Dimethyl sulfoxide in cryopreserved mesenchymal stromal cell therapy products: is there a safety risk to patients?

Abstract

Dimethyl sulfoxide (DMSO) is the preferred cryoprotectant for the cryopreservation of mesenchymal stromal cells (MSCs). As DMSO has been associated with in-vivo toxicity, its potential side effects when administered with MSC therapies are a matter of debate. To contribute to the assessment of the potential patient safety risk that may be posed by typical amounts of DMSO delivered with cryopreserved MSC-based therapy products, published safety data from intravenous and topical applications of DMSO and DMSO-containing MSC products in humans were reviewed. For the intravenous route, 1173 patients treated with 1–24 DMSO-containing MSC infusions were analyzed. For the topical route, for which data from the administration of DMSO-containing MSC products are lacking, the risk of local toxicity was estimated on the basis of the available information from the topical use of DMSO for wound healing purposes, whereas the risk of systemic toxicity was estimated on the basis of a worst-case scenario assuming 100% transdermal absorption. The doses of DMSO delivered via intravenous administration of MSC products were 2.5–30 times lower than the dose of 1 g DMSO/kg typically accepted for hematopoietic stem cell transplantation, and with adequate premedication, only isolated infusion-related reactions, if any, were reported. Published experience with the application of DMSO to skin wounds suggests that DMSO concentrations applied with undiluted DMSO-cryopreserved MSC products are unlikely to cause significant local adverse effects. In the worst-case scenario, assuming complete systemic absorption of DMSO from an MSC product applied to a large wound in a lightweight patient, the systemic exposure to DMSO would be approximately 55 times lower than that from an intravenous dose of 1 g/kg. In conclusion, the available data do not indicate significant safety concerns with the DMSO contained in intravenous or topical MSC products cryopreserved according to current standard protocols.

Background

The immunomodulatory and regenerative capacities of mesenchymal stem cells (MSCs) make them an interesting resource for the treatment of a broad spectrum of currently incurable diseases [1, 2]. MSC preparations for clinical use can be freshly harvested from cell culture immediately prior to administration to the patient, or the cells can be cryopreserved and thawed prior to administration. Although there are indications that immediately thawed MSCs may have lower blood compatibility and functional properties compared to freshly harvested MSCs [3, 4], the use of cryopreserved cells has important advantages: while fresh cells are viable for only several hours to a few days after harvest, cryopreservation allows sufficient time for rigorous quality control testing, enables long-term storage and off-the-shelf availability, ensures that multiple doses can be delivered with minimal inter-batch variability from large-scale cell cultures, and extends the geographic reach of viable cells by creating a larger window of time in which the cells can be shipped from the manufacturing site to the site of clinical use [5, 6].

The cryopreservation of MSCs requires the addition of cryoprotectants to prevent freezing-induced cell damage to ensure acceptable cell recovery, viability and biological function after thawing. Backed by the long-standing use of DMSO in hematopoietic stem cell transfusions, the conventional method of slowly cooling the cells in the presence of 10% (v/v) dimethyl sulfoxide (DMSO) has also become the preferred method for cryopreserving MSC-based products [7, 8]. Although several DMSO-free cryopreservation strategies and protocols have been developed and tested for their suitability and efficiency for the cryopreservation of human MSCs [9–34] (Table 1), none of these approaches has yet been shown to be suitable for clinical application.

Table 1.

