Establishment and characterization of canine mammary tumoroids for translational research

Feasibility of tumoroid culture from freshly resected canine mammary tumors

Canine mammary tumor was collected in the operating room at the time of tumor resection. For the characterization of the tumoroids’ cultures, 6 on 31 patients of the established biobank were included (Additional File 1: Table 1). For all of them, the resection was a primary mammary tumor. Patient’s age ranged between 5 and 14 years old and were all female. Based on the 2010 histologic classification for canine mammary tumors, the 6 tumors were annotated (Additional File 1: Table 1) [25]. The 6 tumors were characterized with the most important and frequent biomarkers of breast cancer: estrogen receptor (ER), progesterone receptor (PR), and HER2. Among the 6 tumors, 4 have a triple-negative signature, signifying the absence of HER2, ER, and PR proteins expression (TM-02, TM-03, TM-05, TM-06), while 2 tumors are of luminal subtype with PR expression (TM-01) or PR/ER expression (TM-04) (Additional File 2: Fig. S1 and Additional File 1: Table 1). In addition, the positive Ki67 labeling of each tumor was evaluated. We found that all 6 tumors showed Ki67 positive cells, with variable levels (Additional File 3: Fig. S2).

After tumor resection, the tumor fragment was divided into three pieces: the first piece was kept fresh for tumoroid culture generation, the second one was frozen without prior fixation, and the last piece was fixed in PFA and cryopreserved (Fig. 1A). Frozen tissue was used for proteomics while fixed tissue was used for histology.

Fig. 1figure 1

Breast cancer tumoroids culture established from canine patient. A Diagram presenting the strategy used to culture tumoroids from a canine mammary tumor. B Representative images of canine tumoroids at different time points of culture. Scale bar (200 μm) is indicated. C H&E staining comparison between tumoroids and the tumor of origin for three different dog patients (n = 3)

Fresh tissue pieces were mechanically and enzymatically dissociated to obtain single cell suspensions which were plated in Matrigel drops and overlaid with optimized mammary tumoroids culture medium. Cultures were followed by microscopy for evidence of tumoroids formation. We successfully generated tumoroid cultures from 31 of 33 tumor samples, an establishment success rate of 94%, with long-term expansion. Indeed, all tumoroids were grown for at least 42 days (6 passages) (Fig. 1B). Majority of tumoroid lines were cryopreserved. The tumoroids morphologically reflected the original tumor they were derived from (Fig. 1C). Tumoroids presented patient-specific heterogeneous morphologies, ranging from compact structures (TM-02) to more irregularly structures (TM-03 and TM-04).

Canine mammary tumoroids can be generated from both fresh and frozen cells and can be cryopreserved with similar histological and molecular features

Next, we wanted to evaluate whether tumoroids could be generated from frozen cells while keeping the same characteristics as fresh tumoroids. From the primary tumor sample, we divided the cell suspension into two parts: one part kept fresh for direct tumoroid formation, named “fresh tumoroids” thereafter, and the second part was frozen for indirect tumoroid formation, named “tumoroids from frozen cells” thereafter. In addition, in order to characterize our biobank, we wanted to make sure that cryopreservation did not affect the tumoroids features. We therefore compared these two types of tumoroid cultures to thawed tumoroids, named “frozen tumoroids” thereafter (Fig. 2A). Tumoroids from these different culture conditions were left in culture during 4–5 weeks (date 1) or 6-–7 weeks (date 2) and compared to study tumoroids drift overtime.

