A Method for Cryogenic Preservation of Human Biopsy Specimens and Subsequent Organoid Culture.

Human tissue–derived gastrointestinal (GI) organoids have revolutionized the study of human biology,1, 2, 3 and are powerful tools for studying human physiology and disease; however, generation of organoids is limited by access to human tissue and a short window of viability for human samples, putting a hard limit on the time and place in which a patient sample can be used for research. These restraints mean that a laboratory must be relatively geographically close to the source of collection to use the sample within the window of viability. Patient-derived organoids also are being used for drug development, stem cell therapies, and personalized medicine; however, it is not always feasible to prospectively develop organoid line efforts given the time and labor involved.5, 6

To overcome these limitations, we sought to develop a practical method to cryopreserve live human biopsy tissue, which then could be stored or shipped frozen and later thawed to generate new cultures of GI epithelium-only organoids (also referred to as enteroids/colonoids). Here, we describe a simple and robust method to cryopreserve human biopsy specimens that subsequently could be thawed and used to generate epithelium-only organoids.

Results/Discussion

Endoscopic biopsy specimens were collected with 8-mm2 biopsy forceps (average biopsy size, 5 mm2), transported to the laboratory, divided further into 2- to 3-mm2 fragments, then transferred into a cryopreservation vial containing freezing medium. We tested 2 freezing media formulations: a complex medium (LWRN, fetal bovine serum, dimethyl sulfoxide, Y27632, CHIR99021; see Supplementary Materials and Methods section)7, 9; and a simple freezing medium (Dulbecco's modified Eagle medium/F12, 10% fetal bovine serum, and 10% dimethyl sulfoxide). By using a Mr. Frosty (ThermoFisher Scientific, Waltham, MA) cell-freezing container, the biopsy fragments were frozen to -80°C overnight. Both freezing media formulations performed equivalently (data not shown), thus we focused on the simple formulation. Upon thawing the cryopreserved biopsy specimens, 4 methods were developed for thawing and subsequent establishment of organoid cultures (Figure 1). Frozen vials were thawed at 37°C, the tissue was washed in recovery media (see Supplementary Materials and Methods section, Figure 1). For techniques 1 and 3, tissue was partially digested with dispase or EDTA followed by mechanical perturbation to loosen adhesion between cell layers (Figure 1, Supplementary Figure 1, Supplementary Figure 2, Supplementary Figure 3). The epithelium was pelleted by centrifugation and then resuspended in recovery medium. For techniques 1 and 3, isolated epithelium was embedded in Matrigel (Corning, Bedford, MA) where it was cultured for 3 days in recovery media.

Figure 1

Schematic of the methods developed that allow successful generation of patient biopsy-derived epithelial organoids after cryopreservation. (A) The process of endoscopic biopsy collection followed by cryopreservation in a simple freezing medium can be accomplished in typical clinical settings with readily available equipment. After cryopreservation, organoid cultures can be established using 3 different techniques. Technique 1 (left) used a dispase digestion to isolate crypts from freshly thawed tissue. Technique 2 (middle–left) adds a step relative to technique 1 in which the entire biopsy specimen is embedded in Matrigel and allowed to repair from the freezing process before tissue digestion and crypt isolation. Technique 3 (middle–right) is very similar to technique 1 but uses a gentle EDTA treatment to separate the epithelium from the mesenchyme. Technique 4 (right) involves isolation of epithelial crypts before cyropreservation so that, upon thawing, cultures can be seeded immediately without additional tissue manipulations. All 4 techniques result in pure organoid cultures 2 weeks after initially thawing the biopsy specimen. (B) Organoids lines were generated using techniques 1 and 2. (C) Organoid lines were generated using technique 3. (D) Organoid lines were generated using technique 4.

Supplementary Figure 1

Organoids derived from frozen tissue have delayed initial growth characteristics. Biopsy specimens from duodenum (columns A and B), ileum (columns C and D), and colon (columns E and F) can be used to derive intestinal organoids before and after freezing. Epithelial fragments from fresh biopsy specimens (A2, C2, E2) are larger and contain recognizable crypts compared with the small fragments from frozen (B2, D2, F2) samples. Growth kinetics are delayed in cultures established from frozen tissue (B2–B5, D2–D5, F2–F5) relative to fresh tissue (A2–A5, C2–C5, E2–E5), but attain normal growth kinetics and size by 10 days in culture (A6–F6). Biopsy images are representative and not necessarily tissue-matched to the subsequent organoid images.

Supplementary Figure 2

Four methods are developed that allow successful generation of patient biopsy specimens. Technique 2 (left; A1 and A2) adds a step relative to technique 1 in which the entire biopsy specimen is embedded in Matrigel and allowed to repair from the freezing process before tissue digestion and crypt isolation. Technique 3 (middle; B1 and B2) is very similar to technique 1 but uses a gentle EDTA treatment to separate the epithelium from the mesenchyme. In technique 4 (right; C1 and C2), whole crypts were isolated, subsequently cryopreserved, and recovered to generate organoids. All 4 techniques result in pure organoid cultures approximately 2 weeks after initially thawing the biopsy specimen. Biopsy images are representative and not necessarily tissue-matched to the subsequent organoid images.

Supplementary Figure 3

Organoids can be derived from frozen biopsy specimens of stomach and colon adenoma. Biopsy specimens from stomach (columns A and B) and colon adenoma (columns C and D) can be used to derive organoids before and after freezing. Organoids from fresh biopsy specimens (A1–A3, C1–C3) are larger compared with the small organoids from frozen (B1–B3, D1–D3) samples, but attain normal growth kinetics and size after the first split for both organoids (A4–B4, C4–D4). Biopsy images are representative and not necessarily tissue-matched to the subsequent organoid images.