Strategies for the DMSO-free cryopreservation of human MSCs

Freezing technique	Strategy	Method	CPA(s)	MSC source	Post-thaw outcome		Reference
					Viability	Recovery
Slow freezing	CPAs other than DMSO	Amino acids	1–10% ectoine	BM (hTERT+)	≤ 72%		[9]
			10% ectoine + 1% proline	BM (hTERT+)	87%		[10]
		Penetrating CPAs (sugar alcohols) +/- small-molecule additives	10% glycerol	Cornea		~ 70%	[11]
			10% glycerol	UC	52%		[12]
			10% ethylene glycol	UC	74%		[12]
			1.25% glycerol + 2 mM creatine	ESC		< 10% DMSO1)	[13]
		Nonpenetrating CPAs (sugars) + penetrating CPAs (sugar alcohols) +/- small-molecule additives	3% trehalose + 5% dextran 40 + 4% polyethylene glycol	AT	~ 95%	~ 95%	[14]
			1 M trehalose + 10% glycerol	AT	77%		[15]
			300 mM trehalose + 10% glycerol + 0.001% ectoine	ESC	92%	88%	[16]
			0.2 M sucrose + 10% glycerol	UC	70%		[12]
			0.2 M sucrose + 10% ethylene glycol	UC	75%		[12]
			150 mM sucrose + 300 mM ethylene glycol + 30 mM alanine + 0.5 mM taurine + 0.02% ectoine	ESC	96%	103%	[16]
			0.05 M glucose + 0.05 M sucrose + 1.5 M ethylene glycol	UC	72%		[17]
			Glucose or sucrose + mannitol or glycerol + creatine (various protocols)	ESC		≥ 10% DMSO1) (at optimal composition)	[18]
			150 mM sucrose + 62.5 mM mannitol + 6.25 mM creatine	ESC		< 10% DMSO1)	[13]
			30 mM sucrose + 5% glycerol + 7.5 mM isoleucine	ESC		~ 10% DMSO1)	[13]
			30 mM sucrose + 5% glycerol + 7.5 mM isoleucine	BM, AT	83%	93%	[19]
		Polymers	10% polyvinyl pyrrolidone	UC	63%		[17]
			7.5% carboxylated poly-l-lysine	BM	> 90%		[20]
			Poly(ethylene glycol)-poly(l-alanine) block copolymer (PEG 5000-PA 500)	Tonsil		87% of the rate achieved with 10% DMSO1)	[21]
	Facilitation of intracellular uptake of nonpenetrating CPAs	Endocytic uptake enabled by 24-hour pre-incubation	300 mM sucrose	Skin	49%	50%	[22]
			300 mM trehalose		49%	50%
			300 mM raffinose		48%	50%
			200 mM sucrose	AT	58%	55%	[23]
		Endocytic uptake enabled by encapsulating the CPA in nanoparticles	Trehalose encapsulated in pH-responsive genipin-crosslinked Pluronic F127-chitosan nanoparticles	AT	88–91%		[24]
			Trehalose encapsulated in cold-responsive poly(N-isopropylacrylamide-co-butyl acrylate) nanoparticles	AT	83%		[25]
		Electroporation-assisted delivery	400 mM sucrose	UC	81%		[26]
			400 mM trehalose	UC	83%		[26]
			400 mM raffinose	UC	89%		[26]
			400 mM trehalose	AT	< 10% DMSO1)	< 10% DMSO1)	[27]
			250 mM trehalose	AT	~ 10% DMSO1)	84%	[27]
			250 mM trehalose	AT	72%		[28]
			250 mM trehalose	UC	83%
			Zwitterionic CPA: 2–10% betaine	UC	83% (at 8% betaine)		[29]
		Ultrasound-facilitated delivery	5% trehalose	BM (hTERT+)	~ 10% DMSO1)		[30]
Vitrification	Encapsulation of cells to inhibit devitrification during thawing	Encapsulation of cells in alginate hydrogel microcapsules	0.5 M trehalose + 1.5 M 1,2-propanediol (for 2.5-µl microcapillaries)	AT	~ 90%		[31]
			1.3 M trehalose + 2 M 1,2-propanediol (for 250-µl straws)	AT	> 80%		[31]
	Nanotechnology-amplified rewarming to minimize devitrification during thawing	Fe3O4 nanoparticle-mediated magnetic induction rewarming	0.5 M trehalose + 2.0 M 1,2-propanediol + 2.0 M ethylene glycol	UCB		72%	[32]
		Near-infrared laser-mediated rewarming using Pluronic F127-liquid metal nanoparticles	0.5 M trehalose + 2.0 M 1,2-propanediol + 2.0 M ethylene glycol	BM	78%		[33]
	Encapsulation of cells + nanotechnology-amplified rewarming	Encapsulation of cells in alginate hydrogel microcapsules + Fe3O4 nanoparticle -mediated magnetic induction rewarming	1 M trehalose + 1 M ethylene glycol + 1.5 M propylene glycol	UCB	~ 84%		[34]

AT, adipose tissue; BM, bone marrow; CPA, cryoprotective agent; ESC, embryonic stem cell-derived; hTERT⁺, human telomerase reverse transcriptase-immortalized; UC, umbilical cord; UCB, umbilical cord blood

¹Refers to a comparison of the tested protocol with a standard cryopreservation protocol using 10% DMSO as the CPA

Unfortunately, DMSO may have unwanted side effects in patients receiving cell therapy [35], which are tolerated due to a lack of suitable alternatives. Because the frequency and severity of DMSO-associated adverse effects are related to the amount of DMSO administered [35], strategies have been developed to remove DMSO from the cell therapy product by repeated cycles of washing and centrifugation (or alternative, mechanical force-reducing methods such as filtration) between thawing and administration to the patient [7]. However, post-thaw washing methods to remove DMSO from a thawed cell therapy product are labor- and/or equipment-intensive. In addition, post-thaw manipulation poses a risk of cell damage and loss, which can affect the product’s functionality and introduce product variability [36]. For these reasons, such procedures should be performed at the manufacturer’s facility [37]; thus, the cells would need to be delivered in a thawed state. Owing to the increased logistical challenges associated with the shipping of thawed cells, this would pose a significant barrier to the use of cell therapies. Therefore, the vast majority of human clinical trials use cryopreserved MSCs that are thawed immediately prior to use [38] and are administered to patients in a suspension containing the cryoprotectant. Nevertheless, although MSC therapies have appeared safe in clinical trials [39, 40], the potential toxicity of DMSO in patients treated with MSCs remains a topic of debate [7].

