The coronavirus disease 2019 (COVID-19) pandemic, caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), is rapidly and continuously spreading globally and can result in serious significant respiratory morbidity and mortality (1, 2). Although most patients with COVID-19 present with mild respiratory tract infection, severe pneumonia and acute respiratory distress syndrome (ARDS) have been described in 19% of reported cases, and the overall mortality is approximately 49% (3). The most common symptoms of SARS-CoV-2 infection are cough, fever, fatigue, headache, myalgia, and diarrhea. Approximately 1 week after the onset of symptoms, patients become severely ill when dyspnea and hypoxemia appear, and progressive respiratory failure subsequently develops (4). The main reason for severe COVID-19 disease is aberrant immune host response accompanied by a “cytokine storm.” More precisely, it involves the excessive activation and dysregulation of CD8⁺ T cells and immoderate pulmonary recruitment of immune cells, as well as the overproduction of pro-inflammatory cytokines such as interleukin 2 (IL-2), IL-6, IL-7, and tumor necrosis factor (TNF) (5). Additionally, endothelial dysfunction and immunothrombosis are also regarded as critical pathogenic mechanisms leading to severe COVID-19 (6) ( Figure 1 ).

Without effective antiviral medications, current therapeutic approaches are limited to aggressive standard supportive care and treatment of any other co-infections. More recently, a growing number of clinical investigations of cell-based therapies, primarily involving mesenchymal stromal cells (MSCs), have demonstrated safety and possible efficacy in the treatment of critical illness, particularly in sepsis and ARDS (7–15). Recent preclinical data in models of sepsis/ARDS and relevant related clinical studies of MSC administration in patients with ARDS may contribute to a better understanding of the potential MSC-based cell therapy approaches for COVID-19. Therefore, our review focused on MSCs as a potential therapeutic agent for COVID-19 on the basis of the current evidence from sepsis and ARDS.

MSCs were originally characterized in bone marrow stromal cells and are now the best-described and most widely used cells for cell-based therapies. According to the International Society of Cellular Therapy, MSCs are defined by the following minimal criteria: 1) must be adherent to plastic; 2) must positively present CD105, CD90, and CD73 and must be lacking CD45, CD34, CD14, CD11b, CD79 alpha, CD19, and HLA-DR surface makers; and 3) must have a differentiation potential in in vitro conditions (16).

MSCs were originally identified in the bone marrow, and the most recent studies have focused on the effects of bone marrow-derived MSCs. However, MSCs can also be isolated from several tissues such as umbilical cord/fetal blood, placenta, amniotic fluid, and adipose tissue (17–21). Recent studies on rodent models have shown that adipose tissue-derived MSCs minimized ischemia–reperfusion lung injury by suppressing oxidative stress and inflammatory reaction (22); improved phagocytosis and the bactericidal functions of macrophages in a model of Pseudomonas aeruginosa pneumonia (23); attenuated radiation-induced lung injury through anti-inflammation, anti-fibrosis, and anti-apoptosis mechanisms (24); and abrogated bleomycin-induced lung fibrosis by significantly reducing collagen deposition at an early stage (25). Moreover, it was also reported that umbilical cord-derived MSCs can ameliorate hyperoxia-induced lung injury in newborn rats with bronchopulmonary dysplasia and improve the bacterial clearance and survival in neonatal septic rats (26, 27).

There are many advantages of MSCs as an attractive therapeutic candidate. Firstly, they express low levels of human leukocyte antigen (HLA)-associated proteins, which lead to the absence of immunosuppression by allogeneic MSCs as stimulators (28). Secondly, MSCs can be easily isolated from host tissue and expanded stably and rapidly in vitro, which enables sufficient quantities for clinical use (29). Thirdly, the homing capacity of MSCs to the injury sites contributes to their persistence in the target tissues (30). Moreover, MSCs can be trapped in the lungs by systemic administration, thus prolonging their persistence in the lungs, which may be of most benefit in lung disease treatment (30). Fourthly, the safety profile of MSCs was demonstrated by phase 1 clinical trials in patients with ARDS and sepsis (7–15). Finally, growing evidence shows that MSCs exert anti-inflammatory and anti-apoptotic effects in response to injury (31).

In COVID-19 patients who develop acute respiratory failure requiring intubation and mechanical ventilation, sepsis and sepsis shock can also develop. Sepsis is a systemic inflammatory response to infection that is characterized by the imbalance between pro-inflammation and anti-inflammation (48, 49). Recently, MSCs have been shown to be effective in modulating such processes, which may be applicable to COVID-19.

