Abstract

   The circadian clock coordinates the daily rhythmicity of biological
   processes, and its dysregulation is associated with various human
   diseases. Despite the direct targeting of rhythmic genes by many
   prevalent and World Health Organization (WHO) essential drugs,
   traditional approaches can't satisfy the need of explore
   multi‐timepoint drug administration strategies across a wide range of
   drugs. Here, droplet‐engineered primary liver organoids (DPLOs) are
   generated with rhythmic characteristics in 4 days, and developed
   Chronotoxici‐plate as an in vitro high‐throughput automated rhythmic
   tool for chronotherapy assessment within 7 days. Cryptochrome 1 (Cry1)
   is identified as a rhythmic marker in DPLOs, providing insights for
   rapid assessment of organoid rhythmicity. Using oxaliplatin as a
   representative drug, time‐dependent variations are demonstrated in
   toxicity on the Chronotoxici‐plate, highlighting the importance of
   considering time‐dependent effects. Additionally, the role of
   chronobiology is underscored in primary organoid modeling. This study
   may provide tools for both precision chronotherapy and chronotoxicity
   in drug development by optimizing administration timing.

   Keywords: chronotoxicity, Cry1, droplet‐engineered liver organoid,
   oxaliplatin
     __________________________________________________________________

   A Chronotoxici‐plate containing droplet‐engineered rhythmic liver
   organoids has been developed for chronotherapy studies and
   chronotoxicity evaluation.

   graphic file with name ADVS-11-2305925-g002.jpg

1. Introduction

   The circadian clock orchestrates daily rhythms in human biochemical,
   physiological, and behavioral functions. Its disruption has been
   implicated in various human diseases, potentially contributing to
   metabolic syndrome. At the molecular level, the clock regulates
   rhythmicity in gene expression, cell division, and DNA repair, hinting
   at potential connections between circadian rhythms and both the
   pathogenesis and therapeutic interventions of cancer.^[ [44]^1 ^] As
   research progresses and understanding deepens, circadian rhythm
   research has transformed from a mere biological model into a crucial
   component of future clinical and pharmaceutical research. Emergencies
   that manifest at specific times of day, such as nocturnal asthma^[
   [45]^2 ^] and glucose peak at dawn,^[ [46]^3 ^] are among the short
   list indications of circadian pathologies. Also, some studies
   demonstrated a higher incidence of heart attacks during morning hours
   and reduced risk associated with afternoon cardiovascular surgeries
   compared to those performed in the morning.^[ [47]^4 ^] These findings
   may be attributed to time‐dependent vulnerabilities linked to
   oscillations in blood pressure and heart rate,^[ [48]^5 ^]
   thromboembolic risk,^[ [49]^6 ^] arrhythmogenicity,^[ [50]^7 ^] and
   inflammation.^[ [51]^8 ^] While the detrimental effects of circadian
   rhythm disruption are well‐documented, the potential health benefits of
   harnessing or rectifying biological timing, including chronotherapy,
   remains underexplored.^[ [52]^9 ^] Balsalobre et al. successfully
   developed a circadian model that replicates the oscillatory expression
   of genes observed in vivo, by treating two‐dimensional (2D) cell line
   (rat‐1 fibroblasts or H35 hepatoma cells) with high concentrations of
   serum.^[ [53]^10 ^] The application of serum shock (SS) stimulated the
   expression of specific genes to mimic homeostasis in vivo.

   Chronotherapy involves drug delivery at the appropriate phase of
   circadian rhythm to achieve optimal efficacy or minimal toxicity. And
   the rationale for this strategy is the performance of mechanism of
   action (MoA),^[ [54]^11 ^] pharmacokinetics and metabolism,^[ [55]^12
   ^] as well as toxicity varying with time.^[ [56]^13 ^] The failure to
   consider chronotherapy can result in the overlooked potential of
   valuable drugs. A notable example is the initial high toxicity observed
   in clinical trials of oxaliplatin, where the benefits of chronotherapy
   were not considered.^[ [57]^14 ^] Synchronizing drug administration
   with the circadian rhythm, specifically at 16:00, has proven effective
   in enhancing therapeutic outcomes while mitigating drug toxicity. A
   significant portion of best‐selling and World Health Organization (WHO)
   essential drugs, including Rituxan, target the products of rhythmic
   genes.^[ [58]^15 ^] These drugs, often with short half‐lives, could
   benefit from timed dosage. Further, the need is pressing for
   researchers to incorporate circadian time in pre‐clinical and clinical
   phases of new drug development, as well as in evaluation of novel
   treatments. Applying chronotherapy to previously unsuccessful drugs,
   following the oxaliplatin example to align the administration time,
   could also yield promising results in the balance of therapeutic
   efficacy and drug toxicity. However, traditional approaches using model
   animals are hampered by complex procedures, high costs, and low
   throughput, hindering large‐scale drug screening and multi‐timepoint
   drug administration strategies.

   Organoids are in vitro self‐assembling, three‐dimensional (3D) cellular
   structures that retain key characteristics of the respective tissues.^[
   [59]^16 ^] They serve as a powerful tool for biomarker research, drug
   screening, and accurate prediction of therapeutic efficacy. When
   combined with high‐throughput assays, organoids can reduce the time and
   cost of experiments. Microfluidic droplets provide monodisperse and
   tailorable templating volumes for cellular compartments, where primary
   or pluripotent cells grow and develop into reproducible organoids,
   termed droplet‐engineered organoids (DEOs).^[ [60]^17 ^] This
   technology provides a high‐throughput, standardized and cost‐effective
   choice to conduct large‐scale drug screening with multi‐timepoint drug
   administration strategies.

   As an emerging tool, multiple attempts have been made on organoid‐based
   circadian rhythm research.^[ [61]^18 , [62]^19 , [63]^20 , [64]^21 ,
   [65]^22 ^] Stokes et al. assessed the circadian transcription rhythms
   in intestinal organoids and tested Bmal1‐dependent proliferation and
   self‐renewal.^[ [66]^21 ^] Rosselot et al. found circadian
   phase‐dependent necrotic cell death responses to Clostridium difficile
   toxin B (TcdB) in intestinal organoids.^[ [67]^19 ^] Notably, there is
   a significant gap in the literature regarding the establishment of
   fully in vitro rhythmic organoids tailored for precision cancer
   chronotherapy. Our work stands out as a promising avenue for
   introducing a novel therapeutic paradigm, moving beyond its role as a
   conventional drug sensitivity and toxicity assessment tool.