Depending on the clinical indication, MSCs are administered systemically via intravascular infusion or locally via direct injection into the target tissue/organ or topical application [41]. Consequently, toxicity from DMSO delivered with MSC products could result from direct systemic exposure, local reactions, and indirect systemic exposure following absorption after local administration. To assess the potential patient safety risk that may be posed by typical amounts of DMSO delivered with cryopreserved MSC-based therapy products, we analyzed and evaluated published safety data from the intravenous and topical use of DMSO for therapeutic purposes and from the application of DMSO-containing cell products in humans. This evaluation and the underlying data may be useful to developers of MSC therapy products when designing cryomedia and formulations, as well as in discussions with regulatory authorities.

Potential toxicity of DMSO from intravenous MSC therapy

Following intravenous administration, DMSO is rapidly (< 10 min) distributed to various organs and tissues [42]. DMSO is metabolized by oxidation to its main metabolite dimethyl sulfone or reduced to dimethyl sulfide [43]. Unchanged DMSO and dimethyl sulfone are excreted in the urine [43], with dimethyl sulfone being the predominant urinary metabolite in humans [44]. Dimethyl sulfide is eliminated through breath, which can be associated with a characteristic “garlic-like” odor [44, 45]. After the intravenous injection of 2.0 g of DMSO in a 50% DMSO solution into humans, the elimination half-life was 4 days, and 80% was eliminated in the urine within 1 week [45].

In addition to the use of DMSO as a cryoprotectant, several attempts have been made in the past to exploit the multiple pharmacological modes of action of DMSO, including free radical scavenging, modulation of innate and adaptive immune functions, and inhibition of platelet activation [46, 47], for potential therapeutic purposes. Reports of these efforts published in the literature include a few small, mostly uncontrolled, pilot-style studies of intravenous DMSO infusion for the palliative treatment of cancer pain [48, 49], treatment of ischemic stroke [50], and hyperosmolar therapy for the control of refractory intracranial hypertension of various etiologies [51–54]. In these settings, repeated DMSO infusions at doses of up to 1.12 g/kg were reported to be associated with no or occasional side effects, including transient mild headache or moderate chills (Table 2). In addition to the dose, the concentration of DMSO in the infusion solution was important for the tolerability and safety of a DMSO infusion (Table 2). For example, following the administration of a 40% (v/v) DMSO solution, hematological disturbances including hemolysis, hemoglobinuria and prolonged bleeding time have been reported, which were not observed when the DMSO concentration in the infusion solution was reduced to 10% (v/v) [53]. In another study, even the infusion of a 28% (v/v) DMSO solution was not associated with any adverse effects, at least at a comparatively low DMSO dose of 0.56 g/kg [50].

Table 2.

Adverse effects of DMSO reported from therapeutic use of intravenous DMSO

Type of study	Indication	N	DMSO concentration (v/v)	DMSO dose per application	Application frequency	Duration of treatment	Adverse effects	Reference
Prospective, open-label, multidose	Refractory cancer pain	12	3.8%	22 g/patient1)	1× daily	8 cycles of 10 days with 2 days break between cycles	Mild headache (8 events), moderate chilling (4 events); all resolved within 2 h	[48]
		10	7.4%	44 g/patient2)
		4	7.4–10.7%	44–66 g/patient3)
Prospective, open-label, multidose	Pain relief in metastatic prostate cancer	7	8.8%	27.5 g/patient4)	1× daily	12 cycles of 5 days with 2 days break between cycles	Mild headache (9 events), moderate chilling (5 events); all resolved within 1–2 h	[49]
		11	7.3%	44 g/patient2)
		2	7.3–10.7%	44–66 g/patient3)
Prospective, uncontrolled	Intracranial hypertension due to severe closed head injury	10	n.s.	55 g/patient5)	As needed	1–8 days	Non cardiovascular effects	[51]
Controlled	Ischemic stroke	11	28%	0.56 g/kg	2× daily	12 days on average	None	[50]
Uncontrolled	Intracranial hypertension of diverse etiologies	2	10%	1.0 g/kg	As needed (up to 8 g/kg/day)	As needed	Hypernatremia	[52]
		4	20%	Titrated against the intracranial pressure
Uncontrolled	Intracranial hypertension of diverse etiologies	6	10%	1.0 g/kg	Every 6 h	1–7 infusions	None	[53]
		5	40%				Hemolysis, hemoglobinuria, bleeding time prolonged
Uncontrolled	Intracranial hypertension due to severe closed head injury	10	28%	1.12 g/kg	Every 6 h	1–10 days	No serious side effects	[54]