Fig. 2figure 2

Study of canine mammary tumoroids drift. A Diagram showing the strategy used to generate the different types of tumoroids. After tumor digestion, a part of the cells were frozen and then thawed to generate the “Frozen Cell Tumoroids.” The remaining cells were used to generate the “Fresh tumoroids.” A part of these tumoroids were cryopreserved and then thawed, corresponding to the “Frozen Tumoroids.” The three types of tumoroids were compared at the same time point post-culture at date 1 (4–5 weeks) or date 2 (6–7 weeks). B Evaluation of the efficiency of tumoroids formation for the 6 cases in triplicates (n = 6). The average of the triplicates is shown, and error bars mean SD. C Representative images of the three types of canine tumoroids at passage 1, 2, and 3 after date 1. Scale bar (200 μm) is indicated. D H&E staining comparison of fresh, frozen, and frozen cells tumoroids. Scale bar = 100 μm (n = 3)

Histologic and molecular drifts of tumoroids after cryopreservation and after long-term culture were studied. First, the efficiency of tumoroids formation was evaluated in all 6 tumors in triplicate. All tumoroids show an exponential growth curve. Almost all of them increase their growth between 230 and 335% compared to the initial time (T0) after 7 days of culture. TM-01 showed a more modest growth, reaching values of about 85% augmentation of growth after 7 days of culture (Fig. 2B and Additional File 4: Table 2). In addition, the culture of tumoroids was successful for each culture condition and after serial passages as well (Fig. 2C). Tumoroid formation was not found visually different between cryopreserved cells and fresh cells. Tumoroid cultures from fresh and frozen cells could be similarly long-term cultured and passaged (Fig. 2C). At the histological level, tumoroids derived from fresh cells, frozen cells, or after cryopreservation retained the same architecture. Figure 2D presents representative images of H&E staining of tumoroids derived from three different tumors. Tumoroids derived from TM-02 were compact while tumoroids derived from TM-03 and TM-04 were more irregular whatever the culture condition and time in culture. The freezing procedure did not affect tumoroids morphology.

The ER, PR, and HER2 expression profiles of breast cancer tumoroids were compared with their original breast cancer tissues. For this, 3 tumors were used: TM-03 (triple-negative subtype), TM-04 (luminal subtype) (Fig. 3), and TM-05 (triple-negative subtype) (Additional File 5: Fig. S3). The results showed that the tumoroids maintain the same expression profile of the tumor of origin. In the case of TM-03 and TM-05, tumoroids present a triple-negative subtype as the tumor they are derived from. In the case of TM-04 tumoroids, we can observe ER and PR positive cells similar to the tumor of origin. This result supports those previously demonstrated in humans and dogs, in which the majority of tumoroids reflect the same histological characteristics as the tumor from which they originate [6].

Fig. 3figure 3

Histology and receptor status (ER, PR, HER2) of breast cancer tumoroids. Comparative histological and immunohistochemical images of breast cancer tumoroids and their original breast cancer tissues

Finally, we verified if the proliferation of the tumoroids in all three conditions was similar. We used tumors of the two different subtypes, a luminal TM-04 and a triple-negative TM-05, to answer this question. The proliferative activity of the tumoroids was determined by the percentage of Ki67 + cells with respect to the total number of cells. Proliferation activity of tumoroids do not show a significant difference between Fresh, Frozen and FrozenCell tumoroids. In TM-04, the percentage of Ki67 + cells remained around 7% for all conditions (Fresh—7.64%, Frozen—7.79% and FrozenCell—7.41%), as well as in the TM-05, around 7% were Ki67 + (Fresh—8.61%, Frozen—6.96% and FrozenCell—6.29) (Fig. 4 and Additional File 4: Table 2).

Fig. 4figure 4

Proliferative activity of the tumoroids. A Comparative quantification of the percentage of Ki67 + cells in tumoroids Fresh, Frozen, and FrozenCell of the tumor TM-04 and TM-05. The average of the triplicates is shown and error bars mean SEM (n = 3, P value ≤ 0.05). B Representative images of canine Ki67 staining (green) of fresh, frozen and frozen cells tumoroids of the tumor TM-05

Similar proteomic profiles are observed between tumoroids generated from fresh and frozen cells over time while cryopreservation seems to trigger a more pronounced molecular drift

We have shown that the freezing procedure as well as the passages did not impact the morphology of tumoroids neither their histological features. In order to understand, whether the frozen tumoroids or tumoroids made from frozen cells kept similar molecular profiles as the fresh tumoroids or the original tumor, we have performed a large-scale unbiased proteomic analysis. The study was carried out on 3 tumors: TM-01, TM-02, and TM-03, of which the three types of tumoroids were made and compared with each other and with the original tumor. For this, the extracted proteins were quantified and the same amount of proteins was used.