   Here, leveraging a high‐throughput DEO fabrication tool, we utilized
   microfluidics and 3D printing technologies to construct a
   “Chronotoxici‐plate”: a 96‐well plate filled with rhythmically
   synchronized droplet‐engineered primary liver organoids (DPLOs). The
   system enabled DPLOs establishment within 4 days, identified
   Cryptochrome 1 (Cry1) as a potential companion diagnostic marker, and
   realized assessment of oxaliplatin toxicity for chronotherapy within
   one week. Further, it demonstrated the effect of rhythm synchronization
   in liver organoid growth, suggesting it as a new dimensional
   intervention to improve primary organoid culture.

2. Results and Discussion

2.1. Establishment of Chronotoxici‐Plate Derived from Mouse Liver

   Circadian biology is increasingly recognized as a crucial factor in
   enhancing drug efficacy and reducing toxicity.^[ [68]^23 ^] However,
   isolated cells exhibit poor synchronization of circadian cycles^[
   [69]^22 , [70]^24 ^] and the large‐scale drug screening with
   multi‐timepoint drug administration is not satisfied. Therefore, there
   is currently a lack of rapid and efficient tools with circadian
   rhythmic features for drug screening. DEOs were formulated in a
   high‐throughput manner in a customized microfluidics setup, i.e.,
   cascade‐tubing microfluidics (CTM). The organoids were individually
   bioprinted into a 96‐well plate and synchronized in rhythm on the 4^th
   day using SS, establishing the circadian drug screening organoids,
   i.e., DPLOs. We designated this organoid‐housed plate as
   “Chronotoxici‐plate” and employed it for subsequent drug screening
   purposes. Cry1 was screened as a marker of chronotherapy. Using the
   screened gene as the drug safety assessment companion, we performed
   in‐depth analysis of rhythmic liver toxicity of oxaliplatin on the
   Chronotoxici‐plate (Figure [71] 1 ).

Figure 1.

   Figure 1
   [72]Open in a new tab

   Schematic illustration of Chronotoci‐plate construction. The
   preservation of intrinsic tissue‐level circadian rhythms poses a
   challenge in isolated cells and primary organoids derived from these
   cells. To overcome this limitation, we designed the Chronotoxici‐plate:
   a formatted tool aimed at restoring rhythmicity for drug safety
   assessment. After four days of culture, the organoids within the plate
   are subjected to SS, then returned to their original environment,
   effectively synchronizing their cellular rhythms. The
   Chronotoxici‐plate represents a high‐throughput kit for the exploration
   of chronotoxicity, making it easier to study of drug responses in a
   time‐dependent manner.

   To provide a fast and efficient tool, Chronotoxici‐plate, for circadian
   research and drug safety assessment, we reconstructed hepatobiliary
   microtissues in vitro^[ [73]^25 , [74]^26 , [75]^27 , [76]^28 ^] using
   primary cells isolated from healthy mouse liver^[ [77]^29 ^] to
   formulate DPLOs and patterned them in Chronotoxici‐plate (Figure [78]
   2a). Each well contained only one DPLO. We utilized SS treatment at Day
   4 to synchronize organoid cell phases, and the DPLOs were featured with
   higher cell density over time (Figure [79]2b). The results of the
   live/dead cell viability test indicated higher rates of cell survival
   at Day 8 for both the SS and control (C) groups, as shown by the mean
   green‐to‐red intensity ratios (MIR) (SS group, MIR[ss] = 62.13; control
   group, MIR[c] = 8.89, Figure [80]2c). The Adenosine triphosphate (ATP)
   luminescence assay revealed a gradual recovery of cell activity
   following an initial decline, at Days 1–4 and Days 5–6 respectively
   (Figure [81]S1, Supporting Information). During the culture process,
   cell metabolic levels and viability rates first decreased from Day 1 to
   Day 4 possibly due to the harmful effect of primary cell extraction and
   adaption to the in‐vitro environment,^[ [82]^30 , [83]^31 ^] followed
   by noticeable cell proliferation from Day 4 to Day 8. This is in line
   with the scenario observed in classical primary liver organoids,^[
   [84]^32 ^] or termed non‐engineered organoids^[ [85]^17 ^] (NEO),
   though the decline in ATP activity was greater and exhibited slower and
   considerably compromised recovery (Figure [86]S2, Supporting
   Information).

Figure 2.

   Figure 2
   [87]Open in a new tab

   Cultivating DPLOs within the Chronotoxici‐plate. a) High‐content
   analysis of each well in Chronotoxici‐plate at Day 1. Every well in the
   96‐plate contained a single DPLO. Scale bar = 500 µm. b) Observation of
   organoid culture process under bright‐field microscopy. DPLOs were
   subjected to SS or kept as control (C) after four days of culture,
   followed by continued cultivation until the eighth day (D8). Scale bar
   = 100 µm. c) Confocal microscopy image depicting the viability of DPLOs
   through whole mount staining using Calcein‐AM/PI live/dead staining.
   MIR: the mean green‐to‐red intensity ratio. Scale bar = 100 µm. d)
   Schematic representation and histological staining images of single
   DPLO and interconnected multi‐DPLOs at Day 4 after establishment. Scale
   bar = 80 µm. e) Oil Red O staining images of single DPLO and
   interconnected multi‐DPLOs at Day 3 and Day 7 with or without SS
   treatment. Scale bar = 50 µm. f, g, and h) Confocal image of a
   cryo‐sectioned DPLO co‐stained for HNF4a (green), CK19 (red), CD68
   (green), albumin (green), CK18 (red), and DAPI (blue). Scale bar =
   50 µm. i, j, and k) qRT‐PCR analysis of the expression levels of
   liver‐related growth factors in DPLOs and AML12 cell lines. i,k) *
   indicates p < 0.05.

   Subsequently, we validated the structure and functionality of DPLOs.
   Hematoxylin‐eosin (HE) staining of the organoids after four‐day
   culturing revealed the presence of tubular structures (Figure [88]2d,
   black arrow) and densely packed cell clusters with organization
   (Figure [89]2d). Oil Red O staining revealed a noticeably higher cell
   density at Day 7 compared to Day 3 and a superior structural
   arrangement with no obvious de novo lipogenesis (Figure [90]2e).

   Immunofluorescence (IF) analysis was conducted on the organoids at Day
   4 demonstrated the presence of HNF4α+ hepatocyte clusters distributed
   within the organoids, co‐existing with a portion of CK19+
   cholangiocytes (Figure [91]2f, Figure [92]S3, Supporting Information).
   Additionally, Kupffer cells, characterized by CD68 expression, were
   also observed in the organoids (Figure [93]2g). Cytokeratin‐18 (CK18),
   an abundant intermediate filament protein in hepatocytes and
   cholangiocytes, was expressed in the tubular structures. Additionally,
   albumin, which reflects hepatocyte functionality and protein synthesis,
   exhibited widely expression in the organoids (Figure [94]2h; Figure
   [95]S3, Supporting Information). The co‐expression of biliary and
   hepatocyte markers (CK19 and HNF4a), one of the characteristics of
   periportal hepatocytes of regenerating livers,^[ [96]^33 ^] has been
   validated as markers for functional liver organoids.^[ [97]^29 ,
   [98]^34 ^] These findings demonstrate the success of structural
   development and functional maturation of DPLOs.