n.s., not specified

¹Reported as 20 ml/patient

²Reported as 40 ml/patient

³Reported as 40–60 ml/patient

⁴Reported as 25 ml/patient

⁵Reported as 50 ml/patient

More extensive experience with the intravenous administration of DMSO to humans has resulted from patients undergoing hematopoietic stem cell (HSC) transplantation. Infusions of DMSO-cryopreserved HSCs have been associated with a number of infusion-related adverse reactions, most of which have been attributed to DMSO-induced histamine release, with chills, gastrointestinal (nausea, vomiting, abdominal pain), cardiopulmonary (hypo- or hypertension, brady- or tachycardia, cough, dyspnea), and neurologic (amnesia, seizures, cerebral infarction) reactions being among the most commonly reported [55]. However, in the setting of HSC transplantation, it is difficult to isolate the effects of DMSO from side effects related to conditioning with high-dose chemotherapy and/or total body irradiation, as well as side effects caused by potential confounding factors such as the amount of non-mononuclear blood cells or cellular debris contained in the stem cell transplant [56–58]. In general, and in accordance with the above findings from the therapeutic use of DMSO, a maximum dose of 1 ml or 1 g of DMSO per kg body weight per infusion is considered acceptable for HSC transplantation [59]. A maximum DMSO dose of 1 ml or 1 g/kg body weight per HSC infusion is also advocated by relevant scientific societies, including the European Society for Blood and Marrow Transplantation (EBMT) [60] and the Association for the Advancement of Blood & Biotherapies (AABB) in collaboration with, among others, the International Society for Cell and Gene Therapy (ISCT) [61], as well as regulatory authorities, e.g., the Federal Institute for Vaccines and Biomedicines (Paul Ehrlich Institute) in Germany [62].

Unlike DMSO dosing, to our knowledge there are no consensus or regulatory recommendations regarding the rate at which DMSO should be infused. In addition, recommendations regarding the infusion of DMSO-containing HSC preparations are, naturally, based not only on the potential toxicity of DMSO, but also on other potential adverse reactions associated with HSC infusion including immunologic complications or transfusion-related circulatory overload [59, 61, 62]. On the other hand, there is an incentive to infuse DMSO-containing HSC products as quickly as possible to minimize the contact time between the thawed cells and DMSO, since DMSO can adversely affect the viability and functionality of the thawed cells [59, 62]. In general, it is recommended to start the infusion slowly to recognize any signs or symptoms that may indicate complications and to increase the infusion rate as tolerated by the patient [61, 62].

Due to the aforementioned challenges of distinguishing the potential adverse effects of DMSO from the toxicity of the HSC transplantation procedure and the underlying disease, observations from HSC transplantations cannot be directly applied to therapeutic MSC infusions. Furthermore, since MSCs are used to treat a much wider range of conditions with varying degrees of severity, potential adverse effects from DMSO may be more relevant in patients receiving MSC therapy than for those undergoing HSC transplantation. Especially in patients with less severe conditions, adverse effects from DMSO will more strongly impact the risk-benefit balance of the therapy than in patients receiving life-saving treatments such as HSC transplantations. For these reasons, to assess the potential toxicity of DMSO in the context of MSC therapies, it is more meaningful and highly warranted to analyze studies of DMSO-cryopreserved MSC products. However, in the field of MSC-based products, there are no systematic analyses of the amount of DMSO administered to patients or the rate at which the DMSO was infused. Although many clinical trials have used cryopreserved MSCs or mesenchymal precursor cells (MPCs) that were thawed immediately prior to use, information on the composition of the cryopreservation medium and reconstitution of the product is often underreported in clinical trial reports [63]. Evaluable information could be derived from clinical trials or case reports of intravenous use of approved products for which information on DMSO content and instructions for product reconstitution are publicly available, and from studies of investigational products that have provided this information [64–82]. On the basis of these data, a total of 1116 patients (448 children and 668 adults) treated with 1–24 infusions of DMSO-containing MSC or MPC products or a DMSO-containing placebo were analyzed (Table 3).

Table 3.

