Ten human colon and six rectal tissues were obtained from 16 patients undergoing surgery at the University Center for Gastrointestinal and Liver Disease (Clarunis), Basel, Switzerland. Written informed consent was obtained from all patients. The study was performed in accordance with the Helsinki Declaration and approved by the ethics committee (Ethics Committee of Basel, EKBB, no. 2019-02118). Data were collected retrospectively in a non-stratified and non-matched manner including patient age, sex, treatment, American Joint Committee on Cancer (AJCC) clinical stage, bowel preparation, and antibiotic prophylaxis ( Table 1 ).

Prior to surgery, patients received a bowel preparation that combines oral antibiotics and mechanical cleansing of the colon (using a macrogol solution for cleansing) as follows: 1 day before surgery patients underwent bowel preparation and received oral antibiotic treatment of 2 500 mg tablets of Neomycin three times a day and 2 500 mg tablets of Metronidazole three times a day (16:00–20:00–24:00). Moreover, 30’ before surgery, intravenous antibiotics were administered. Patient-specific treatments are summarized in Table 1 .

Human tissue samples were collected in DMEM (without any antibiotics), placed on ice, and processed after a maximum of 3 h post-surgery. The tissue was then divided into 5 pieces ( Figures 1 and 2A ), attempting to retain the tissue homogeneity regarding macroscopic tissue morphology and size among them. Subsequently, while the “no wash” condition was transferred to a 2-ml Eppendorf and kept on ice in full advanced medium for the entire washing procedure, the remaining 4 pieces were moved into 4 different wells of a 6-well plate where they were each then washed with the 4 different solutions ( Figure 2B ).

All the solutions were stored at 4°C.

Each washing solution was carefully added into the well with a serological pipette to avoid any possibility of spill-over between the different conditions until the tissues were fully covered. After 5 min, solutions were removed with a glass Pasteur pipette linked to a vacuum pump (also in this case, different pipettes were used for the 4 different conditions) and the washing was performed for 2 additional times ( Figure 2B ).

Following the washing steps, all tissue pieces were processed for the generation of PDOs according to standard protocol (11). Our success rate for PDO generation was 62.5%. Briefly, tissues were minced into small pieces and subsequently enzymatically digested in 1 ml of advanced DMEM/F-12 FULL medium containing 2.5 mg/ml collagenase IV (Worthington, #LS004189), 0.1 mg/ml DNase IV (Sigma, #D5025), 20 μg/ml hyaluronidase V (Sigma, #H6254), 1% BSA (Sigma, #A3059), and 10 μM LY27632 (Abmole Bioscience, #M1817) for 1 h at 37°C under slow rotation and vigorous pipetting every 15 min ( Figures 2C, D ). Of note, the DMEM used to prepare the enzymatic digestion cocktail for the “no washed” tissue was the same as where the tissue itself was kept in during the washing step. The tissue lysate was filtered through a 100 μM cell strainer, centrifuged at 300 g for 10 min, and then treated with Accutase (Sigma, #A6964) for 15 min at room temperature in order to dissociate the remaining fragments ( Figure 2E ). After 5 min of centrifugation at 300 g, the cell pellet was suspended in PBS and cells were counted using trypan blue (GIBCO, #15250061), Countess™ Cell Counting Chamber Slides (Invitrogen, #C10228) in Countess™ II FL Automated Cell Counter (Invitrogen, #AMQAF1000) ( Figure 2F ). Then, the same amount of cells from each washing condition was mixed with growth factor reduced Matrigel (Corning, #356231) and seeded as drops in a tissue-culture dish ( Figure 2G ). After polymerization of the Matrigel, a medium supplemented with growth factors for CRC tissue was added to the cells (14) ( Figure 2H ). Organoids were monitored every day for a period of 7 to 9 days for the presence of microbial contaminants. Organoid medium was changed every 3 days and, when needed, organoids were passaged after dissociation with 0.25% Trypsin-EDTA (GIBCO, #25200056).

Immunohistochemical staining was performed on a Benchmark immunohistochemistry staining system VENTANA BenchMark Special Stains system (Roche) using anti-CDX2 (Clone EPR2764Y, catalog number 760-4380), anti-CK20 (Clone SP3, catalog number 790-4431), and primary antibodies as substrate. Images were acquired using an Olympus BX46 microscope and evaluated by an experienced pathologist (CE).