   Organoid cultures have the potential to partially recapitulate tissue
   phenotypes and reproduce their paracrine activity compared to 2D cell
   cultures.^[ [99]^35 ^] The viability and preservation of in vivo‐like
   phenotypes in liver primary cells isolated from tissue have been shown
   to be challenging when cultured in a 2D environment (Figure [100]S4,
   Supporting Information). This difficulty underscores the inherent
   limitations of traditional 2D culture systems in replicating the
   complex in vivo conditions essential for maintaining the functional
   characteristics of hepatic cells.^[ [101]^36 , [102]^37 ^] We also
   examined the expression of important paracrine factors involved in
   liver growth and development in DPLOs where hepatocytes isolated from
   the normal liver of a 3‐month‐old mouse, and AML12 (alpha mouse liver
   12) cells of 2D culture after SS treatment. Our results revealed
   significantly higher expression levels of hepatocyte growth factor
   (Hgf),^[ [103]^38 , [104]^39 , [105]^40 ^] transforming growth factor
   beta (Tgf),^[ [106]^41 , [107]^42 ^] and fibroblast growth factor‐7
   (Fgf‐7)^[ [108]^43 , [109]^44 ^] in DPLOs compared to the AML12 cells
   (Figure [110]2j–l) This highlights the advantage of
   droplet‐encapsulation‐derived organoids in reproducing tissue paracrine
   activity and their potential to simulate organ‐like features in vitro.

2.2. Rhythmicity Gaining in DPLOs in Chronotoxici‐Plate

   To validate the synchronization of cellular rhythms in DPLOs of
   Chronotoxici‐plate, we sampled organoids every 6 hours starting from
   the end of SS treatment.^[ [111]^45 , [112]^46 , [113]^47 ^] The
   organoids were labeled as SS‐0 h, SS‐6 h, SS‐12 h, SS‐18 h, and
   SS‐24 h. The control group organoids were also sampled from the same
   batch of DPLOs at 0 h but without SS treatment. RNA sequencing analysis
   revealed that gene expression levels in samples from different time
   points after SS were similar to the control group, i.e., the
   non‐SS‐treated group (Figure [114]S5,S6, Supporting Information). It
   suggested that SS did not induce widespread sustainable upregulation or
   downregulation of genes. A comparative analysis was conducted between
   the SS‐treated groups at various time points and the control group
   (SS‐0 h/C, SS‐6 h/C, SS‐12 h/C, SS‐18 h/C, SS‐24 h/C). The analysis
   revealed 1145 common differentially expressed genes (DEGs). While DEGs
   uniquely expressed in a specific SS group were proposed as genes
   subject to circadian rhythm, as shown in Figure [115] 3a. To evaluate
   the rhythmicity, the sample similarity of well‐acknowledged 12
   clock‐controlled genes^[ [116]^48 ^] (CCGs) was assessed in the five
   comparison groups (SS‐0 h/C, SS‐6 h/C, SS‐12 h/C, SS‐18 h/C, SS‐24 h/C)
   through gene expression profiling. The results revealed a oscillation
   pattern of initial decrease, subsequent increase, and final decrease
   from 0 to 24 hours (Figure [117]3b,c). A similar oscillation pattern
   was observed in the CCGs and the whole transcriptomic profiling from 6
   to 24 hours (Figure [118]3d,e). Collectively, it was revealed that
   rhythmic gene expression was recapitulated in DPLOs of
   Chronotoxici‐plate.

Figure 3.

   Figure 3
   [119]Open in a new tab

   Synchronization of DPLOs via SS treatment. a) Venn diagram illustration
   of the overlap of differentially expressed genes (DEGs) among different
   groups from RNA sequencing. b) Sample Similarity Heatmap of
   clock‐controlled genes (CCGs) determined by RNA sequencing comparing
   samples from different time points after SS (SS‐0 h, SS‐6 h, SS‐12 h,
   SS‐18 h, SS‐24 h) with the control group. c) Temporal dynamics of CCGs
   expression from RNA sequencing in response to SS treatment compared to
   the control group. d) Sample similarity heatmap of gene expression
   profiles determined by mRNA sequencing across samples from various time
   points following SS compared to the control group. e) Temporal dynamics
   of gene expression profiles in response to SS treatment compared to the
   control group. f) Gene Ontology enrichment analysis of DEG sets by mRNA
   sequencing comparing SS DPLOs to control DPLOs. The analysis revealed
   significant enrichment of rhythmic‐related pathways. A threshold of
   Q‐value (corrected P‐value) ≤ 0.05 was applied, defining GO terms that
   exhibited significant enrichment among the candidate genes. The Rich
   ratio represents the ratio between the number of DEGs in a specific
   metabolic pathway and the total number of genes annotated to that
   pathway. Higher values indicate a greater degree of enrichment within
   the pathway. g and h) Gene Ontology enrichment analysis of DEG sets by
   mRNA sequencing comparing SS DPLOs to control DPLOs or the initial time
   point after SS. The analysis revealed significant enrichment of liver
   development and cell proliferation. A threshold of Q‐value (corrected
   P‐value) ≤ 0.05 was applied, defining GO terms that exhibited
   significant enrichment among the candidate genes.

2.3. Liver Development in Rhythmic DPLOs

   To further investigate pathway alterations in synchronized DPLOs, gene
   ontology (GO) enrichment analysis was performed between different time
   points of the SS groups and the control group. Under the criteria of Q
   value < 0.05, enriched pathways related to circadian rhythms were
   identified, including negative regulation of circadian rhythm
   (GO:00 42754), regulation of circadian rhythm (GO:00 42752), circadian
   regulation of gene expression (GO:00 32922), and rhythmic process
   (GO:00 48511) (Figure [120]3f). Additionally, enrichment was observed
   in pathways associated with regulation of cell proliferation
   (GO:00 42127), positive regulation of cell migration (GO:00 30335), as
   well as positive regulation of hepatocyte proliferation (GO:2 000 347),
   hepatocyte growth factor receptor signaling pathway (GO:00 48012), and
   liver development (GO:0 001889), indicating enhancement of cell
   proliferation, migration, and potential liver development in DPLOs
   after synchronization (Figure [121]3g,h). Besides, the survival of
   DPLOs after synchronization surpassed the control group by Day 6 and
   showed a significantly higher cell viability rate at Day 8
   (Figure [122]2b,c). It may suggest a positive effect of rhythmic
   synchronization on DPLOs in terms of liver development and survival.

