Intestinal crypts isolated from duodenal (CD, n = 5; controls, n = 5), jejunal (CD, n = 9; controls, n = 13), ileal (CD, n = 43; controls, n = 15), and colonic biopsy samples (CD, n = 12; controls, n = 8) from controls (n = 34) and patients with CD (n = 51) were cultured in Matrigel with maintenance medium and then organoid-forming efficiency was evaluated. Patient characteristics for enrolled patients are listed in Table 1. In both groups, the organoid-forming efficiency of duodenal crypts was the highest, followed by jejunal, ileal, and colonic crypts. The organoid-forming efficiency of the ileal crypts obtained from patients with CD was significantly lower than that of ileal crypts obtained from controls on day 3 (53.4% ± 4.0% vs. 39.6% ± 2.4%, p = 0.005), day 5 (38.9% ± 3.0% vs. 21.1% ± 1.7%, p < 0.001), and day 7 (29.2% ± 2.8% vs. 15.7% ± 1.5%, p < 0.001). The organoid-forming efficiency on day 7 of jejunal crypts obtained from CD patients was numerically lower than that of those obtained from controls (46.8 ± 2.6% vs. 38.1% ± 3.8%, Fig. 2).
Table 1
Characteristics of enrolled patients and number of attempts to construct organoids.
Total number of enrolled patients | 85 | Number of attempts to construct organoids | 110* |
Patients with CD (n) | 51 | Number of sub-culture (≥ passage 3) enabled organoids | 61 |
Age, years (mean ± S.D) | 36.2 ± 13.4 | CD-derived organoids (n) | 69 |
Male (n) | 43 | Duodenal organoids (n of attempts / n of subculture) | 5 / 4 |
Montreal classification (n) | | Mucosal healing / no duodenal involvement | 5 / 4 |
Age at diagnosis (A1/A2/A3) | 0 / 42 / 9 | Active ulcer | 0 / 0 |
Location (L1/L2/L3) | 1 / 34 / 16 | Jejunal organoids (n of attempts / n of subculture) | 9 / 8 |
Behavior (B1/B2/B3) | 20 / 27 / 4 | Mucosal healing / no jejunal involvement | 4 / 4 |
Medication at sampling | | Active ulcer | 5 / 4 |
None (sampling before treatment), (n) | 6 | Ileal organoids (n of attempts / n of subculture) | 43 / 23 |
Immunomodulator user, (n) | 25 | Mucosal healing / no ileal involvement | 8 / 5 |
Anti-TNFα user, (n) | 25 | Active ulcer | 35 / 18 |
Sampling modality | | Colonic organoids (n of attempts / n of subculture) | 12 / 2 |
Colonoscopy | 11 | Mucosal healing / no colonic involvement | 4 / 1 |
Single-balloon enteroscopy (n) | 40 | Active ulcer | 8 / 1 |
Controls (n) | 34 | Control-derived organoids† | 41 / 24 |
Age, years (mean ± S.D) | 50.6 ± 17.9 | Duodenal organoids (n of attempts / n of subculture) | 5 / 4 |
Male sex (n) | 24 | Jejunal organoids (n of attempts / n of subculture) | 13 / 11 |
| | Ileal organoids (n of attempts / n of subculture) | 15 / 8 |
| | Colonic organoids (n of attempts / n of subculture) | 8 / 1 |
*patients having samples from two (n = 21) and three (n = 2) different sites. |
Organoids grown from intestinal crypts were sub-cultured in maintenance medium and the control organoids became uniformly spheroid (> 90% of total organoids) after 2–4 passages. However, organoids derived from patients with CD had both enteroid and spheroid forms in the early passages and became uniformly spheroids after 4–6 passages. After six passages, the organoids derived from the controls and patients with CD exhibited consistent spheroid features and culturing behaviors (Fig. 3). The shapes of organoids cultured in the maintenance medium were similar regardless of their location; however, those cultured in differentiation medium tended to have different shapes, depending on their origin. The murine enteroids tended to have more budding structure compared to human enteroids. Ileal organoids typically exhibited budding, whereas jejunal organoids formed thick-walled structures (Fig. 4).
Previous studies have identified the cytotoxicity of intestinal organoids occurred in a concentration-dependent manner in response to TNFα [18–20]. To address the appropriate concentration of TNFα and the interval of administration to assess the epithelial regenerative ability, control organoids were tested for the changes in expression of the ISC and progenitor marker after TNFα treatment and organoid survival at the various concentrations of TNFα. The organoid viability—measured using MTT and the enteroid/spheroid ratio—decreased significantly with a gradual increase in the TNFα concentration. At TNFα concentrations of ≤ 10 ng/ml in the differentiation medium, changes in cell viability and morphology were negligible; however, the changes observed at TNFα concentrations of ≥ 30 ng/ml were notable (Supplementary Figures S1). The expression of LGR5 (active ISC), BMI1 (reserve ISC), HES1 (absorptive progenitor), and ATOH1 (secretory progenitor) increased within 24 h after TNFα treatment and then decreased (Fig. 5).