Bacteria detection in contaminated medium was conducted using the Remel™ Gram Stain Kit (Thermo Fisher, #R40080) following the manufacturer’s instructions. Briefly, a smear of bacteria-containing medium was placed on a glass slide and air-dried for 5 min. The smear was then quickly fixed with the flame of a Bunsen burner and stained with a crystal violet solution for 1 min at room temperature. Subsequently, the slide was rinsed under running tap water to remove excess crystal violet. Gram iodine mordant was applied for 1 min and briefly washed in tap water. To remove non-specific crystal violet staining, a Gram decolorizer solvent was applied for 30 s and then quickly rinsed under running tap water until the water ran clear. Finally, the slides were stained with Gram Safranin for 30 s and allowed to dry and then coverslipped. Images were acquired using an Olympus BX46 microscope and evaluated by an experienced pathologist (CE).

After harvesting fungi from the culture, they were briefly washed with PBS and then fixed in 10% formaldehyde overnight. Fungi was embedded in paraffin; 5-µm slices were cut and put on a slide. The staining was performed by incubating the deparaffinized slides with 0.5% periodic acid solution for 5 min, then stained with Schiff’s reagent for 10 min, followed by counterstaining with hematoxylin solution for 2 min. All steps were performed at room temperature, and slides were rinsed with tap water after each step. Images were acquired using an Olympus BX46 microscope and evaluated by an experienced pathologist (CE).

As for PAS staining, we produced 5-µm slices of embedded fungi and placed them on glass slides. The staining was performed using the automatized “VENTANA BenchMark Special Stains system” from Roche and the GMS II Staining Kit (Roche, #860-028). Images were acquired using an Olympus BX46 microscope and evaluated by an experienced pathologist (CE).

To check mycoplasma contamination, we performed a slightly modified PCR protocol using specific primers as described previously (28). Briefly, contaminated and uncontaminated culture medium was harvested and boiled at 95°C for 5 min or kept at 4°C up to 1 week before processing. PCR reaction was conducted with AmpliTaq Gold™ 360 Master Mix (Thermo Fisher #4398881) as previously described (28). Electrophoretic run was performed on 1.5% agarose gel and bands height were compared with TrackIt™ 100 bp DNA Ladder (Invitrogen, #10488058).

Statistical analyses were performed using Wilcoxon tests (Prism GraphPad 8), Spearman correlation, and Chi-square test. All statistical tests were 2-sided and p < 0.05 was considered statistically significant.

PBS solution, P/S, or combination of complex antibiotics have been used in washing solutions prior to tissue processing for CRC PDOs (10, 11, 25, 26). However, lack of standardization makes it difficult to conclude which is the best protocol in preventing microbial contamination. To determine the optimal components of the washing solution for CRC-PDO generation, we used a cohort of 16 CRC patients and compared the antimicrobial efficacy of 4 different solutions compared to the no-washing step: PBS, PBS supplemented with P/S, PBS supplemented with Primocin, or PBS supplemented with P/S and Primocin ( Figures 1 and 2 ). The samples included in the study were mostly from male patients (56%) with a diagnosis of colon carcinoma (62.5%), and a median age of 73.5 years (range: 34–86 years). All patients included in the study received antibiotic prophylaxis on the day of the surgery ( Table 1 ).

Tissue samples were divided into 5 pieces and randomly assigned to a washing condition ( Figures 1 and 2 ). All pieces underwent three washing steps on ice, for a total of 15 min, with the corresponding washing solution, with the exception of the no-wash condition (control) that was kept on ice without changing the solution. The following steps for the PDO generation, including tissue dissociation, cell seeding, and medium composition, did not differ among the washing conditions. The growth of PDOs was monitored every day and inspected for the presence of microbial contamination using an inverted microscope. Microbial contamination was usually detected between day 3 and 5 post tissue processing. In some cases (3 out of the 10 cases contaminated in the no-wash condition), the detection of microbial contamination matched a change in the color of the culture medium, indicating a change in the pH of the medium due to bacterial outgrowth. Representative images of the changes in the color of the medium and detection of bacterial contamination are shown in Figure 3A .

The presence of microbial contamination was detected in 62.5% of the samples that were not washed prior to tissue processing, while no contamination was detected in samples washed with solution containing Primocin ( Figure 3B and Table 2 ). Interestingly, while washing steps performed with PBS-P/S solution reduced the risk of contamination by 37.5%, compared to the PBS alone, in 25% of tissues, microbial contamination was detected ( Figure 3B and Table 2 ). CRC-PDO medium contains both P/S and Primocin at the same concentration used in the washing solution; thus, supplementing the culture medium is not sufficient to prevent bacteria contamination. Of note, clinical parameters such as size of the tumor or presence of ulcer were not associated with the occurrence of contamination (p = 0.1149, p = 0.6990, and p = 0.1432, respectively).