2.4. Cry1 is a Rhythmic Marker in DPLOs

   To simplify and expedite the assessment of rhythmicity in DPLOs of the
   Chronotoxici‐plate, we identified the rhythmic marker in DPLOs. Based
   on the rhythmic expression of 12 CCGs observed in DPLOs
   (Figure [123]3b,c), protein‐protein interaction networks (PPI) analysis
   revealed close interactions among all CCGs, particularly the close
   interactions between Cry1 and other genes (Figure [124] 4a). Analysis
   of CCG expression at different time points after synchronization showed
   Cry1 expression oscillated around the expression values in the control
   group and exhibited the best fit to the periodic function (p‐value =
   0.06158, Figure [125]4b,c).

Figure 4.

   Figure 4
   [126]Open in a new tab

   Identification of Cry1 as a potential companion diagnostic indicator.
   a) Network representation of CCGs interactions using mRNA sequencing
   results. The color of the nodes indicates the number of connections,
   while the color of the edges reflects the degree of interaction. b)
   Heatmap of CCGs expression determined by mRNA sequencing comparing
   samples from different time points after SS with the control group. c)
   Fit of CCGs expression profiles at different time points after SS to a
   cosine function, and the p‐value was calculated.^[ [127]^19 , [128]^22
   , [129]^70 , [130]^71 , [131]^72 , [132]^73 ^] d) Temporal dynamics of
   the ratio between Cry1 and housekeeping gene Actin expression levels
   following SS treatment in western blot analysis. Each dot stands for
   the relative protein expression of Cry1, which were fitted to the
   cosine function, and the p‐value was calculated. e,f,h, and i) Fit of
   Cry1 expression by qRT‐PCR analysis at different time points in DPLOs
   or AML12 cell lines with or without SS to a cosine function, and the
   p‐value was calculated. Each dot stands for the relative mRNA
   expression of Cry1. g) Comparison of the fit between the expression of
   Cry1 gene, with or without SS, and a sine function. ∗∗ indicates p <
   0.01.

   Subsequently, we performed multiple replicates at the protein and mRNA
   levels to investigate the effects of synchronization on DPLOs of the SS
   group and the control group. At the protein expression level, Cry1
   exhibited robust periodicity in expression ratio with the housekeeping
   gene Actin (p‐value = 0.03027, Figure [133]4d). At the mRNA expression
   level, following synchronization, DPLOs showed a lower p‐value for Cry1
   gene expression fitting the sine function (p‐value = 0.09105,
   Figure [134]4e) compared to the control group (p‐value = 0.79646,
   Figure [135]4f). Multiple repeated experiments consistently
   demonstrated a significant effect of SS in reducing the p‐value for
   Cry1 expression and sine function fitting in DPLOs (Figure [136]4g).
   This demonstrated Cry1 emerges as a representative rhythmic marker in
   DPLOs. In contrast, SS in AML12 cells cultured in 2D showed a larger
   p‐value (p‐value = 0.17076) for Cry1 expression fitting the sine
   function (Figure [137]4h,i), indicating a less pronounced effect of
   SS‐induced synchronization on 2D cells than on DPLO cells. This
   discrepancy may be attributed to the activation of numerous
   rhythmic‐related pathways (Figure [138]3f–h) and paracrine expression
   (Figure [139]2i–k) of 3D culture following SS treatment. These results
   underscore the importance of droplet‐encapsulation in generating
   rhythmically synchronized organoids.

2.5. Chronotoxici‐Plate is a Circadian Tool for Drug Toxicity Assessment

   Oxaliplatin, as a typical example of chronotoxicity, was marred by
   excessive toxicity in initial clinical trials, but subsequent
   successful chronotherapy investigations have propelled oxaliplatin to
   become one of the primary therapeutic options for cancer on a global
   scale.^[ [140]^49 ^] To evaluate the effect of rhythmic DPLOs in
   Chronotoxici‐plate on drug safety assessment, we treated
   Chronotoxici‐plate with oxaliplatin for 36 hours at different time
   points following synchronization. Glutamic‐oxaloacetic transaminase
   (GOT) and glutamic‐pyruvic transaminase (GPT) isoenzymes are known as
   crucial markers for assessing liver function.^[ [141]^50 , [142]^51 ^]
   Therefore, we measured the mRNA expression levels of cytosolic
   glutamic‐oxaloacetic transaminase‐1 (Got1), mitochondrial
   glutamic‐oxaloacetic transaminase‐2 (Got2), glutamic‐pyruvic
   transaminase (Gpt), along with the companion of the representative
   circadian gene Cry1. We assessed the expression of the rhythmic marker
   Cry1 at the time points of drug delivery (6 h, 12 h, 18 h), which
   showed the pattern of an initial decrease followed by an increase
   (Figure [143] 5a). With the oscillated expression of rhythmic marker
   Cry1 in varied time‐point after synchronization, the results revealed a
   slightly lower expression of GOT1 in SS‐6h‐O (the group treated with
   oxaliplatin for 36 hours at 6 hours after SS synchronization) followed
   by a higher expression in SS‐12h‐O and SS‐18h‐O, compared to the
   non‐SS‐treated group. GOT2 showed a slightly higher expression
   initially and then a slightly lower expression compared to the
   non‐SS‐treated group, with only minor differences observed. The
   expression level of GPT initially exceeded and showed a significant
   decrease in 12 h, followed by a slight increase above 18 h
   (Figure [144]5a). Proteomic analysis of DPLOs (Figure [145]S7,
   Supporting Information) after oxaliplatin treatment beginning at 6, 12,
   and 18 hours after SS synchronization displayed similar trends for GOT1
   and GPT as observed in the mRNA analysis (Figure [146]5b), as well as
   enzyme activity assays (Figure [147]5c). Furthermore, significant
   differences in enzyme activity indicative of liver damage were observed
   at SS‐6 h and SS‐12 h, which highlighted distinct impacts of drug
   administration at different time points after SS treatment on DPLOs
   (Figure [148]5c). In comparison to the non‐SS treated group, the
   SS‐12 h group exhibited notably higher GOT and GPT activities,
   suggesting a potential heightened sensitivity to the drug.

Figure 5.