Treatment-related adverse events (AEs) with DMSO-containing MSC/MPC products or a DMSO-containing placebo

Product (Indication)	Cell dose [×106/kg]	DMSO				Premedication	Number of patients Total (Children: Adults)	Number of doses		Treatment-related AEs	Refer-ence
		Concentration		Dose per infusion [mg/kg]	Infusion rate [g/min]			Total	Per patient Range (Median)
		Before recon-stitution	After recon-stitution
Remestemcel-L (aGVHD)	2	10% (v/v)	3.75% (v/v)	33	0.16–0.241)	Hydrocortisone + diphenhydramine	16 (0:16)2)	322)	2	None	[64]
			3.75% (v/v)		0.16–0.241)	Hydrocortisone + diphenhydramine	10 (10:0)3)	993)	3–20 (8.5)	None	[65]
			3.75% (v/v)		0.16–0.241,4)	Hydrocortisone + diphendydramine4)	241 (241:0)	n.s.	1–24 (11)	12 serious AEs possibly related, including 2 infusion reactions (resolved without sequelae)	[66]
			n.d.		n.d.	Hydrocortisone + diphendydramine5)	54 (54:0)	535	n.s.	16 AEs in 9 patients possibly related (6 of them serious), including 3 infusion reactions	[67]
			n.d.		n.d.	Hydrocortisone + diphenhydramine	244 (27:217)	n.s.	1-166)	5 infusion reactions	[68]
	8	10% (v/v)	3.75% (v/v)	132	0.16–0.241)	Hydrocortisone + diphenhydramine	15 (0:15)7)	307)	2	None	[64]
						Hydrocortisone + diphenhydramine	2 (2:0)8)	238)	2–21 (11.5)	None	[65]
Remestemcel-L (other indications9))	2	10% (v/v)	3.75% (v/v)10)	33	n.d.	Hydrocortisone + diphenhydramine11)	2 (2:0)	4	2	None	[69]
	2			33	0.16–0.24	Hydrocortisone + diphenhydramine	11 (0:11)	22	2	None	[70]
	2			33	0.16–0.24	n.s.	110 (0:110)	21712)	1–2 (2)12)	No infusion reactions	[71]
	0.5–5			8–80	0.08	n.s.	53 (0:53)	53	1	None	[72]
Temcell® (aGvHD)	2	10% (w/v)13)	3.75% (w/v)	30	0.1514)	Hydrocortisone and / or chlorpheniramine	14 (1:13)	121	3–12 (8)	No infusion reactions	[73]
						Hydrocortisone and / or chlorpheniramine	25 (5:20)	231	4–16 (8)	No infusion reactions	[74]
						n.s.	12 (12:0)	> 12815)	> 4–12 (12)15)	No infusion reactions, no treatment-related AEs	[75]
MSC-FFM (aGvHD)	0.6–4.5	10% (v/v)	10% (v/v)	Up to 40016)	Up to 0.3317)	n.s.	92 (61:31)	280	1–9 (3)	No infusion-related reactions	[76]
	1–2			37-22018)		n.s.	156 (33:123)	n.s.	(4)	5 AEs related to treatment or DMSO-containing placebo19)	[77]
Rexlemestrocel-L (Diabetic kidney disease)	≈ 1.520)	7.5% (v/v)	0.28% (v/v)	3.3	0.007	n.s.	10 (0:10)	10	1	No treatment-related AEs, no infusion reactions	[78]
	≈ 320)		0.56% (v/v)	6.6	0.015		10 (0:10)	10	1	No treatment-related AEs, no infusion reactions	[78]
Rexlemestrocel-L (Type 2 diabetes)	0.3	7.5% (v/v)21)	n.d.	0.66	n.d.	n.s.	15 (0:15)	15	1	No treatment-related acute AEs	[79]
	1			2.2			15 (0:15)	15	1		[79]
	2			4.4			15 (0:15)	15	1		[79]
Orbcel-C™ (ARDS)	≈ 122)	5% (v/v)	0.25% (v/v)	5.922)	0.006–0.018	Chlorphenamine	3 (0:3)	3	1	1 mild infusion-related pyrexia	[80]
	≈ 222)		0.5% (v/v)	11.822)	0.012–0.036	Chlorphenamine	3 (0:3)	3	1	1 mild infusion-related pyrexia	[80]
	≈ 422)		1% (v/v)	23.722)	0.024–0.073	Chlorphenamine	3 (0:3)	3	1	No infusion reactions	[80]
	≈ 422)		1% (v/v)	23.722)	0.024–0.073	Chlorphenamine	30 (0:30)	30	1	1 mild infusion-related pyrexia	[81]
Orbcel-M™ (Diabetic kidney disease)	0.823)	10% (v/v)	10%	4423)	0.22–0.44	Acetaminophen + hydrocortisone + chlorphenamine	12 (0:12)	12	1	1 moderate bronchospasm shortly after DMSO-containing placebo infusion (resolved without sequelae)	[82]

n.d., not derivable from reported data; n.s., not specified

¹For patients ≥ 35 kg. In patients < 35 kg, a fixed minimum infusion time of 60 min reduces the infusion rate proportional to the body weight