To understand if there was any molecular drift over time, we have analyzed the proteome of tumoroids at two dates of tumoroid passage (D1 and D2), 1796 proteins were identified in the six conditions: D1-Fresh, D1-FrozenCell, D1-Frozen, D2-Fresh, D2-FrozenCell, and D2-Frozen (62% of all the proteins identified) (Fig. 5A) (Additional File 6: Table 3). The D1-Frozen tumoroids seem to be the most different compared to all the other conditions as shown on Fig. 5A. In fact, 343 proteins were identified in all conditions except in D1-Frozen. This may be due to a lack of protein diversity in this condition, as all samples were quantified to have the same amount of proteins. However, if we observe more closely the proteins lacking in the condition of D1-Frozen (Fig. 5B), we found several proteins involved in cellular metabolic process and energy pathways such as the metabolic process of cellular aromatic compound (with HMOX2, FAH, ACSL4, PMVK, AARS2, GGH, FECH, PRPSAP1, HMGCS1, and ADA) and the cellular catabolic process (with HMOX2, PLA2G4A, MANBA, FAH, HEXB, ARG2, BLMH, SARDH, ECI1, BCKDHB, and ADA). In addition, we found proteins involved in the inhibition of apoptosis (API5, RIPK3, FIS1, FAF1, and CTNNBL1), in cell growth (GOLGA3, SOD2, COL5A2, SYNE2, ARVCF, GOLGA2, PARVA, CGN, ADD3, TIMP3, DSG2, VCAN, PAK4, ELMO3, FKBP15, WASL, SYNPO, TRIOBP, ACTR10, GAN, ACTN4, LGALS7, ANK2, ITM2B, RRS1, ARPC1A, TNXB, SPON1, FGL2, CAV1), and cell communication (ATP6AP2, ITGAV, LDLR, PPFIBP2, SBF1, SORBS2, ABR, NCKIPSD, NAE1, PRPF6, EPB41L2, BMP2K, APP, RAB4A, MTOR, RAP1B). Very interestingly, proteins involved in immune response processes were enriched (NRP1, PROCR, ALCAM, CD109, LBP, ST6GAL1, LRRC8A, CFB), these proteins which were not found in the D1-Frozen condition have a proliferative, immune, and anti-apoptosis profile, demonstrating a lack of these biological processes in the D1-Frozen condition.

Fig. 5figure 5

Proteomics analysis of CMT tumoroids. A Venn diagram representing specific proteins identified in the Fresh, Frozen, and FrozenCell tumoroids at date 1 and date 2 (n = 3 for each condition). B Biological processes of the proteins identified in common except in Frozen D1 tumoroids (n = 3). C Matrix correlation studies between the two dates in the tree different conditions (n = 3 for each condition). D PCA analysis of the proteomics data from the tree different tumoroid conditions (n = 3 for each condition). E Venn diagram representing specific proteins in the Fresh, Frozen, and FrozenCell tumoroids (n = 3 for each condition). F FunRich biological process distribution of the specific proteins identified in Fresh, Frozen, and FrozenCell tumoroids (n = 3 for each condition)

Likewise by a Pearson correlation analysis, hierarchical clustering of all the samples based on the correlation coefficients between them revealed higher similarity between Fresh and FrozenCell tumoroids at dates 1 and 2. Frozen tumoroids were more different, specifically at date 1 (Fig. 5C). The similarity of D1-Frozen with the other conditions was less than 87%, while all the other conditions showed more than 95% similarity. The duration of the tumoroids culture did not seem to have a big impact on their proteomic profiles. The fact that D1-Frozen tumoroids were more distinct suggests that the tumoroids should be preferentially left in culture long enough to recover after freezing, which was not observed from D1-FrozenCells.