   Figure 5
   [149]Open in a new tab

   Drug safety assessment in Chronotoxici‐plate. a) qRT‐PCR analysis of
   liver injury‐related genes Got1, Got2, and Gpt, as well as the
   circadian rhythm gene Cry1 expression level, in DPLOs following SS
   treatment or keeping as a control at different time points with a
   period of 36‐hour drug exposure. b) iTRAQ analysis of Got1 and Gpt in
   DPLOs following SS treatment at different time points with a 36‐hour
   drug exposure period. c) Differences in GOT and GPT enzyme activities
   in DPLOs with or without SS treatment following a 36‐hour exposure to
   drug or blank conditions. d) Relative expression of Got1, Got2, Gpt,
   and Cry1 by mRNA sequencing comparing samples from different time
   points after SS (SS‐0 h, SS‐6 h, SS‐12 h, SS‐18 h, SS‐24 h) with the
   control group. e) IC50 comparison of oxaliplatin applied on DPLOs
   treated with or without SS at different time points for a 36‐hour
   period. f) Protein profile heatmap of oxaliplatin or non‐oxaliplatin
   applied on DPLOs at 6 hours or 18 hours after SS group or control
   group. g and h) Comparison of DEGs by protein quantitation analysis in
   DPLOs treated with or without oxaliplatin over a 36‐hour period
   beginning at 6 hours or 18 hours after SS, as well as in the control
   group. i and j) KEGG pathway enrichment analysis of DEG sets by protein
   expression comparing oxaliplatin treatment and serum treatment DPLOs to
   control DPLOs. A threshold of Q‐value (corrected P‐value) ≤ 0.05 was
   applied, defining KEGG pathway that exhibited significant enrichment
   among the candidate genes. k) Temporal dynamics of drug target gene
   expression related to circadian rhythm and liver function for the top
   100 best‐selling drugs.^[ [150]^15 ^] * indicates p < 0.05, **indicates
   p < 0.01, ***indicates p < 0.001.

   In RNA‐seq analysis, compared to the control group, the synchronized
   DPLOs without drug treatment displayed rhythmic changes in Cry1
   expression, but no corresponding rhythmic changes were shown in Got1,
   Got2 and Gpt (Figure [151]5d). It suggested that Got1, Got2 and Gpt did
   not have intrinsic rhythmicity. Furthermore, proteomic analysis of
   drug‐treated DPLOs revealed that the expression levels of GOT and GPT
   in the control group were intermediate between those in the SS‐6 h and
   SS‐18 h DPLOs (Figure [152]S8, Supporting Information). Thus, the
   varying expression of GOT and GPT was induced by the different timing
   of oxaliplatin administration rather than SS itself.

   Subsequently, cell viability assay was conducted to assess the drug
   safety. The results showed no significant differences in cell viability
   changes at different time points in the non‐synchronized groups (C‐6 h,
   C‐12 h). However, the rhythmic groups displayed substantial variations
   in cell viability after drug administration at different time points
   (SS‐6 h, SS‐12 h). Specifically, SS‐6 h DPLOs exhibited lower
   sensitivity to oxaliplatin compared to the SS‐12 h DPLOs. Compared to
   the non‐synchronization group, the SS‐12 h groups demonstrated
   significantly lower half maximal inhibitory concentration (IC[50])
   values, indicating their higher sensitivity to the drug, while SS‐6 h
   DPLOs exhibited similar sensitivity to the drug as the
   non‐synchronization group (Figure [153]5e). AM‐PI staining yielded
   consistent results with the IC50 curves at an oxaliplatin concentration
   of 10^−4 mM (Figure [154]S9, Supporting Information). Together, it
   proved the liver toxicity of oxaliplatin was rhythm‐dependent, and
   SS‐DPLOs possessed the ability to evaluate the rhythm‐dependent
   toxicity.

   Taking count of the expression of rhythmic marker Cry1, the sensitivity
   of DPLOs on oxaliplatin is in negative correlation with the expression
   of Cry1 (Figure [155]5a). The results shed light on the potential of
   Cry1 as a valuable marker for determining the optimal drug
   administration timepoint with minimized toxicity in clinical therapy.

2.6. Mechanisms of Varying Toxicity of Oxaliplatin on Chronotoxici‐Plate

   To further investigate the effects of multi‐timepoint drug
   administration on rhythmic DPLOs, we performed Isobaric Tags for
   Relative and Absolute Quantitation (iTRAQ) analysis on DPLO treated
   with oxaliplatin at different time points (Figure [156]S7, Supporting
   Information). The results revealed significant differences in protein
   expression levels at the 6‐hour after synchronization
   oxaliplatin‐treated group (SS‐6h‐O) with the 18‐hour group (SS‐18h‐O),
   compared to their respective untreated groups (SS‐6 h, SS‐18 h).
   Besides, there were minimal differences observed between the
   oxaliplatin‐treated group without synchronization (C‐6h‐O) and the
   untreated group (C‐6 h) (Figure [157]5f; Figure [158]S10, Supporting
   Information). These findings indicate that rhythmic DPLOs possess
   varying levels of toxicity at different timepoint, consistent with the
   IC50 conclusions (Figure [159]5e). When comparing differentially
   expressed proteins (DEPs) in DPLOs with or without oxaliplatin
   treatment and synchronization, it was observed that SS‐6h‐O (compared
   to SS‐6 h, 540 DEPs) had more DEPs than SS‐18h‐O (compared to SS‐18 h,
   111 DEPs), and both were higher than the DEPs in C‐6h‐O compared to
   C‐6 h (Figure [160]5g). Additionally, the number of DEPs in SS‐6h‐O
   versus C‐6h‐O and SS‐18h‐O versus C‐18h‐O was significantly greater
   than in the corresponding untreated groups (Figure [161]5h). Thus,
   oxaliplatin treatment resulted in a greater number of DEPs in rhythmic
   DPLOs compared to the untreated groups (Figure [162]S11, Supporting
   Information). Rhythmic DPLOs demonstrated enhanced sensitivity in
   reflecting drug safety, and the impact of oxaliplatin on DPLOs varied
   across different time points.

   Furthermore, DEPs enrichment analysis was performed based on the
   proteomic results of different groups. When comparing the groups with
   and without synchronization treatment, enrichment of DEPs was observed
   across various pathways, including the rhythmic process pathway
   (Figure [163]5i). Additionally, the cell proliferation and cell killing
   pathways were enriched, which is consistent with the inhibitory effect
   of oxaliplatin^[ [164]^52 , [165]^53 ^] on DNA synthesis and cell
   proliferation (Figure [166]5i). When comparing the drug‐treated and
   untreated groups, we observed changes in multiple pathways after drug
   administration, with SS‐6 h group showing more pronounced alterations
   (Figure [167]5j; Figure [168]S11, Supporting Information), including
   changes in the cell‐killing pathway. However, such changes were not
   observed in the SS‐18 h group. Oxaliplatin administered at 6 hours
   after synchronization (SS‐6h‐O) showed a more profound impact on DPLOs
   compared to 18 hours (SS‐18h‐O), implying time‐dependent variations in
   oxaliplatin toxicity (Figure [169]5j). Moreover, the number of enriched
   pathways between the control groups with and without oxaliplatin
   treatment (C‐6 h/C‐6h‐O) were minimal, possibly due to the lower
   sensitivity of non‐rhythmic DPLO in evaluating drug safety, as previous
   described (Figure [170]5j).