Characterized by the acute onset of non-cardiogenic pulmonary edema, hypoxemia, and diffuse alveolar–capillary membrane damage, ARDS is a common cause of respiratory failure in critically ill patients (97). Recently, a growing number of studies have suggested that MSCs may be a potential therapy for ARDS by improving lung permeability, increasing alveolar fluid clearance, and promoting lung epithelial and endothelial repair.

According to statistical data generated early in the COVID-19 pandemic, 14% of cases were classified as severe COVID-19 and 5% were determined as critical with organ failure (3). Severe pneumonia and ARDS have been described in the reported COVID-19 cases. In parallel with proper ventilation management, dexamethasone, antiviral agents, convalescent plasma, and IL-1/6 inhibitors, MSCs are also considered a promising candidate therapy for COVID-19-related ARDS (4, 119). It was reported that MSCs from the bone marrow, amniotic fluid, and adipose tissue were all resistant to SARS-CoV-2 infection due to the low expressions of angiotensin-converting enzyme 2 (ACE-2) and transmembrane protease serine subtype 2 (TMPRSS2) on the cell surface under both steady-state and inflammatory conditions (120, 121), which supported the potential applicability of MSCs for COVID-19 treatment ( Figure 1 ).

It was reported that MSCs were well tolerated in ARDS patients without severe adverse events in phase 1 and phase 2a clinical trials (13, 15). Even though adverse effects such as shivering, muscle contraction, and increased lactic acid dehydrogenase arose after the infusion of MSCs in a small minority of patients (122–124), multiple clinical studies have identified the tolerability and efficacy of MSCs in the treatment of COVID-19-related ARDS. Current clinical trials advise that it is safe to administer MSCs extracted from the umbilical cord, placenta, or menstrual blood to severely and critically ill COVID-19 patients (122–132), always intravenously, with one to three infusions and 10⁶–10⁸ cells per dose.

Leng et al. (128) reported the earliest clinical trial (ChiCTR2000029990) of MSC therapy against COVID-19. In one dose of 1 × 10⁶ cells/kg, clinical grade MSCs were intravenously transplanted to seven patients with COVID-19 pneumonia (one of the critical type, four of the severe type, and two of the common type), and three patients in the control group received placebo treatment. Without any observed adverse effects, the symptoms (high fever, weakness, and shortness of breath) and oxygen saturations significantly improved within 2–4 days after MSC transplantation in all patients. More importantly, it was suggested that MSCs were valuable in suppressing virus-induced cytokine storms. Compared with the placebo control group, the overactivated cytokine-secreting immune cells disappeared in 3–6 days and regulatory dendritic cells dramatically increased in the group that received MSC treatment. Meanwhile, the serum pro-inflammatory cytokine TNF-α was significantly decreased and the anti-inflammatory cytokine IL-10 was remarkably increased after MSC transplantation. Of note is that all seven patients were put forward for this pilot study after no improvements were observed under standard treatment, and thus, it is not surprising that patients in the placebo group showed negative results. Therefore, the functions of MSCs should be identified in a larger population.

In a multicenter, open-label, non-randomized, parallel controlled phase 1 exploratory trial in China (ChiCTR2000029606) (125), 44 severe or critical COVID-19 patients were enrolled. Among them, 26 patients received allogeneic, menstrual blood-derived MSC therapy (three infusions totaling 9 × 10⁷ MSCs, one infusion every other day), while 18 patients received only concomitant medication. After 1 month of follow-up, it was observed that the mortality between the two groups was significantly different (7.69% in the MSC group vs. 33.33% in the control group, p = 0.048). Compared with the control group, the symptoms of patients in the experimental group were quickly alleviated, with expiratory dyspnea significantly improving during MSC infusion on day 1 (p = 0.016), day 3 (p = 0.040), and day 5 (p = 0.031). Chest imaging results also showed improvements in the experimental group in the first month after MSC infusion. In addition, Lanzoni et al. (132) reported a double-blind, phase 1/2a, randomized controlled trial of UC-MSC therapy against COVID-19. A total of 24 patients who were diagnosed with COVID-19-related ARDS within 24 h were randomized 1:1 to either a UC-MSC treatment group (two intravenous infusions of 100 ± 20 × 10⁶ UC-MSCs, on days 0 and 3) or a control group (two infusions of vehicle solution). The results reflected that UC-MSC treatment resulted in significant reductions of inflammatory cytokines on day 6 and improved survival (91% vs. 42%, p = 0.015) during a 31-day follow-up. In another double-blind, multicenter, phase 1 randomized controlled trial (NCT04457609) (130) in 40 critically ill COVID-19 patients, 20 patients received a single intravenous infusion of 1 × 10⁶/kg UC-MSCs in 100 ml saline (0.9%) solution (SS) and 20 patients received 100 ml 0.9% SS as the control group. The median time from intensive care unit (ICU) enrollment to MSC treatment was 8 days (range = 2–30 days). During 30 days of observation after MSC administration, the results revealed that UC-MSC treatment was significantly associated with a higher survival rate, with particular benefit in patients with comorbidities. In conclusion, MSC therapy is a safe approach against COVID-19 disease and shows promising benefits for severely ill COVID-19 patients at a progression stage. However, the observation periods of these studies were short and the sample size was small; thus, whether MSC therapy has long-term benefits warrants further research.