Despite the contamination, we observed healthy organoids also in the presence of bacteria. However, the bacterial contamination may be responsible for matrix drop degradation (loss of extracellular matrix support for the CRC-PDO growth), thus indicating that preventing microbial contamination is a critical step of tissue processing and organoid generation ( Figure 3C ). A comparison of Matrigel-embedded PDO derived from P107 at day 3 and day 4 after using PBS as a washing solution: at day 3, Matrigel drop borders are still intact despite bacterial contamination (dark area on the bottom right, Figure 3C ). Inversely, at day 4, bacterial overgrowth led to Matrigel degradation, suggested by the curved border of residue from the degraded drop, indicated by the 4 red arrows. As a consequence of the degradation, while PDOs at day 3 are not surrounded by bacterial suspension, at day 4, bacteria cover all areas of the well ( Figure 3C ).

Taken together, our results indicate that implementation of multiple washing steps (3×, 5 min each) with PBS-Primocin solution prior to tissue dissociation is able to eliminate the inherent high risk of microbial contamination in CRC-PDO cultures. Moreover, we show that the presence of P/S and Primocin in the PDO culture medium is not sufficient to prevent bacterial contamination in CRC-PDO cultures.

To determine the impact of the additional washing steps on PDO generation, we checked cell viability immediately before embedding in Matrigel using the Trypan blue staining assay. The presence of P/S alone in the washing solution leads to a significantly lower percentage of viable cells compared to a wash performed with PBS alone and Primocin-PBS ( Figures 4A, B ; Wilcoxon test, p = 0.0240 and 0.0061, respectively). Likewise, washing steps performed with P/S-PBS showed a high tendency to negatively impact cell viability compared to the no-washing condition ( Figure 4C ; Wilcoxon test, p = 0.0856). No significant difference in cell viability was observed when comparing PBS-P/S solution and PBS-P/S-Primocin ( Figure 4D ; Wilcoxon test, p = 0.2458), suggesting that the presence of Primocin is not sufficient to overcome the negative impact of P/S. Moreover, no difference was noticed when comparing cell viability assessed after using other washing solutions that didn't contain P/S ( Figures 4E, F ).

Our results suggested that the presence of P/S in the washing solution could impact cell viability and therefore the success rate of CRC-PDO generation.

To characterize the type of microbial contamination, we performed GRAM staining in the presence of a suspected bacterial contamination, and a PAS and Grocott staining when contamination appeared related to fungi. GRAM staining was consistent with the presence of GRAM-positive chain-forming bacillus and GRAM-positive cocci ( Figure 5A ). These bacteria seem to be capable of aerobic metabolism (29, 30), since they were kept in a classic incubator for cell culture where the O₂ percentage is maintained at 18.6% (31).

Although mycoplasma is not normally present within the intestinal flora, it may be present in neoplastic disease and, if present, can alter organoid formation by affecting their structure, number, and size (32, 33). Therefore, we investigated the presence of mycoplasma contamination in the organoids generated from each washing condition. None of the organoids were positive for mycoplasma contamination ( Figure S1 ).

Bacteria contamination was found to be the major contaminant in our patient-derived organoid cultures. Indeed, bacterial and fungal contamination was detected in only 1 (P135) out of the 10 cases with microbial contamination. To perform a morphological characterization of the fungi, we performed PAS and Grocott staining. We observed yeast, hyphae, and pseudohyphae structures consistent with Candida albicans ( Figure 5B ). Interestingly, this contamination appeared only when tissues were not previously washed with Primocin-containing PBS ( Figure 5C ), suggesting that, as for bacteria contamination, growth medium supplemented antibiotics with is not sufficient to protect cultures from fungal-yeast contamination.

In our study, we generate 10 PDOs from different patients. In order to evaluate whether PDOs recapitulate primary tumor morphology and marker expression features, we performed H&E and immunohistochemistry staining. Histologic and immunophenotypic analysis was performed by an expert pathologist with expertise in gastrointestinal pathology confirming that our 10 PDOs recapitulated and maintained the histologic profile and cellular morphology of the tumors from which they originated ( Figure 6 ).