   To evaluate the potential of Chronotoxici‐plate as an effective tool
   for assessing the efficacy of chronotherapy on those drugs, we selected
   liver‐related genes from the circadian‐related drugs among the Top 100
   best‐selling drugs and generated a gene expression correlation heatmap
   (Table [171]S1, Supporting Information). If a gene's expression varies
   over time in a rhythmic manner, its relative expression compared to the
   non‐rhythmic state can be represented by a periodic function symmetric
   about the x‐axis. Consequently, if the ratio of relative expression
   between two consecutive time points spaced over 6 hours was positive,
   then the next 6‐hour interval would be negative, and vice versa. The
   ratio between time points spaced 12 hours apart was always negative.
   The heatmap reveals that these genes exhibit periodic changes in
   rhythmic DPLO (Figure [172]5k), providing a foundation for exploring
   the optimal administration time and mechanisms of the aforementioned
   prevalent drugs. Furthermore, this approach also holds the promise for
   effective toxicity screening of drug candidates.

3. Conclusion

   The crucial role of circadian biology in amplifying drug efficacy and
   minimizing toxicity is becoming increasingly acknowledged. The
   biological state of target cells is instrumental in modulating an
   organ's response to a drug at specific times, thereby affecting
   pharmacodynamics.^[ [173]^1 ^] Overlooking these time‐dependent effects
   may lead to premature dismissal of potentially useful drugs^[ [174]^23
   ^] or provoke unnecessary side effects in patients.^[ [175]^14 ^]

   Highlighting the significance of considering circadian rhythms, the
   case of oxaliplatin is instructive. Initially, Rhone‐Poulenc‐Rorer^[
   [176]^54 ^] in France halted Phase I clinical trials due to excessive
   toxicity. However, the drug found a new lease of life when Debiopharm,
   a Swiss company, focused on assessing its safety and efficacy using
   chronopharmacology. Following successful clinical trials,^[ [177]^49 ,
   [178]^55 , [179]^56 , [180]^57 ^] oxaliplatin is now globally
   recognized as a vital treatment for cancer. This underscores the
   necessity for incorporating circadian rhythms into clinical trials and
   the development of novel therapies.

   The comprehensive screening of drug administration timings could have a
   significant impact on the pharmaceutical industry, especially when
   considering drugs that have previously failed in clinical trials.
   Integrating chronotherapy considerations into the process of new drug
   development presents a unique opportunity for advancing medical
   treatment. However, developing high‐throughput and cost‐effective
   models that accurately represent the effect of the circadian clock on
   drug efficacy remains a significant challenge.

   To tackle this challenge, we utilized 3D organoid printing technology
   to establish Chronotoxici‐plate, which incorporated in vitro
   synchronization treatment on DPLOs. This innovative tool enables
   circadian drug screening within a mere 4‐day period and comprehensive
   safety assessments within a week (Figure [181]3, [182]4). It represents
   a high‐throughput solution for studying drug administration in the
   context of circadian rhythms, facilitating a more rapid and accurate
   drug screening process. To simplify the incorporation of circadian
   assessments in clinical therapy, Cry1 has been identified as a
   potential companion diagnostic marker.

   The culture of primary cells in vitro presents unique challenges due to
   the limited cell interactions in a 2D environment. These limitations
   lead to lower cell proliferation, diminished viability, and weakened
   self‐assembling abilities. In particular, the reduced interactions
   among hepatocytes result in hampered formation of bile canaliculi and a
   decrease in essential signaling pathways necessary for standard
   hepatocyte function.^[ [183]^58 ^]

   Our study, however, revealed that circadian organoids exhibit certain
   variations in biological activity compared to their non‐circadian
   counterparts (Figures [184]2, [185]4). These results indicate that the
   incorporation of circadian rhythmicity into primary organoid culture
   could provide promising avenues for enhancing organoid development.
   Future research will seek to illuminate the underlying mechanisms that
   contribute to the beneficial effects of rhythmicity on organoid
   development.

   Further, our evaluation of oxaliplatin's safety using DPLOs in the
   Chronotoxici‐plate demonstrated its usefulness as a sensitive tool for
   assessing drug safety in the realm of chronopharmacology. The
   observation of a ten‐fold difference in IC50 values at different
   administration times aligns with the conclusion of clinical trials^[
   [186]^23 ^] indicating varying toxicity levels of oxaliplatin depending
   on the time of administration (Figure [187]5d). Several studies have
   demonstrated the influence of circadian rhythms on the pharmacological
   effects of oxaliplatin, but few have investigated the underlying
   mechanisms of this rhythm and its interaction with oxaliplatin within
   mammalian organism.^[ [188]^59 , [189]^60 , [190]^61 , [191]^62 ^] The
   relationship between these two may be a breakthrough in the mechanism
   of action of oxaliplatin and deserves more in‐depth study.

   The study also revealed the rhythmicity of popular drug targets in the
   Chronotoxici‐plate (Figure [192]5j), highlighting the potential to
   develop chronotherapy for existing drugs. Moreover, 3D organoid
   printing technology enables automated and high‐throughput organoid
   production and patterning in multi‐well plates. Compared to traditional
   animal models, DPLOs in Chronotoxici‐plate could offer over a 1000‐fold
   increase in chronotherapy screening efficiency, which supports the
   completion of chronotoxicity assessment within 7 days.^[ [193]^63 ,
   [194]^64 , [195]^65 , [196]^66 , [197]^67 , [198]^68 ^]

   Despite these advancements, there are still several challenges and
   future research directions that require further exploration. First, it
   would be invaluable to expand the application of Chronotoxici‐plate to
   organoids derived from a variety of tissue types – such as kidney,
   heart, or even tumor tissues – would be invaluable. This could provide
   a means for conducting organ‐specific chronotherapy evaluations and
   exploring the pan‐organ applicability of Cry1 as a companion diagnostic
   marker.

   Second, recognizing the nuanced differences in circadian behaviors
   between rodents and humans – such as the polyphasic sleep patterns in
   rodents^[ [199]^69 ^] – prioritizing organoids derived from human cell
   sources, such as induced pluripotent stem cells (iPSCs) or primary
   human tissues, should be considered. This strategy would serve to
   enhance the clinical translational potential of Chronotoxici‐plate.

   Finally, there are numerous commonly used drugs with significant
   chronotherapeutic potential that have not yet been evaluated for
   optimal timing of administration. This highlights the need for more
   comprehensive examinations of pharmaceuticals in future studies,
   further emphasizing the importance of this developing field.