In a study (NCT04288102) by Shi et al. (131), a randomized, double-blind, placebo-controlled phase 2 trial was performed in 100 severe COVID-19 patients who were randomly assigned to receive either UC-MSCs (4 × 10⁷ cells per infusion) or placebo on days 0, 3, and 6. According to the analyses of both radiologists and lung imaging artificial intelligence software, the results of high-resolution chest CT imaging revealed that the proportions of solid component lesion volumes were significantly reduced in the MSC treatment group, but not in the placebo group, after 28 days of follow-up. The results of a 6-min walk test also indicated better restoration in patients treated with UC-MSCs. However, with a median time from symptom onset to study baseline of approximately 45 days, most of these patients were at the convalescent stage when joining this trial, which might account for the lack of significant reduction in the duration of oxygen therapy, dyspnea scores, cytokine levels, or chemokine levels. Nonetheless, COVID-19 patients might benefit from MSC therapy, even in the convalescent stage.

Several clinical studies were performed to assess the safety and effectiveness of MSCs in COVID-19 disease ( Table 1 ). However, these preclinical trials seemed to be less convincing due to the small sample sizes ranging from 5 to 100. Because of the emergency nature of the COVID-19 outbreak and the ethical limitations, some of these studies failed to strictly follow the principles of standard clinical trials, such as randomization, blindness, and comparisons. In addition, the patients enrolled in these clinical studies were at different stages of severe COVID-19 disease, which resulted in greater heterogeneity and different results. In particular, given that intravascular coagulation and thromboembolism are considered as leading causes of fatality in COVID-19, the application of MSCs remains controversial because variable levels of highly procoagulant tissue factor (TF/CD142) are expressed by MSCs (133). Furthermore, the time to starting MSC treatment, the cell dosage, and the interval duration have not been fully determined. Hence, further high-quality randomized clinical trials are needed to provide evidence regarding the application of MSCs in COVID-19-related ARDS, and hundreds of clinical trials registered on (https://clinicaltrials.gov/) may offer assistance with this.