Knowing that the time in culture did not impact too much their molecular profile, we wanted to verify whether if the culture condition impacted or not their proteome. For that, we have compared the proteomic profiles of tumoroids from three culture conditions: fresh, frozen, and tumoroids made from frozen cells. First of all, the principal component analysis (PCA) based on the LFQ values of the protein identification showed that the samples were grouped by tumor and not according to the type of culture condition (Fig. 5D). This sample grouping by PCA means that there was a high level of similarity between the biological replicates of each condition but also between the tumoroids without influence of their culture condition. Furthermore, a Venn diagram showing the number of common and unique proteins in all conditions showed that a majority of proteins were identified in all three conditions of culture (2389 proteins, representing 90% of all proteins). However, some proteins were found specifically expressed in each condition: 27 identified specifically in fresh tumoroids, 23 in frozen tumoroids, and 34 in tumoroids made from frozen Cells (Fig. 5E) (Additional File 7: Table 4). Based on the GO term enrichment analysis of the biological processes using FunRich software, we observed that these proteins, specifically expressed in each condition, were linked to different biological processes (Fig. 5F). An effect on the cellular metabolic and energy pathways was found in Frozen and FrozenCell tumoroid conditions compared to Fresh tumoroids (with: AMY1A, SDR9C7, CDA, ARG1, FKBP, NDUFB10, DDO, ADH5, and SIAE). In addition, in the Frozen and FrozenCell condition, we have identified proteins involved in apoptosis like the apoptosis facilitator BCL2L13, ATG5, and TXNRD2. In the Fresh tumoroids, a higher number of proteins linked to cell communication and to signal transduction were identified. Interestingly, some of the specific proteins identified in fresh tumoroids were involved in the immune response, such as GZMB (Granzyme B) expressed by cytotoxic T and NK cells, the cell adhesion molecule Siglec1 (Sialoadhesin) expressed by macrophages, as well as CD163 (Cluster Differentiation 163), a marker of anti-inflammatory macrophages, AMBP (alpha-1 microglobulin/bikunin) precursor of a glycoprotein synthesized by lymphocytes and CD177 (CD177 molecule), a marker of neutrophil activation.

In order to better understand the differences linked to the culture conditions, an analysis of the variation of abundance of common proteins to all conditions (2389 proteins) was carried out, using a multiple sample test ANOVA with an FDR of 0.05. A total of 489 proteins showed significantly different expression between the three groups. These specific variations were analyzed by hierarchical clustering and then illustrated by a Heatmap (Fig. 6A). Six clusters of proteins were identified: one cluster representing the specific underexpressed proteins and one representing the specific overexpressed proteins for each condition (Additional File 8: Table 5). Based on over- and under-expressed proteins, fresh tumoroids and tumoroids made from frozen cells showed more similarities compared to frozen tumoroids, as observed before. In order to characterize the impact of these proteins, the GO terms analysis of each cluster was carried out with Cytoscape and ClueGO software, allowing to generate the networks connecting the proteins overexpressed (in red) and underexpressed (in green) to their biological process.

Fig. 6figure 6

Proteomics analysis of Fresh, Frozen, and FrozenCells tumoroids. A Hierarchical clustering of the most variable proteins between the 3 conditions (n = 3 for each condition, ANOVA with permutation-based FDR < 0.05). Network of proteins overexpressed (red) or underexpressed (green) in Fresh (B), FrozenCell (C), and Frozen (D) tumoroids and their associated GO terms. The networks were enriched through addition of STRING network to the identified proteins using ClueGO application on Cytoscape

The results showed that in the fresh condition (clusters 3–4) (Fig. 6A, B), the cellular respiration and the amino acid metabolism were underexpressed compare to frozen tumoroids and FrozenCell tumoroids, which could be explained by the cryopreservation. In another hand, the assembly of the cell-substrate junction and the RNA translation by RNA polymerase appeared to be overexpressed in the Fresh condition compared to the other conditions, and the biological process of cell adhesion which can be explained by a conservation or a faster reformation of the extracellular matrix and the cell compaction. For the FrozenCell condition (clusters 1–2) (Fig. 6C), there is a higher abundance of proteins linked to chromatin remodeling and cellular metabolism as we observed before (UBA52, PSMC1, PSMD5, RAB7A, HSPA9, HSPA5, HSPA8, DDB1, MDH2, SLC25A12). Again, a high metabolic activity can be a consequence of freezing. In addition, the chromatic remodeling involved in the cell division cycle can be linked to a process of multiplication and recovery from freezing that seems important in this condition.