4. Methods

4.1. Cell Culture

   AML12 (alpha mouse liver 12) cells were used as the control of liver
   organoids. Cells were cultured in a medium comprising Dulbecco's
   modified Eagle's medium (DMEM/F12) (Gibco) supplemented with 10% fetal
   bovine serum (FBS) (Gibco), 1% (w/v)
   Penicillin‐Streptomycin‐Amphotericin B Solution (P/S/A) (Procell), 1×
   Insulin‐Transferrin‐Selenium (ITS) (Thermo Fisher), and 40 ng ml^−1
   dexamethasone (Glpbio).

   Primary mouse liver tissues were obtained from BALB/c mice (male, 6–8
   weeks old) euthanized by cervical dislocation. The tissue processing
   method was previously described^[ [200]^74 ^] The cells were then
   resuspended in DMEM/F12 supplemented with 10% (w/v) FBS and 1%
   (w/v) P/S/A. The cell number was counted for organoid fabrication.

4.2. DPLO Precursor Fabrication and Chronotoxici‐Plate Construction

   The cells isolated from mouse liver were suspended in Matrigel
   (Corning) on ice to avoid the crosslink of Matrigel and the
   cell‐Matrigel mixture was loaded in a 1 mL syringe and installed in an
   injection pump. A fluorocarbon oil (HFE‐7000, 3 M Novec) was loaded
   into a 10 mL injection syringe and installed in another injection pump.
   Both pumps were placed in a 0–4 °C refrigerator. The two phases were
   co‐injected simultaneously through polytetrafluoroethylene
   (PTFE) tubing (the inner diameter ID = 560 µm, Woer) into a third piece
   of PTFE tubing via a 3‐way hand‐made PDMS connector. The Matrigel phase
   was injected at the flow rate of 8.0 µL mi^−1n and subsequently sheared
   into monodisperse droplets by the fluorocarbon oil at the flow rate of
   48.0 µL min^−1.^[ [201]^75 , [202]^76 ^] The tubing, along with the
   Matrigel droplets collected in it was heated to 37 °C for 25 min to
   complete solidification. The tubing was then connected to a 3D printing
   device. On one end, the tubing‐in‐the‐printer was linked to a pump
   filled with oil, while the other end was connected to a three‐port
   connector equipped with an optical sensor. One outlet of this connector
   was connected to a controlled compressed gas port, and the other was
   linked to the printing nozzle. When the optical sensor at the nozzle
   identified an organoid in transit, the three‐port connector's
   compressed air valve was actuated, propelling the organoids into the
   designated 96‐well culture plate beneath, containing primary liver
   organoid culture medium^[ [203]^29 ^] (DMEM/F12 supplemented with 10%
   (w/v) FBS, 1 mM nicotinamide, 0.1 µM dexamethasone, 1% P/S/A, 1× ITS,
   5 ng mL^−1 EGF and 50 ng ml^−1 HGF) individually (Figure [204]1). All
   organoid precursors were cultured at 37 °C in an incubator, supplied
   with 5% (v/v) CO2. Medium was changed every 2 days. Subsequently, the
   organoids were either harvested for further analysis or conditioned
   with drugs.

4.3. Non‐Engineered Organoids (NEOs) Fabrication

   The cells isolated from mouse liver were suspended in Matrigel
   (Corning) on ice to avoid the crosslink of Matrigel and aliquotes of
   cell suspension were placed in a 96‐well plate format at 20 µl drop^−1
   so that a small dome of Matrigel formed in the center of each well, and
   plates were then inverted incubated at 37 °C for 25 min until the
   Matrigel solidified. Subsequently, primary liver organoid culture
   medium was added to each well. The culture medium was changed every
   48 h. Afterward, the organoids were either harvested for further
   analysis or conditioned with drugs.

4.4. Serum Shock

   The organoid reached a stable condition after ≈4 days. Then the primary
   liver organoid culture medium was exchanged with serum‐rich medium
   (DMEM/F12 supplemented with 50% FBS^[ [205]^10 ^]), after
   30 minitues,^[ [206]^77 ^] the medium was replaced with the
   aforementioned culture medium.

4.5. Cell Viability Assay

   The viability of cells in the organoids or organoids precursor was
   assessed using a LIVE/DEAD staining kit (Yeasen, China). The organoids
   were washed with 1 × PBS buffer and then immersed with 2.0 × 10^−6 M
   Calcein AM (the live cell staining dye) and 4.5 × 10^−6 M PI (the dead
   cell staining dye) for 15 min, followed by gentle 1×PBS buffer washing.
   All microscopic images were captured under a confocal microscope
   (N‐STORM). All images were generated using z‐plot analysis in ImageJ.

   The viability of cells in organoids was also evaluated using
   CellTiter‐Glo® 2.0 Cell Viability Assay by quantitating the amount of
   ATP present, which indicates the presence of metabolically active
   cells. Following removal of the culture medium, the organoids were
   washed with 1×PBS buffer and then immersed with CellTiter‐Glo® 2.0
   Reagent. Finally, the luminescence value was measured by Spark
   multimode microplate reader (TECAN, SPARK).

4.6. Measurement of RNA concentration and real‐time quantitative PCR
(RT‐qPCR)

   Total RNA from organoids or 2D cells was extracted using Eastep® super
   total RNA extraction kit (Promega) and transcribed to cDNA by the
   PrimeScript™ RT reagent Kit (TAKARA). cDNA was used for qPCR using the
   GoTaq® qPCR Master Mix (Promega, USA) with RT‐qPCR instrument (Bio‐Rad,
   CA). The primers used for the qPCR reaction are shown in Table [207]S1
   (Supporting Information). To ensure uniform baseline values across all
   curves, each data point was normalized by the value at t = 0. This
   normalization eliminated variations attributed to distinct RNA
   extraction amounts, facilitating comparisons between different groups.
   Additionally, we evaluated the statistical significance (p‐value) of
   the curve fits were evaluated for both SS‐treated and control (C)
   groups.

4.7. Western Blot

   Cells cultured in Matrigel were collected and lysed with RIPA buffer
   (MCE) containing protease and phosphatase inhibitors (MCE). Protein
   concentrations were determined by the BCA Protein Assay Kit (23 227,
   Thermo Fisher Scientific). Western blotting was conducted according to
   a standard protocol. Briefly, the proteins were loaded on 10% (w/v)
   SDS‐polyacrylamide gels and electrophoretically transferred to a PVDF
   membrane (Millipore). The primary antibodies were used as follows:
   β‐actin (1:1000, Cell Signaling Technology, 8H10D10), CRY1 (1:1000,
   Proteintech, 13474‐1‐AP). Detection was performed using a
   Chemiluminescent Western Blot detection kit (4AW012‐1000, 4A Biotech).
   The western blot results were analyzed using ImageJ (version 1.52).

4.8. Drug Safety Evaluation on DPLOs

   Primary liver organoids were printed on 96‐well plates or 24‐well
   plates (for iTRAQ analysis). After 4‐day development, organoids began
   to be exposed to 0, 0.1, 0.01, 0.001, 0.0001 mM oxaliplatin for 36 h at
   different time points. Cell viability assay and other evaluating was
   performed as described above.