²A total of 31 patients were treated in this trial, with 16 patients receiving 2 cell doses according to the label

³Of a total 12 patients treated in this trial, 10 patients received 3–20 cell doses according to the label

⁴According to a preliminary report [85]

⁵According to the study protocol [130]

⁶15, 83, 59, and 6 patients received ≤ 4, 5–8, 9–12, and 13–16 remestemcel-L infusions, respectively, and 13, 42, 25, and 1 patient received ≤ 4, 5–8, 9–12, and 13–16 placebo infusions

⁷Of a total of 31 patients treated in this trial, with 15 patients receiving 2 infusions of 4× the cell dose as per label, delivering DMSO doses of 132 mg/kg

⁸Of a total 12 patients treated in this trial, 2 patients received 2 and 21 infusions, respectively, of 4× the cell dose as per label, delivering DMSO doses of 132 mg/kg

⁹Including COVID-19-associated multisystem inflammatory syndrome in children [69], COVID-19-associated ARDS [70, 71], and acute myocardial infarction [72]

¹⁰Assuming that the product was reconstituted according to the label

¹¹According to the registry entry on clinicaltrials.gov [131]

¹²Of a total 110 patients, 107 received two infusions, and 3 received one infusion

¹³Specified as 1.5 g DMSO/15 ml or 1.08 g DMSO/10.8 ml

¹⁴For patients > 50 kg. In patients ≤ 35 kg, a fixed minimum infusion time of 10 min reduces the infusion rate proportional to the body weight

¹⁵Four patients received additional doses, the number of which is not reported in the publication

¹⁶When administered according to the label, DMSO doses of 37–220 mg/kg are delivered. The actual DMSO range was wider, as patients were reported to have received a wider range of cell doses (0.6–4.5 × 10⁶ cells/kg) than specified on the label (1–2 × 10⁶ cells/kg). A preliminary report of 69 patients [89] reported maximum DMSO doses of 300–400 mg/kg, which were administered to 3 children

¹⁷The infusion rate may be slowed depending on the patient’s cardiovascular status (to avoid cold-related bradycardia induced by the ice-cold infusion solution)

¹⁸According to the label, each patient (15–90 kg) receives one freezing bag (containing different amounts of MSCs) per infusion. Each freezing bag contains 3.3 g of DMSO, resulting in DMSO doses ranging from 37 mg/kg (for a 90-kg patient) to 220 mg/kg (for a 15-kg patient)

¹⁹Including chills, BK virus cystitis, C-reactive protein elevation (2 events) and nausea

²⁰On the basis of a mean patient body weight of ≈ 100 kg

²¹Specified in the supplementary material to the report of Packham et al. [78]

²²On the basis of a mean patient body weight of 93 kg, as reported in the trial by Gorman et al., 2023 [81]

²³Assuming a high (for safety reasons) patient body weight of 100 kg

Remestemcel-L

Remestemcel-L (Mesoblast Inc., New York, NY, USA) is a human bone marrow-derived MSC product approved in Canada as Prochymal^® under the Notice of Compliance with Conditions (conditional approval) policy and most recently fully approved in the U.S. as Ryoncil^® for the treatment of acute graft-versus-host disease (aGvHD) in pediatric patients refractory to corticosteroid therapy [83, 84]. Remestemcel-L consists of ex vivo cultured human MSCs cryopreserved at approximately 6.6 × 10⁶ cells/ml (Prochymal^®: 100 × 10⁶ cells/15 ml [83]; Ryoncil^®: 25 × 10⁶ cells/3.8 ml [84]) in a solution consisting of 70% (v/v) Plasma-Lyte A, 20% (v/v) of a 25% human serum albumin (HSA) solution and 10% (v/v) DMSO, which is equivalent to 0.11 g DMSO/ml (density of DMSO = 1.1 g/ml). The recommended dose of remestemcel-L is 2 × 10⁶ cells/kg body weight [83, 84]; therefore, the amount of DMSO infused with a single cell dose is 33 mg/kg body weight (contained in 0.3 ml/kg of the product). According to the label [83], Prochymal^® is diluted after thawing to a concentration of 2.5 × 10⁶ cells/ml by adding 25 ml of a diluent to 15 ml of the product [83]; the reconstituted solution then contains 3.75% (v/v) DMSO [64–66], which is equivalent to 0.04 g DMSO/ml. For patients ≥ 35 kg, the infusion rate is 4–6 ml/min [83], delivering 0.16–0.24 g DMSO/min. To patients < 35 kg, the total infusion volume is given over 60 min [83], which reduces the infusion rate and thus the rate of DMSO delivery proportional to body weight. Ryoncil^® is reconstituted by diluting the required volume of the product with 40 ml of a diluent [84]; therefore, the concentration of DMSO in the infusion solution and thus the rate of DMSO delivery vary depending on the patient’s body weight. To reduce the potential for infusion reactions, patients receiving remestemcel-L should be premedicated with hydrocortisone and antihistamines [83, 84].