Based on these results, organoids derived from controls and CD patients —cultured over the long-term (≥ 6 passages)—were treated with 30 ng/ml TNFα every 24 hours for 10 days (Fig. 6A). The epithelial regenerative ability of intestinal organoids was evaluated using organoid reconstitution, MTT, EdU, and wound healing assays.
The organoid reconstitution rate of TNFα-treated organoids was significantly lower than that of TNFα-free organoids (jejunal organoids: 68.9% ± 12.4% vs. 47.4% ± 13.6%, p < 0.001; ileal organoids: 55.2% ± 12.3% vs. 32.8% ± 10.1%, p < 0.001). There was no significant difference in the organoid reconstitution rate between TNFα-free controls and CD patient-derived organoids; however, the organoid reconstitution rate of TNFα-treated CD patient-derived organoids was significantly lower than that of TNFα-treated control organoids (jejunal organoids: 55.5% ± 11.5% vs. 39.3% ± 10.6%, p = 0.011; ileal organoids: 40.2% ± 6.9% vs. 25.3% ± 6.6 %, p = 0.027; Figs. 6B and 6C).
The organoid viability was assessed using MTT; results showed that the formazan absorbance values of viable TNFα-treated organoids were significantly lower than those of TNFα-free organoids (OD of jejunal organoids: 0.96 ± 0.16 vs. 0.71 ± 0.12, p < 0.001; OD of ileal organoids: 0.83 ± 0.16 vs. 0.56 ± 0.16, p < 0.001). In the TNFα-enriched condition, the viable cells of jejunal and ileal CD patient-derived organoids was significantly decreased compared with those of the control organoids (OD of jejunal organoids: 0.79 ± 0.10 vs. 0.62 ± 0.07, p = 0.135; OD of ileal organoids: 0.66 ± 0.13 vs. 0.45 ± 0.10, p = 0.019; Fig. 7).
Two hours after EdU administration, EdU + cells were confirmed in the buds, which represented the intestinal crypt (Supplementary Figure S2). The number of EdU + cells was higher in TNFα-free organoids than in TNFα-treated organoids (91.9 ± 32.6 vs. 32.3 ± 22.2, p < 0.001). Although there was no significant difference in the number of EdU + cells between the control and CD patient-derived organoids in the steady state (82.7 ± 33.1 vs. 102.4 ± 30.5, p = 0.653), the number of EdU + cells were significantly lower in TNFα-treated CD patient-derived organoids than in TNFα-treated control organoids (49.0 ± 19.1 vs. 15.6 ± 7.7, p = 0.001, Fig. 8).
The wound healing assay showed that the unhealed wound area in TNFα-treated CD patient-derived organoids was significantly larger than that in control organoids at 8 h (50.5% ± 10.5% vs. 84.7% ± 12.3%, p < 0.001), 16 h (11.0% ± 4.8% vs. 64.3% ± 14.0%, p < 0.001), and 24 h after insert removal (1.7% ± 1.5% vs. 37.0% ± 8.5%, p < 0.001). There was no significant difference in the wound healing between the organoid derived from controls and CD patients. The unhealed wound area at 24 h was not significantly different between the TNFα-free and TNFα-treated control organoids (1.7% ± 1.5% vs. 16.0% ± 4.6%, p = 0.257), however showed significant difference between the TNFα-free and TNFα-treated CD patient-derived organoids (2.3% ± 2.5% vs. 37.0% ± 8.5%, p < 0.001). In addition, the unhealed wound area at 24 h of TNFα-treated CD patient-derived organoids was significantly larger than that in TNFα-treated control organoids at 24 h after insert removal (16.0% ± 4.6% vs. 37.0% ± 8.5%, p = 0.044). The wound-healing ability of TNFα-treated CD patient-derived organoids was significantly lower than that of TNFα-free control and CD patient-derived organoids, and TNFα-treated control organoids (Fig. 9)
RNA-seq was performed on endoscopic biopsy tissue samples from controls and patients with CD and organoids derived from these samples. The clustering heatmap and principal component analysis identified that the gene expression profile was clearly separated between the endoscopic biopsy tissue samples and the organoids derived from these samples. Furthermore, among the gene expression profile of organoids, there was a clear distinction between TNFα-free and -treated organoids (Figs. 10A and 10B). The epithelial lineage-specific gene expression was evaluated in the tissue samples and organoids derived from these samples (Fig. 10C). In endoscopic biopsy tissue samples, the expression of genes associated with the intestinal microbiota (NOD2, DEFA5, DEFA6, PLA2G2A, MUC2, and NARP6) and differentiated cells [enterocytes (ECs): SI, APOC3, ALPI, and APOA1; goblet cells (GCs): MUC2; Paneth cells (PCs): WNT3, ARG2, DLL1, and DLL4; and enteroendocrine cells (EECs): CCK, CHGA, CHGB, and NEUROG3)] was increased in comparison to organoids. The expression of EC markers such as VIL1, KLF5, and KRT5 and GC markers such as TFF3 and MUC13 were up-regulated in biopsy tissue samples and TNFα-free control organoids (day 6 and 9 organoids). In the TNFα-treated control organoids, the expression of ISC markers, such as LGR5, OLFM4, and TNFRSF19, was increased compared with those of TNFα-treated CD patient-derived organoids. The alterations of ISC properties can be affect cell proliferation and wound healing ability