Modeling mammalian hibernation to improve organ cold preservation: Using the intestine as an example

Dear Editor,

Organ transplantation is the ideal treatment for patients suffering from terminal organ failure.¹ As the first step in organ transplantation, cold preservation of organs is crucial for transplantation outcomes. However, donor organs are vulnerable to the effects of static cold storage (SCS), the standard clinical cold preservation method, resulting in a safe cold ischemic time of < 6 h for hearts and lungs, < 10 h for intestines, < 12 h for livers, and ~24 h for kidneys.²

In nature, many mammals employ hibernation to survive a harsh environment.³ During hibernation, some mammalian hibernators can drop their body temperature from ~37 °C to ~4 °C and experience alternating torpor-arousal phases, while their organs are protected from such harmful insults.⁴ Therefore, mammalian hibernators may serve as ideal models for studying organ cold preservation.

The intestine is extremely susceptible to SCS, because of its hollow and monolayer epithelial structure.⁵ Prolonged SCS breaks down intestinal tight junctions and further disrupts mucosal barriers, causing mucosal edema, villous shedding and bacterial translocation.⁶ In addition, cold storage injury can lead to the loss of intestinal stem cells and thus diminish the damage repair and renewal capacity of intestinal allografts.⁷ Hence, in this study, we used the intestine as an example to seek molecular clues regarding the cold adaptation of the Syrian hamster (Mesocricetus auratus), a mammalian hibernator.

We successfully induced a torpid state in Syrian hamsters by simulating their hibernation condition (e.g., temperature, light, etc.) (Fig. 1a; Supplementary information, Fig. S1a, b, and Table S1). Specifically, in this study, the active state designates that Syrian hamsters never experienced hibernation induction, whereas the torpid state designates that Syrian hamsters were induced into deep torpor. Syrian hamsters entered deep torpor after ~12 weeks of induction and were sampled after 5–6 torpor-arousal cycles. In the torpid state, the body temperatures of Syrian hamsters drastically decreased from ~36 °C to ~5 °C. Thus, Syrian hamsters could cope well with multiple rounds of cooling-rewarming stresses.

Fig. 1. Molecular mechanisms of intestinal cold adaptation during mammalian hibernation.

a Images showing active and torpid Syrian hamsters and a representative body temperature trace showing torpor-arousal cycles. b H&E staining of mouse, active and torpid Syrian hamster intestines after 0 h and 48 h SCS. Intestinal injury was evaluated by Chiu/Park’s histology score (n = 3 individuals). Data are presented as mean ± SD; t-test, P values are indicated. These images are also presented as a part of Supplementary information, Fig. S1d. c H&E staining of active and torpid Syrian hamster intestines used for scStereo-seq. d Spatial visualization of intestinal cell types and their ratios. e Violin plots showing tight junction and stemness scores in snRNA-seq data; t-test, P values are indicated. f Images showing shIOs in different culture media (CM) and passages (P). g Calcein AM/PI staining of live (green) and dead (red) cells in mIOs and shIOs (n = 3 replicates). h Venn plot showing gene numbers that are upregulated in stem/progenitor and TA cells of torpid Syrian hamster intestines and shIOs under cold stress. i Violin plot showing Ace2 expression in snRNA-seq data. j Western blot assay (GAPDH as reference) and IF staining (n = 3 replicates) for ACE2 in active and torpid Syrian hamster intestines. Data are presented as mean ± SD; t-test, P value is indicated. k Calcein AM/PI staining of live (green) and dead (red) cells in mIOs treated with and without harpagoside (Har) (n = 3 replicates). l H&E staining of mouse intestines was compared between UW and UW with harpagoside groups. Intestinal injury was evaluated by Chiu/Park’s histology score (n = 3 individuals). Data are presented as mean ± SD; t-test, P value is indicated. m Schematic illustrating the B⁰AT1-dependent function of ACE2. n Spatial expression pattern of Slc6a19. o Survival ratio of mIOs examined by the ATP level between negative control (NC) group and Ace2&Slc6a19 overexpression (OE) group (n = 6 replicates). Data are presented as mean ± SD; t-test, P value is indicated. p IF staining for the apoptosis indicator CC3 in mouse intestines. Cin: cinromide (B⁰AT1 inhibitor). q Barplot showing Ace2 expression in wild type and Ace2-KO mouse intestines by RT-qPCR (Actb as reference, n = 3 individuals). Data are presented as mean ± SD; t-test, P value is indicated. r H&E staining of Ace2-KO mouse intestines compared between UW and UW with harpagoside groups. Intestinal injury was evaluated by the Chiu/Park’s histology score (n = 3 individuals). Data are presented as mean values ± SD; t-test, P value is indicated. s IF staining for the apoptosis indicator CC3 in Ace2-KO mouse intestines. t IF staining for the apoptosis indicator CC3 in human intestines. u IF staining for ZO-1 in mouse kidneys showing the tight junctions. v H&E staining of mouse pancreases showing islet injuries. Black arrow indicates damaged islets.