The first two published clinical trials of remestemcel-L evaluated two different dose levels, a “low” dose of 2 × 10⁶ cells/kg (the later standard dose) and a high (fourfold) dose of 8 × 10⁶ cells/kg for the upfront treatment of newly diagnosed, untreated aGvHD in adults [64] and standard treatment-refractory aGvHD in children [65]. At both dose levels, the cell product was reconstituted to the same cell and DMSO concentration, so that patients receiving the high doses received four times the amount (132 mg/kg) of DMSO per infusion. In these trials, a total of 131 standard-dose infusions were administered to 26 patients (16 adults and 10 children), and a total of 53 high-dose infusions were administered to 17 patients (15 adults and 2 children). No infusion-related or other treatment-related toxicities were observed in any patient in either trial [64, 65].

In subsequent studies, including a multicenter expanded-access program [66] (preliminary report in [85]), a single-arm phase 3 registrational trial [67], and a randomized placebo-controlled phase 3 trial [68], children, adolescents and adults with treatment-refractory aGvHD were enrolled to receive eight standard-dose cell infusions (or a DMSO-containing placebo in the latter trial) given over 4 weeks. Patients with a partial or mixed response were eligible to receive 4 additional infusions given over 4 weeks, and patients who experienced an aGvHD flare after achieving a complete response were eligible to receive 8 additional infusions given over 4 weeks. In these three studies, a total of 539 patients received at least one infusion (median 11, mean 9.9, and 5–8 infusions in 50% of patients reported for the first, second and third study, respectively). From the expanded-access program (241 patients), a total of 12 serious adverse events were considered possibly related to remestemcel-L; however, it could not be excluded that patients may have responded to concomitant treatments other than remestemcel-L. These events included 2 infusion-related reactions in 1 patient, both of which resolved without sequelae [66]. In the single-arm phase 3 trial (54 patients who received a total of 535 infusions), a total of 16 adverse events in 9 patients, including 3 infusion reactions, were possibly related to remestemcel-L [67]. In the randomized placebo-controlled phase 3 trial (244 patients), 5 infusion-related reactions were reported, in 3 patients receiving remestemcel-L and in 2 patients receiving placebo [68].

An additional 123 patients were reported to have received a total of 243 infusions of remestemcel-L at label doses in an expanded-access program for the treatment of coronavirus disease 2019 (COVID-19)-associated multisystem inflammatory syndrome in children (4 infusions in 2 patients) [69], and a series of consecutive cases (22 infusions in 11 patients) and a randomized placebo-controlled trial (217 infusions in 110 patients) for the treatment of COVID-19-associated acute respiratory distress syndrome (ARDS) [70, 71], with no infusion-related reactions or treatment-related adverse events observed.

In addition, in a dose-escalation trial of remestemcel-L to treat acute myocardial infarction, 53 patients received a single dose of 0.5, 1.6, or 5 × 10⁶ cells/kg or (DMSO-containing) placebo at a rate of 2 ml/min [72], delivering DMSO doses of 8, 26, and 80 mg/kg (contained in 0.2, 0.64, and 2 ml/kg of the reconstituted product), respectively, at a rate of 0.08 g/min. No treatment-related adverse events were observed [72].

Temcell^®

Temcell^® HS Inj. (JCR Pharmaceuticals Co. Ltd, Ashiya, Japan) is a human bone marrow-derived MSC product approved in Japan for the treatment of aGvHD following HSC transplantation [86], which was developed through the in-licensing of the manufacturing technology for remestemcel-L from Osiris Therapeutics. Temcell® consists of ex vivo cultured, cryopreserved MSCs delivered in freezing bags at 72 × 10⁶ cells/10.8 ml containing 540 mg HSA and 1.08 g DMSO (commercial product) or at 100 × 10⁶ cells/15 ml containing 750 mg HSA and 1.5 g DMSO (investigational product JR-031) [73, 74, 86]. In both cases, this is equivalent to 0.1 g of DMSO/ml. Prior to use, the product is thawed and diluted with 18 ml (commercial product) or 25 ml (investigational product) of physiological saline [73, 86] to obtain a final infusion suspension containing 2.5 × 10⁶ cells/ml and 0.0375 g of DMSO/ml. Temcell® is infused at 2 × 10⁶ cells/kg body weight [86]; therefore, the amount of DMSO infused with a single cell dose is 30 mg/kg (contained in 0.8 ml of the reconstituted product). For patients > 50 kg, the infusion rate is 4 ml/min [86], delivering 0.15 g of DMSO/min. To patients ≤ 50 kg, the total infusion volume should be infused slowly over at least 10 min [86], which reduces the infusion rate and thus the rate of DMSO delivery proportional to body weight. To reduce the potential for infusion reactions, patients should be premedicated with either hydrocortisone or chlorpheniramine [86].