Finally, in frozen tumoroids (clusters 5–6) (Fig. 6D), many proteins related to protein translation, extracellular matrix assembly, vesicle-mediated transport, and RNA nuclear export were found overexpressed. The biological process of vesicle-mediated transport can be subdivided on transport of extracellular vesicles, vesicle budding from membrane, vesicle targeting, vesicle coating, and COPPII-coated vesicles cargo with the presence of distinct protein such as ARCN1, AP2A1, DYNC1H1, AP2B1, ANXA7, and SEC13. In addition, the biological process of protein translation, protein-RNA nuclear export, and telomerase RNA localization are represented by proteins like eukaryotic translation initiation factor (4A-III, 3 subunit A, 3 subunit L, 3 subunit B, 3 subunit E, among others) and 40S, 60S ribosomal protein (RPL10, RPL13A, RPL14, RPL15, RPS11, RPS13, RPS18, RPS28, RPS3, RPS8, RPSA). These analysis highlight a dysfunction in the translation pathways that we know contribute to cancer progression, for example, in the deregulation of ncRNAs that leads to aberrant protein translation in cancers [26], and which seems more active in frozen tumoroids.

On the contrary, the underexpressed proteins are related to the metabolism of amino acids or nucleotides, negative regulation of cytokines, immune effector process, and the organization of the cytoskeleton, cell adhesion, and death. Different metabolic pathways were touched, such as dicarboxylic acid metabolic process, purine ribonucleotide biosynthetic process, pyruvate metabolic process, and generation of precursor metabolites and energy. Regarding the organization of the cytoskeleton and cell adhesion, different isoforms of laminin, collagen, catenin and Coronin-1B were found to be underexpressed in this condition. Apoptosis and cell death proteins (CYP1B1, HSPA1, ARL6IP5, TRAP1, among others) were also found underexpressed. In addition, we have identified underexpressed proteins linked to a regulation of cytokines and to the immune effector process, in which we find proteins: CD44, thrombospondin-1, SAMHD1, TINAGL1, GAA, LGALS9 (Galectin), among others.

The Cytoscape and ClueGO analysis shows that Fresh and FrozenCell conditions have limited amount of underexpressed and overexpressed proteins, while frozen condition shows three time more proteins with significant variation. On the other side, even if the FrozenCell is closer to Fresh Tumoroids than Frozen condition, the degree of similarity stays high.

The proteome of canine mammary tumoroids is very similar to the tumor they originate and therefore represent a faithful breast cancer model

We next wanted to determine whether the three different types of tumoroids were similar to the tumor of origin, since the tumoroids will be used as a model of breast cancer.

For this, a Venn diagram (Fig. 7A) was made and showed the number of common and unique proteins in all conditions (Additional File 9: Table 6). It can be observed that a majority of proteins were identified in all three conditions of culture (2138 proteins, representing 74% of all proteins). However, there were specific proteins for each condition: 4 identified specifically in fresh tumoroids, 15 in frozen tumoroids, 15 in tumoroids made from frozen cells, and 154 specific proteins that were found specifically in the tumor. These 154 proteins are involved in different biological processes such as cell growth and/or maintenance, cell communication, signal transduction, and immune response (Fig. 7B). Interestingly, an immunological profile can be observed in the tumor compared to the tumoroids. We found several proteins involved in the complement signaling pathway (complement factor I, complement component 4 binding protein, complement component 5, complement component 7, and complement component 8) involved in immunological response and phagocytosis and found overexpressed in different types of cancer, such as breast cancer [27]. In addition, proteins such as CD93, CD34, CLEC4G, haptoglobin, and joining chain of multimeric IgA and IgM have been identified and are all implicated in immune response. The AOC3 protein was also identified and has been recently described to play a role in the reduction of immune cell recruitment and impacting the promotion and progression of lung cancer [28].