We wondered whether Syrian hamster intestines could survive cold stress better than those of their nonhibernator counterparts (i.e., mice). We perfused Syrian hamster and mouse intestines with University of Wisconsin (UW) solution, the standard organ cold preservation solution, and preserved them at 4 °C for 12 h, 24 h and 48 h, respectively (Supplementary information, Fig. S1c, d). Intestinal histology and injury were evaluated according to Chiu/Park’s standard,^8,9 in which a higher score indicates a more severe injury (Fig. 1b). The mouse intestines exhibited evident injuries at villus tips after 12 h SCS, large gaps down villus sides and visible crypt injuries after 24 h, and massive villus shedding and severe crypt damage after 48 h. By contrast, no visible crypt injuries occurred in either active or torpid Syrian hamster intestines within 48 h SCS treatment. However, the active Syrian hamster intestines showed small gaps at villus tips after 48 h, whereas the torpid ones presented no evident injury. Therefore, even active Syrian hamster intestines possess better cold adaptation abilities than mouse intestines, which may partially reflect the inherent characteristics of Syrian hamsters. Notably, such cold adaptation ability is further enhanced in the torpid state.

To study the cellular and transcriptomic dynamics between the torpid and active states, we constructed a spatiotemporal hibernation atlas of the Syrian hamster intestine (Supplementary information, Fig. S2a–d). We first profiled intestinal transcriptomes from 3 active and 3 torpid Syrian hamsters by single-nucleus RNA sequencing (snRNA-seq), by which the transcription states of intestines were better locked by flash freezing in liquid nitrogen immediately after dissection. In total, we obtained 30,465 single nuclei and annotated 12 cell populations, namely, stem/progenitor cells, absorptive transient amplifying (TA) cells, secretory TA cells, Paneth cells, enteroendocrine cells, tuft cells, goblet cells, enterocytes, endothelial cells, smooth muscle cells, fibroblasts and immune cells (Supplementary information, Fig. S2e, f, and Table S2). We then performed single-cell spatial enhanced resolution omics sequencing (scStereo-seq) coupled with snRNA-seq for mapping the spatial organizations of cell populations (Fig. 1c; Supplementary information, Tables S3, S4). The integration result revealed similar epithelial cell compositions between the torpid and active states (Fig. 1d).

Considering the importance of tight junctions and stemness for intestinal cold preservation, we calculated their scores in the snRNA-seq dataset (Fig. 1e). Importantly, both scores exhibited no significant differences between active and torpid intestines, which was also supported by the immunofluorescence (IF) staining for representative tight junction-related proteins (i.e., ZO-1 and OCLN) and stemness-related proteins (i.e., OLFM4 and HMGA1) (Fig. 1e; Supplementary information, Fig. S2g–j). Therefore, Syrian hamster intestines can maintain tight junctions and stemness under cold stress, suggesting the Syrian hamster as an ideal model for studying intestinal cold adaptation.

Next, we attempted to establish a hibernation model using Syrian hamster intestinal organoids (shIOs). However, both mouse and human culture media failed to support shIOs (Fig. 1f). We compared the differences between mouse and human culture media and modified the culture medium by adding or removing each constituent (Supplementary information, Fig. S3a). Finally, shIOs were successfully cultured by adding A8301 to mouse culture medium, which was also supported by snRNA-seq analysis (Fig. 1f; Supplementary information, Fig. S3b, c). We then utilized shIOs to model intestinal hibernation via cooling (4 °C) and rewarming (37 °C) experiments, and mouse intestinal organoids (mIOs) were used as controls (Fig. 1g; Supplementary information, Fig. S4). Notably, mIOs presented rough surfaces with cell shedding and increased in size after 12 h cooling, and underwent widespread cell death and failed to regenerate after rewarming. By contrast, shIOs could bear long-term cold stress, maintain a tight and intact 3D structure for up to 24 h and quickly restore proliferation after rewarming. Therefore, shIOs can be used to model hibernation-like torpor-arousal and will be useful tools to study intestinal cold adaptation.