We also tested a second thawing technique that omitted enzymatic digestion (technique 2) (Figure 1, Supplementary Figure 2). In this case we thawed the biopsy specimen, washed with recovery media, and embedded whole tissue fragments in Matrigel. These cultures were grown in recovery media for 1 week, where we observed epithelial cysts growing from the biopsy fragment (Supplementary Figure 2A). Tissue was removed from Matrigel and the epithelium was dissociated enzymatically using dispase, similar to technique 1, and cultured further (Figure 1). We also found that whole fresh crypts could be isolated, cryopreserved, and recovered to generate organoids (technique 4) (Figure 1, Supplementary Figure 2C).

By using techniques 1 and 2, we established 20 frozen biopsy-derived organoid lines from 4 different regions of the GI tract (stomach, duodenum, ileum, and colon) from healthy patients, and 1 organoid line from adenomatous tissue (Figure 1B). We generated 3 patient-specific organoid lines using technique 3 (Figure 1C), and 7 patient-specific organoid lines using technique 4 (Figure 1D). All techniques had a 100% success rate, showing the robustness of the method. The initial growth of organoids from frozen samples is delayed when compared with freshly isolated epithelium (Figure 2A), but frozen organoids eventually catch up with fresh tissue organoids and are indistinguishable after the first passage (Figure 2A and B).

Figure 2

Organoids derived from fresh and frozen tissue appear morphologically indistinguishable after passage, and transcriptional profiles of organoids derived from fresh and frozen biopsy specimens are nearly identical. (A) The size of organoids derived from different sections of intestine were measured at 4, 7, and 10 days of growth and after the first split. (B) Organoids derived from duodenum (representative biopsy specimen shown at top left, organoids columns A and B), ileum (biopsy top middle, organoids columns C and D), and colon (representative biopsy specimen shown at top right, organoids columns E and F) have different morphologic characteristics when derived from fresh tissue (A1–A4, C1–C4, and E1–E4). Duodenum-derived organoids grow as smooth cystic structures regardless of whether they originate from fresh or frozen tissue (A1–A4 and B1–B4). Organoids from fresh ileum and colon grow with distinct budded morphology whereas those derived from patient-matched frozen tissue (D1–D4 and F1–F4) lack these morphologic distinctions during the early phases of culture establishment. After the first passage and for the remainder of the observed lifetime of the culture, fresh and frozen organoid morphologies cannot be distinguished (A4–F4). (C and D) Clustering organoid RNA sequencing samples by (C) source of variance with principal component analysis or by (D) sample similarity using Pearson correlation. Both analyses generate groups from the same gastrointestinal region regardless of derivation from fresh or frozen tissue biopsy specimens. (E) Differential expression analysis comparing fresh-derived duodenum organoids with frozen-derived duodenum organoids (left), or fresh-derived colon organoids with frozen-derived colon organoids (middle), show a near lack of significant differences in gene expression. In marked contrast, comparing organoids from different regions of the small intestine, such as duodenum and ileum (right), yields more than 9% of genes showing significantly different expression. *Significant difference denotes an adjusted P value of ≤ .01 and fold difference ≥2 (log2 fold difference ≥1 or ≤-1). †Biopsy images are representative and not necessarily tissue-matched to the subsequent organoid images.

To determine if cryopreservation caused any molecular differences within the organoids, we performed RNA sequencing on a cohort of fresh tissue organoids (n = 10) and cryopreserved tissue-derived organoids (n = 5). Principal component analysis showed that the strongest driver of variability was the GI region from which organoids were derived (Figure 2C). Similarly, using Pearson correlation, fresh and frozen organoids from the same region of the intestine clustered together, and there was no clustering based on fresh vs frozen status (Figure 2D). Finally, we performed differential expression analysis to compare organoids derived from fresh vs frozen biopsy specimens from the same region of the intestine (ie, duodenum or colon). Strikingly, frozen and fresh organoids were nearly identical, with only 46 (0.24% of all genes expressed) and 72 genes (0.35% of all genes expressed) showing significant expression differences in the duodenum and colon, respectively (log2 fold change ≥ 1 or log2 fold change ≤ -1; P ≤ .01) (Figure 2E). By contrast, comparing duodenum vs ileum using the same method showed 1777 (9.32% of all genes expressed) differentially expressed genes (Figure 2E).

Organoid generation from cryopreserved biopsy specimens was robust and patients ranged in age from 2 to 70 years and included both males and females. We kept a subset of samples frozen for times ranging from 3 days to 10 months with successful organoid cultures derived in each case (Figure 1B–D). To test the idea that cryopreserved biopsy specimens can be shipped long distances, 2 patient samples were cryopreserved at the University of Michigan and shipped on dry ice to Baylor College of Medicine, where organoid cultures were established successfully from both patients (Figure 1, samples 104 and 105). Biopsy specimens from 3 patients were shipped on dry ice from Ankara, Turkey, to the University of Michigan, where organoids were established successfully from these samples (Figure 1B, samples 132, 133, and 134). These results indicate that GI tissue can be cryopreserved, shipped long distances, and cultured from a diverse human demographic.

In summary, we have shown human biopsy specimens from multiple regions of the GI tract can be cryopreserved and, upon thawing, used to establish long-term organoid cultures. We speculate that this technically simple process will be adaptable to biopsy specimens from other organs/tissues, and that the major obstacle will be the identification of tissue-specific recovery conditions after thawing. Frozen tissue samples now can be shipped across the globe, effectively freeing patients, hospitals, clinics, researchers, and diagnostics laboratories from the necessity of geographic proximity.