Fig. 7figure 7

Proteomics analysis comparing the primary tumor to their derived tumoroids. A Venn diagram representing specific proteins in tumor of origin, Fresh, Frozen, and FrozenCell tumoroids (n = 3 for each condition). B Biological processes of the specific proteins identified in primary tumors using FunRich and ClueGO (n = 3 for each condition). C Hierarchical clustering of the most variable proteins between the tumor of origin and the 3 tumoroid conditions (n = 3 for each condition, ANOVA with permutation-based FDR < 0.05). Biological processes distribution of underexpressed (D) and overexpressed (E) proteins in tumors compared to tumoroids using FunRich and ClueGO (n = 3 for each condition)

To better understand the differences between tumor and tumoroids, an analysis of the variation of abundance of common proteins was carried out. A total of 512 proteins showed significantly different expression between the four groups. The HeatMap of the specific variations of abundance (Fig. 7C) shows few variations between the three types of tumoroids. Frozen tumoroids were more different compared to the two other culture conditions confirming the previous results. Interestingly, a small cluster of overexpressed proteins was observed in Tumor, Fresh, and FrozenCell tumoroids, while underexpressed in Frozen tumoroids. This result shows again that the conditions more similar to the tumor of origin are the Fresh and FrozenCell tumoroids.

We next focused on the two clusters that showed the significant differences between the tumor of origin and the tumoroids (Additional File 10: Table 7). The HeatMap shows two clusters of proteins over- or underexpressed in tumor compared to the different tumoroids. Functional annotation and characterization of these two clusters of proteins were performed using the FunRich software. The results showed that the biological processes underexpressed in the tumor compared to tumoroids are different processes involved in cellular metabolic and energy process such as cellular respiration (with NDUFV1, GPD2, CS, NDUFS1, SDHA) and carbohydrate derivative metabolic process (GPD2, ENO2, CAD, UGGT1, HK1, PFKP, FASN, AMPD2, MOGS), transport with two principal pathway detected: the protein import into nucleus (with presence of interacting protein CSE1L, IPO7, KPNB1, NUP93, TNPO1), and endoplasmic reticulum to Golgi vesicle-mediated transport (with the complex of protein like COPA, COPB1, COPB2, COPG2, SEC13, SEC23A, USO1), apoptosis, and signal transduction (Fig. 7D).

On the contrary, we can observe proteins overexpressed in the tumor compared to tumoroids involved in cell growth and maintenance, actin filament-based process (through ACTG1, ACTN1, ARPC2, EZR, FLNA, GSN, MYH11, MYH14, MYH9, MYL6, MYO1C, TLN1, TPM3, VIM), cell communication, signal transduction, and immune response (Fig. 7E).

These results confirm that in tumoroids we find a higher metabolic activity, especially in frozen tumoroids as demonstrated above. In addition, we can observe apoptotic proteins such as BCL2-associated athanogene 6, heat shock 60 kDa protein 1, PH domain and leucine rich repeat protein phosphatase 2, cell cycle and apoptosis regulator 2, underexpressed in the tumor and consequently overexpressed in the tumoroids. Also, in the tumor, we can find an important proliferative profile demonstrated by the overexpression of proteins involved in cell communication and linked to the organization of the cytoskeleton and cell growth (actin beta-like 2, collagen type VI, tubulin, beta 4B class IVb, lamin A/C, actin alpha 2, actin-related protein, among others). Finally, we can find an immune profile more present in the original tumor.