Subsequently, we attempted to identify the underlying molecular clues regarding intestinal cold adaptation (Supplementary information, Fig. S5a–d). We focused mainly on stem/progenitor and TA cells, as they fulfil a continuous demand for differentiated cells, and their dysfunction results in complete disruption of the intestine. We first made a comparison between the torpid and active states in both the snRNA-seq and scStereo-seq datasets. Second, we compared the transcriptomes of shIOs and mIOs after 24 h at 4 °C (Supplementary information, Table S5). Finally, we intersected the upregulated genes under cold stress in these three datasets and identified 13 shared genes as cold adaptation-related candidate genes, namely, Ace2, S100a6, S100g, Slc15a1, Lgals3, Dpep1, Ethe1, Pmp22, Tnip1, Sdcbp2, Prr13, Krt8 and Fabp6 (Fig. 1h; Supplementary information, Fig. S5e–g and Table S6). Among them, Ace2 attracted our attention for three reasons: first, Ace2 was highly upregulated in almost all cell types in the torpid state (Fig. 1i, j); second, the ACE2 protein is important for the maintenance of epithelial barrier permeability;^10,11 third, ACE2 has been widely studied in human diseases such as hypertension and coronavirus infection, and numerous effective agonists are available.^12,13 Hence, we selected Ace2 as a priority candidate gene in this study.

Using mIOs that failed to survive cold stress, we tested 4 ACE2 agonists, namely, harpagoside, methazolamide, imatinib and DIZE, because of their effectiveness in binding to ACE2.¹² Surprisingly, harpagoside successfully protected mIOs from cold stress, although the other 3 agonists also moderately mitigated cold injuries (Fig. 1k; Supplementary information, Fig. S6a, b). Moreover, after 48 h at 4 °C, harpagoside promoted the growth of mIOs at 37 °C, suggesting preservation of stemness (Supplementary information, Fig. S6c). Indeed, expression levels of tight junction- and stemness-related genes significantly decreased in mIOs under cold stress but were well sustained after the addition of harpagoside (Supplementary information, Fig. S7a).

Notably, ACE2 activation could also mitigate tissue injury caused by intestinal cold preservation (Fig. 1l; Supplementary information, Fig. S7b, c). Compared with the standard UW solution, UW solution with harpagoside preserved most villi and crypt structures after 48 h and had less lamina propria injury after 72 h. Meanwhile, cell inflammation and apoptosis were mitigated by harpagoside, as evaluated by staining for the inflammatory cytokines TNF-α and HMGB1 and the apoptosis indicator cleaved caspase-3 (CC3) (Supplementary information, Fig. S7d). In addition, harpagoside also improved absorptive and peristaltic functions of cold-stored intestines (Supplementary information, Fig. S7e–g). Altogether, these results support the important role of ACE2 in intestinal cold protection.

There are two main roles of ACE2 in the intestine: one is to metabolize angiotensin (Ang) II into the beneficial peptide Ang 1-7 in the renin angiotensin system (RAS), and the other is to stabilize the neutral amino acid transporters (e.g., B⁰AT1), which can promote tight junction formation.¹⁴ For the RAS-dependent role of ACE2 (Supplementary information, Fig. S8a), we found that the Ang 1-7 receptor encoded gene, Mas1, was barely expressed in both active and torpid intestines, which was also supported by immuohistochemical staining for MAS1 (Supplementary information, Fig. S8b, c). Moreover, the addition of Ang 1-7 failed to improve intestinal protective effects under cold stress (Supplementary information, Fig. S8d–g). Thus, the RAS-dependent role of ACE2 does not seem to be involved in intestinal cold adaptation.

