DCDs, formerly known as “non-heart-beating donors”, exhibit notably different pathophysiology from that of brain death donors. DCDs can be divided into “controlled” and “uncontrolled” categories according to the Maastricht classification [44]. Although there has been considerably renewed interest in DCD donors to increase the pool of available organs, there are challenges associated with the additional hypoxic damage in DCD grafts, such as the increased incidence of primary non-function, early dysfunction, and cholangiopathy in controlled DCD transplantation [59]. Additionally, uncontrolled DCD grafts that experience warm ischaemic time of more than 30 min are excluded from transplantation worldwide, owing to the higher incidence of primary non-function, which results in a great loss of donated organs [45].

Innovative strategies and research are needed to improve the application of DCD liver grafts. However, animal studies have primarily focused on graft preservation, including ex vivo machine perfusion and modified perfusates in rat or swine models [4650]. To our knowledge, only Liu et al. have reported syngeneic liver transplants from DCD mice; however, the surgery in their study was performed without hepatic artery reconstruction, which differs from clinical transplant procedures [51]. To strictly adhere to clinical surgery procedures, we first attempted liver transplantation from C57Bl/6 to C57Bl/6 mice; however, because of hepatic artery variations, the transplant surgery success rate was only approximately 30 %, which was not suitable for establishing an animal model. Because CBF1 mice have heterosis and do not reject liver grafts from C57Bl/6 mice, we then attempted liver transplantation from C57Bl/6 to CBF1 mice. Unexpectedly, the success rate of transplantation with hepatic artery reconstruction reached 90 %, and the survival rate and graft histology in this mouse model were consistent (Additional file 1: Figure S1). On the basis of a large amount of preliminary data, we established an optimal arterialized mouse NHB liver transplantation model characterized by 10 min of NHB time and 120 min of cold ischaemic time. The NHB5min group showed no significant differences compared with the NHB0min group in terms of serum ALT/AST levels, Suzuki scores, and 14-day survival rates, thus indicating that warm ischaemia for less than 5 min did not induce obvious graft injury or damage. According to the calculation of Giraud et al. [52], 5 min of warm ischaemia in mice is equivalent to 42.8 min in humans. Thus, human DCD liver grafts with a cardiac arrest time of less than 42.8 min may be acceptable, a result consistent with the clinical criterion that DCD grafts used within 30 min of warm ischaemia are safe [45]. Furthermore, mRNA levels of pro-inflammatory cytokines associated with the Th1/Th2 response, with the exception of IL2 and IL4, decreased with increased NHB duration. The exact mechanism underlying these results remains unclear, but prolonged NHB time may potentially induce severe injury or even apoptosis of intrahepatic non-parenchymal cells, ultimately resulting in decreased mRNA levels for the majority of Th1/Th2 cytokines. However, it is unknown why IL2 and IL4 are exceptions to this trend, and further investigation is required.

MSCs have powerful immunomodulatory and regenerative capacities, thus making them a promising therapeutic option for treatment of clinical diseases. However, clinical trials investigating MSCs in liver transplantation are still in process, and the ultimate effects of MSCs on prolonged graft survival, reduced transplantation side effects, and concomitant therapy have yet to be evaluated [53].

In our study, MSC infusion dramatically ameliorated NHB-induced graft injury and increased recipient survival post-transplantation. Mechanistically, MSCs inhibited Kupffer cell apoptosis by producing PGE2 in an ERK1/2-dependent manner. Together, protected Kupffer cells and infused MSCs suppressed Th1/Th17 cells, which further decreased chemokine expression and inhibited inflammatory cell infiltration, thereby resulting in a reduction in NHB-induced graft injury. Furthermore, 10 min of NHB graft injury was reversed by MSC infusion. Graft injury in the MSC infusion group was not significantly different compared with grafts in the NHB0min group, thus suggesting promisingly clinical application of MSCs in uncontrolled DCD grafts. Uncontrolled DCD liver grafts with warm ischaemia time of less than 85.6 min may achieve outcomes comparable to those of controlled organs if combined with MSC infusion. Additionally, the lack of tumourigenesis observed during a prolonged observation time of 3 months indicates the safety and effectiveness of MSC infusion. However, additional clinical studies must be performed to verify the feasibility of our proposed method.

Kupffer cells, known as liver-resident macrophages, account for approximately 30 % of liver non-parenchymal cells and more than 50 % of resident macrophages throughout the entire body [5456] and play an important role in immunomodulation, phagocytosis, biochemical attack, and liver IRI [57]. Kupffer cells create an inflammatory milieu by generating cytotoxic ROS, cytokines such as TNFα and IL1β, and neutrophil-associated chemokines such as CXCL1, CXCL2, CCL2, and CXCL10 [5863]. In contrast, IL10 released by Kupffer cells may induce anergy in T cells and protect tissues against excessive injury [64, 65].

Interestingly, in our study, the number of F4/80-positive Kupffer cells decreased rapidly with the extension of warm ischaemia time, whereas MSC infusion restored Kupffer cells by inhibiting apoptosis. The protected Kupffer cells suppressed Th1/Th17 cells and depressed the intrahepatic infiltration of neutrophils and T cells by reducing cytokine and chemokine secretion, thus resulting in reduced graft injury. Moreover, the protective effects of MSCs on NHB liver grafts were dependent on the presence of Kupffer cells. Together, our data provide the first evidence of the extremely important role of Kupffer cells in determining NHB liver graft quality. In the future, Kupffer cells or F4/80 expression levels may be valuable in evaluating the quality of DCD liver grafts and predicting the prognosis of DCD liver transplantation.

The paracrine effects of MSCs are critical components of their immunomodulatory functions. Among numerous soluble factors, PGE2 is the most well-characterized, affecting Th17 differentiation and regulating modulatory macrophages and dendritic cells [66, 67]. In vitro data revealed the important role of PGE2 in suppressing Kupffer cell apoptosis. We also elucidated the involvement of the TLR4-ERK1/2-Fas/FasL-caspase3 signalling pathway in this process. Notably, the underlying mechanism revealed in our study broadens insight and allows for a better understanding of how MSCs protect NHB grafts (Additional file 2: Figure S2); however, other mechanisms involved in this protection require further investigation.