Canine mammary tumoroids can be used to test human drugs and cryopreservation of tumoroids does not impact drug response

In order to evaluate canine mammary tumoroids as useful tools for translational in vitro drug screening studies, we performed cell viability assays in presence of a chemotherapeutic agent used in human medicine, paclitaxel. Tumoroids were treated with paclitaxel for 7 days before cell viability was measured. Representative images of tumoroids derived from TM-05 tumor are shown in Fig. 8A demonstrating drug sensitivity depending on the drug concentration. Using 6 concentrations of paclitaxel, we generated dose–response curves (Fig. 8B and Additional File 4: Table 2). First, we could demonstrate that tumoroids derived from fresh canine mammary tumors responded well to paclitaxel with an IC50 ranging from 0.1 to 1 μM. 0.1 μM paclitaxel was needed to kill 50% of tumor cells for TM-04 and TM-06, while around 1 μM was needed for TM-05 demonstrating higher resistance to paclitaxel.

Fig. 8figure 8

Drug response of canine and human tumoroids to paclitaxel. A Representative bright field images showing the morphology of the three types of tumoroids after 7 days treatment with paclitaxel at different concentrations. Scale bar (100 μm) is indicated. B Quantification of the tumoroids viability following paclitaxel treatment. Tumoroids were generated from three different canine mammary tumors (TM-04, TM-05, and TM-06), and drug response was compared between fresh, frozen, and frozen cell tumoroids. Data are the means ± SD (n = 3 for each condition). C Tumoroids were generated from human and canine luminal and triple-negative mammary tumors, and drug response was compared between human and canine tumoroids. Different concentrations of drug were used and compared to non-treated tumoroids. Data are the means ± SD (n = 3 for each condition)

We next compared the paclitaxel response between fresh tumoroids, frozen tumoroids, and tumoroids made from frozen cells in order to verify whether cryopreserved tumoroids could represent faithful models for drug testing. Killed curves from these three culture conditions were similar for the three tumors tested (Fig. 8B). As we have observed before with the proteomic analysis, fresh tumoroids and tumoroids derived from frozen cells were the most similar in term of paclitaxel responses. Nonetheless, tumoroids derived from frozen cells appear to become slightly more resistant at higher concentrations of paclitaxel (Fig. 8A, B) compared to fresh tumoroids. Indeed, an increase of viability of tumoroids derived from frozen cells can be observed for each tumor at a concentration of 100 μM. A 50% viability of tumoroids was measured in this condition, while only 35% of cells were viable in fresh tumoroids (Fig. 8B). Finally, the response to paclitaxel of frozen tumoroids appears to be slightly different compared to the two other conditions, even if not significant. For TM-04 and TM-06, fresh tumoroids and tumoroids made from frozen cells seem to be more sensitive whatever the concentration of paclitaxel used compared to frozen tumoroids. For TM-05, the three curves are more similar. A 50% viability decrease of fresh tumoroids and tumoroids made from frozen cells is observed between 0.1 and 1 μM of drug, while between 1 and 10 μM of paclitaxel are needed to kill 50% of tumor cells in the frozen tumoroids condition (Fig. 8B).

In the end, we have shown that paclitaxel response between luminal subtype CMT tumoroids and human breast tumoroids was similar with 0.1 μM of paclitaxel needed to kill 50% of tumor cells. In contrast between triple-negative subtype CMT tumors and human breast tumors, although they show a similar response trend, they respond differently to the range of paclitaxel concentrations. CMT tumors (TM-05 and TM-06) show an amount between 0.1 and 1 μM paclitaxel needed to kill 50% of tumor cells, whereas the human tumor shows around 10 μM paclitaxel (Fig. 8C and Additional File 4: Table 2). Further study on triple-negative tumors will be necessary to make more detailed conclusions. Unfortunately, triple-negative tumors are the most complicated in terms of tumoroids growth and the collection of a cohort of only human triple-negative tumors would be challenging in terms of time [29].

In conclusion, canine tumoroids respond well to a chemotherapeutic agent used in human medicine. The way the tumoroids are preserved has little impact on drug response. It seems, however, that tumoroids made from frozen cells best mimic the drug response of fresh tumoroids.

留言 (0)

沒有登入
gif