For the B⁰AT1-dependent role of ACE2 (Fig. 1m), we detected upregulation of gene expression of Slc6a19, which encodes B⁰AT1, in torpid intestines (Fig. 1n). Moreover, simultaneous overexpression of Ace2 and Slc6a19 significantly enhanced the cold resistance ability of mIOs (Fig. 1o; Supplementary information, Fig. S8h). By contrast, the B⁰AT1 inhibitor cinromide severely undermined the protective effect of harpagoside (Fig. 1p). Furthermore, to evaluate stemness, we cultured mIOs from intestinal tissues after 24 h SCS at 4 °C. After 96 h culture, more mIOs were generated from intestinal tissues in UW solution with harpagoside than in the standard UW solution, and the addition of the B⁰AT1 inhibitor indeed impaired the number and quality of mIOs (Supplementary information, Fig. S8i, j).

Moreover, we utilized Ace2-knockout (KO) mice to further confirm the causality of the ACE2-B⁰AT1 axis in intestinal cold protection. Both RT-qPCR and IF staining indicated the successful KO of Ace2 (Fig. 1q; Supplementary information, Fig. S8k). Notably, harpagoside failed to mitigate the cold injury in Ace2-KO intestines, and the stronger apoptosis signals in Ace2-KO intestines suggested their worse cold-resistance ability than control intestines during SCS at 4 °C (Fig. 1r, s). In addition, we also tested the ACE2 inhibitor ML364 in shIOs and detected an increased cell death rate compared with the control group after 48 h cooling (Supplementary information, Fig. S8l, m). Taken together, these results reinforce the important role of ACE2-B⁰AT1 axis in intestinal cold adaptation.

Given that human intestines are much larger and thicker than mouse intestines, we investigated whether harpagoside would preserve human intestines under cold conditions. After 48 h SCS, villous injury and cell apoptosis were observed in the intestines stored in standard UW solution but were not evident in the intestines stored in UW solution with harpagoside, indicating the efficacy of harpagoside in human intestines during SCS (Fig. 1t; Supplementary information, Fig. S9a).

Since ACE2 is widely expressed in many organs of the body,¹⁵ we also tested the protective effect of harpagoside on the cold preservation of mouse kidney and pancreas (Supplementary information, Fig. S9b, c). In the kidneys, harpagoside rescued the intercellular tight junctions and abnormal cytoskeleton distribution after 24 h SCS and rewarming, as shown by the ZO-1 and β-actin expression patterns, respectively (Fig. 1u; Supplementary information, Fig. S9d). In the pancreas, harpagoside protected the integrity of islets after 24 h SCS and rewarming (Fig. 1v; Supplementary information, Fig. S9e). Hence, these results suggest that harpagoside may also improve the cold preservation of the kidney and pancreas.

In conclusion, Syrian hamster intestines possess excellent cold adaptation ability and can maintain their tight junctions and stemness under cold stress, suggesting that the Syrian hamster constitutes an ideal animal model for studying intestinal cold adaptation. Moreover, we successfully constructed shIOs and used them to model hibernation-like torpor-arousal, indicating that shIOs will be useful tools to study intestinal cold adaptation easily and deeply. By combining both in vivo and in vitro hibernation models, we demonstrated the importance of ACE2 in intestinal cold adaptation and further suggested that ACE2 activation stabilizes B⁰AT1 for neutral amino acid uptake and regulates the intestinal mucosal barrier and cell stemness. Notably, among the 4 ACE2 agonists examined, harpagoside presented the best efficacy in intestinal cold protection. This observation may indicate that these molecules exhibit distinct action mechanisms or varying efficacies under room temperature and low-temperature conditions. Further investigations are required to screen more efficient ACE2 agonists, thereby enhancing the cold preservation of organs. In addition to Ace2, we also identified twelve other upregulated and one downregulated candidate genes that are related to cold adaptation. In the future, we will investigate their roles and other candidates in intestinal cold adaptation. Collectively, our study provides a new research paradigm for improving organ cold preservation and studying mammalian hibernation.

Author contributions

J.D., M.L., R.X. and G.C. designed the study; W.Y.H., W.D., C.W., W.J.H., Y.L., C.L., Y.C., R.L., Y.Z., J.L. and X.Z. performed the experiments; Z.H. conducted the bioinformatic analyses. J.D. and M.L. wrote the manuscript with the help of other authors.

Data availability

All data generated in this study are deposited in the Genome Sequence Archive (GSA) database (accession number: CRA015672).

Competing interests

The authors declare no competing interests. J.D. is responsible for the potential patents of the study in the future.

Supplementary information

The online version contains supplementary material available at 10.1038/s41422-025-01149-w.