4.9. Measurement of Glutamic‐Oxalacetic Transaminase/ Glutamic‐Pyruvic
Transaminase (GOT/GPT) Enzyme Activities

   Organoids were exposed to 0.1 mM oxaliplatin for 36 h at designated
   time points (12 h,18 h) after SS treatment. After being collected and
   washed for 3 times with PBS, the organoids were suspended in cold
   physiological saline, and later were subsequently disrupted via
   ultrasonic cell homogenizer (Scientz‐IID) to extract lysed proteins.
   The activities of GOT and GPT were measured using detection kits
   purchased from Jiancheng Bioengineering Institute (Nanjing, China). The
   process followed the manufacturer's instructions. Optical density (OD)
   values were measured using an enzyme‐labeled instrument (TECAN, SPARK).

4.10. Histologic Analysis and Immunostaining and of Rhythmic DPLOs

   DPLOs were cultured in the primary liver organoid culture medium for 4
   days and subjected to SS. Subsequently, the DPLOs were fixed using a
   paraformaldehyde‐glutaraldehyde fixative (0.5%/0.625%, Biosharp) for 15
   minutes at room temperature, followed by three‐time washes with PBS.
   Dehydration of the DPLOs was performed using a 20% saccharose solution.
   The samples were then immersed in an optimal cutting temperature (OCT)
   compound and incubated for 12 hours, after which they were rapidly
   frozen in liquid nitrogen. The samples were cryo‐sectioned into 10‐µm
   slides. HE staining was conducted following the manufacturer's
   instructions, and images were captured using an Olympus VS200 Research
   Slide Scanner.

   For immunofluorescence, the slides were blocked with 10% (w/v) normal
   donkey serum in PBS supplemented with 0.1‐1% Triton X‐100 for 1 hour at
   room temperature. Subsequently, the slides were incubated at 4 °C for
   12 hours with mouse Cytokeratin 19 (1:100, XH3643123, Invitrogen),
   rabbit Albumin (1:100, 0806–9, Huabio), mouse Cytokeratin 18 (1:100,
   M0407‐19, Huabio), rabbit HNF4a (A20865, 1:100, ABclonal), rat
   anti‐CD68 (1:100, 14‐0681‐82, Invitrogen). Slides were washed three
   times with PBS and then incubated with secondary antibodies, including
   anti‐rabbit secondary antibody (1:500) (Alexa Fluor 647, Abcam),
   anti‐mouse secondary antibody (1:500) (Alexa Fluor 488, Abcam), and
   anti‐rat secondary antibody (1:500) (Alexa Fluor 488, Abcam), at room
   temperature for 1 hour. Afterward, the slides were mounted using DAPI
   (1:1000), and the images were captured using a confocal microscope
   (N‐STORM). Negative control experiments were conducted in parallel.

4.11. Oil Red O Staining of DPLOs

   Oil Red O staining was applied to evaluate the de novo lipogenesis in
   DPLOs samples. The samples were divided into 2 groups, one was cultured
   in the primary liver culture medium for 3 days, followed by subsequent
   collection. In contrast, the second group was cultured for an
   additional day, after which it underwent SS treatment or kept as
   control, and continued to be cultured until day 7. Subsequently, the
   DPLOs were fixed using a paraformaldehyde‐glutaraldehyde fixative
   (0.5%/0.625%, Biosharp) for 15 minutes at room temperature, followed by
   three‐time washes with PBS. Dehydration of the DPLOs was performed
   using a 20% saccharose solution. The samples were then immersed in an
   optimal cutting temperature (OCT) compound and incubated for 12 hours,
   after which they were rapidly frozen in liquid nitrogen. The samples
   were cryo‐sectioned into 10‐µm slides. Oil Red O staining (G1261,
   Solarbio) was conducted following the manufacturer's instructions, and
   images were captured using an Olympus VS200 Research Slide Scanner.

4.12. RNA‐Seq Analysis

   The mRNA sequencing of the control group and SS groups was performed in
   BGI‐Shenzhen, China. Briefly, the control group was taken out after
   four‐day development and treated with the normal primary liver organoid
   culture medium, while the SS groups were treated using SS medium for
   0.5 h in 37 °C and then cultured in the normal culture medium before
   being taken out at specific time points. The sequencing was performed
   on BGISEQ platform and SOAPnuke v1.5.2^[ [208]^78 ^] was used to filter
   out the low‐quality reads. And HISAT v2.0.4.^[ [209]^79 ^] was used to
   the alignment of reads to the reference genome
   GCF_0 00001635.26_GRCm38.p6.

4.13. iTRAQ Quantification Proteomics Analysis

   The protein quantification of control group and SS group with or
   without oxaliplatin at different time points was performed with
   Isobaric tags for relative and absolute quantitation (iTRAQ) technology
   in BGI‐Shenzhen, China. Totally 906 357 spectrums were generated,
   35 136 peptides and 5879 proteins were identified with 1% FDR. The
   quantification process followed the company's protocol.

4.14. Statistical Analysis

   All the statistical analysis was performed using Origin 2021b
   (OriginLab Co, Northampton, MA). In measuring the RNA concentration of
   RT‐qPCR, we normalized each data point by the value at t = 0. A
   statistically significant difference was considered when the value of p
   < 0.05 (* indicates p < 0.05, ∗∗ indicates p < 0.01, ***indicates p <
   0.001), using unpaired two‐failed Student's t‐test (comparing two
   experimental groups). Error bars in all figures are the standard
   deviations obtained from at least three independent measurements unless
   otherwise stated. All representative fluorescence images, H&E staining,
   and bright field images shown in the main text and supplementary
   information were independently repeated at least three times.

4.15. Ethics Statement

   All the animal procedures involved in the study were reviewed and
   approved by the Animal Experimentation Ethics Committee at Tsinghua
   Shenzhen International Graduate School, Tsinghua University (Project
   2023/22).

Conflict of Interest

   The authors declare no conflict of interest.

Author Contributions

   J.Z., Y.‐C.H., and W.W. contributed equally to this work. J.Z., J.L.
   and S.M. raised the innovation, designed the experiments, data analysis
   and organized the experiments. J.Z., W.W., Y.H., J.L., Z.Y., H.Y., Z.W
   and Y.Z performed the experimental work. J.Z., W.W., Y.H. and K.H.
   contributed to Figure preparation. S.M., J.Z., J.L. contributed to
   paper writing and revising. S.M. and J.L. conceived and advised on this
   work. All authors contributed to discussion.

Supporting information

   Supporting Information
   [210]ADVS-11-2305925-s001.pdf^ (1.1MB, pdf)

Acknowledgements