The homing capacity and the systemic distribution of MSCs contribute to the persistence of MSCs in the lungs (26). Moreover, interesting work has been undertaken on therapeutic factors, such as Ang-1, KGF, and HGF, in ARDS and sepsis. However, MSCs alone or “protected mediators” alone may not be sufficient to rectify the disorder of the microenvironment, particularly in severe ARDS and septic shock. Therefore, there is a strong rationale for gene-modified MSCs with critical molecules to “rescue” lung injury and sepsis. It was reported that Ang-1-transfected MSCs could further improve both alveolar inflammation and permeability in septic mice (134, 135). In addition, ROR2-overexpressing MSCs led to more significant effects than did the native MSC treatment in ARDS mice, including the retention of MSCs in the lung, differentiation into alveolar type 2 (ATII) cells, and promotion of lung repair (113). Moreover, gene-modified MSCs also showed further benefits in LPS-induced sepsis and ALI treatment than did native MSCs alone, including overexpressing β-catenin, KGF, HGF, IL-10, the E-prostanoid 2 receptor, CXCR4 receptor, soluble IL-1 receptor-like-1 (sST2), ACE2, TGFβ1, CXCR7, LL-37, heme oxygenase-1, angiotensin II type 2 receptor (AT2R), Miro1, and 7ND (dominant-negative inhibitor of CCL2) (114, 115, 136–148) ( Table 2 ). Aside from the above-mentioned coding genes, the modification of non-coding genes, microRNAs, is an alternative approach. For instance, Wei et al. (150) reported that miR-377-3p released by UC-MSCs suppressed the expression of the target gene RPTOR, and RPTOR silencing played a pivotal role in enhancing autophagy-related proteins. Administration of UC-MSCs with overexpression of miR-377-3p activated the process of autophagy in alveolar epithelial cells under LPS treatment and finally dramatically ameliorated inflammation both in vitro and in vivo. Besides, miR-124-3p transferred by MSC-derived exosomes was observed to improve oxidative stress injury and suppress inflammatory response, thus was beneficial to attenuate traumatic ALI in rats (151). It was also reported that the transduction of miRNA let-7d (with anti-fibrotic effects) promoted recovery from bleomycin-associated lung injury (149). However, there are no reports on which genes should be overexpressed or knocked out in different models of sepsis and ARDS, and it is not clear whether double or triple knockouts of specific genes overexpressed in MSCs are better than a single gene. There is also no report of whether genetically modified MSCs are safe or effective for use in the human body. Therefore, further preclinical studies are required to elucidate the exact role of gene-modified MSCs in different microenvironments.

Despite the promising effects of MSCs in both preclinical and phase 1 clinical trials, the adverse effects of MSC treatment should not be ignored. MSCs are highly adherent to membrane oxygenators, which led to a rapid decline in oxygenator performance during extracorporeal membrane oxygenation (ECMO) in a sheep model of ALI (152, 153). In addition, human MSC administration showed a lack of efficacy in influenza-mediated lung injury in mice (64) and in improving the survival of mice with staphylococcal toxic shock syndrome (154). Senescent MSCs were also unable to show protection in a murine model of LPS-induced lethal endotoxemia (155). In some reports, MSCs were indicated to enhance neutrophil recruitment and its bactericidal functions through TNF-α and IL-1β signaling (156, 157). However, these findings highlight MSCs as a potential risk factor because aberrant neutrophil activation contributes to disease severity and local tissue damage in COVID-19 through the amplification of hypercytokinemia and the formation of neutrophil extracellular traps (NETs) (158). Furthermore, the encouraging findings from animal experiments do not guarantee their efficacy in clinical trials. It was reported that biomarkers of inflammation (IL-6 and IL-8) at 5 days and the length of hospital stay, ventilator-free days, and ICU-free days at 28 days were similar between the MSC treatment and saline placebo groups in a randomized controlled trial of MSCs for ARDS (14).

Clinical trials of MSCs for ARDS and COVID-19 treatment demonstrated safety and tolerance (see Table 1 ). However, both published trials contained very few patients, and much work remains to be undertaken before MSCs can be used in the clinic and even for COVID-19. Firstly, there are various sources of MSCs (including tissues and donors), and the best source needs to be identified. Secondly, the preparation of MSCs (including the culture conditions, passage, and storage) should be standardized and consistent. Thirdly, the best MSC dose, timing, and delivery remain to be further clarified. Finally, the etiologies of sepsis and ARDS vary; thus, optimal treatment with MSCs may depend on different microenvironments. Future studies should take the above suggestions into full consideration and carefully investigate them.

A growing number of preclinical studies have shown encouraging findings for MSC-based therapy in sepsis and ARDS treatment. Phase 1 clinical trials have also reported that MSCs are safe and well tolerated in ARDS patients. Nevertheless, experimental studies and clinical trials are still needed to fully understand the optimal MSC application in sepsis/ARDS and COVID-19.

ZX, YH, and JZ prepared the original draft and reviewed and edited the manuscript. WH, XL, YL, NZ, and LS reviewed and edited the manuscript. XD prepared the tables and figures. All authors contributed to the article and approved the submitted version.

The study was funded by the National Natural Science Foundation of China (81870069, 81970071, and 82070084), the Natural Science Foundation of Guangdong Province (2020A1515011459 and 2021A1515012565), the Science and Technology Program of Guangzhou (202102010157), the State Key Laboratory of Respiratory Disease Independent Program (SKLRD-Z-202108), Emergency Key Program of Guangzhou Laboratory (EKPG21-17) and Natural Science Foundation of Guangdong Province (2020A1515011459).

YL is an editor of Frontiers in Immunology.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.