi-Bing Zeng: Writing – review & editing, Methodology, Investigation, Formal analysis. Josh-Haipeng Lei: Writing – review & editing, Methodology, Investigation, Formal analysis. Ka-Pou Wong: Writing – review & editing, Resources. Tai-Ip Chan: Writing – review & editing, Resources. Chon-Wa Lam: Writing – review & editing, Resources. Wen-Li Zhu: Writing – review & editing, Resources. Wai-Kuok Chu: Writing – review & editing, Resources. Feng Hu: Writing – review & editing, Resources. Guang-Hui Luo: Writing – review & editing, Resources. Kin-Iong Chan: Writing – review & editing, Resources. Chu-Xia Deng: Writing – review & editing, Writing – original draft, Validation, Supervision, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article.