Two clinical trials of Temcell^® have been reported in the literature: a phase 1/2 trial in 1 child (4 years) and 13 adults (29–62 years) [73] and a phase 2/3 trial in 5 children and 20 adults (5–66 years) [74] with steroid-refractory aGvHD. Patients were enrolled to receive 8 cell infusions over 4 weeks. Patients with a partial or mixed response were eligible to receive 4 additional infusions given over 4 weeks. In addition, in the phase 2/3 trial, patients with a complete response and a subsequent flare at week 10 were eligible for retreatment with 8 cell infusions given over 4 weeks [86]. Furthermore, a retrospective analysis of 12 consecutively enrolled pediatric patients who received at least one cycle of 4–12 (median 12) infusions at a Japanese children’s medical center was published [75]. In these studies, a total of more than 480 infusions were administered to 51 patients, with no infusion-related reactions observed [73–75]. There is also a review of a large body of real-world evidence including 381 consecutive aGVHD patients treated with Temcell^® in the clinical setting during the first 3 years after its approval in 2016 [87]. However, this review analyzed only efficacy outcomes and did not include safety data.

MSC-FFM

MSC-FFM (Obnitix^®, medac GmbH, Wedel, Germany) is a human bone marrow-derived MSC product currently tested in a phase III clinical trial and marketed in Germany under the Hospital Exemption for the treatment of treatment-refractory aGvHD [88]. The product is formulated in ready-to-use freezing bags as Obnitix^® 30 (30 × 10⁶ cells at 1 × 10⁶ cells/ml), Obnitix^® 60 (60 × 10⁶ cells at 2 × 10⁶ cells/ml) or Obnitix^® 90 (90 × 10⁶ cells at 3 × 10⁶ cells/ml). The single infusion dose is 1 freezing bag Obnitix^® 30, Obnitix^® 60, or Obnitix^® 90 for patients weighing 15–30, 31–60, or 61–90 kg, respectively, resulting in cell doses of 1–2 × 10⁶ cells/kg. The cryomedium consists of 0.9% sodium chloride solution containing 5% HSA and 10% (v/v) DMSO, which is 3.3 g DMSO per freezing bag, resulting in DMSO doses ranging from 37 mg/kg (for a 90-kg patient) to 220 mg/kg (for a 15-kg patient). The minimum infusion time is 10 min, which is equivalent to 3 ml/min, delivering a maximum of 0.3 ml ≙ 0.33 g DMSO/min. The infusion rate may be slowed depending on the patient’s cardiovascular status (to avoid cold-related bradycardia induced by the ice-cold infusion solution). The recommended infusion schedule consists of four infusions given at weekly intervals. Premedication with dimetindene, acetaminophen, dihydrocodeine and/or prednisolone may be considered [88].

Published experience comes from post-marketing surveillance reports of 61 children and adolescents (5 months to 18 years) and 31 adults (18–66 years) treated with MSC-FFM for steroid-refractory aGvHD [76] and 33 children and adolescents (0–17 years) and 123 adults (19–79 years) treated for ruxolitinib-refractory aGvHD [77]. The former population (92 patients) received a total of 280 infusions of 0.6–4.5 × 10⁶ cells/kg. No infusion toxicity was observed [76]. In a preliminary report of 69 of these patients receiving a total of 212 doses, with the 3 smallest children (12–14 kg) receiving maximum DMSO doses of 300–400 mg, only 2 adverse events were noted: one case of nausea/vomiting, presumably due to DMSO, and one case of headache, presumably due to the ice-cold infusion solution [89]. In the ruxolitinib-refractory population (156 patients), who received a median of 4 infusions of 1–2 × 10⁶ cells/kg, 5 adverse drug reactions were reported in 3 adult patients: chills, BK virus cystitis, C-reactive protein elevation (2 events), and nausea. None of the reactions observed with MSC-FFM led to drug discontinuation or dose reduction [77].

Rexlemestrocel-L

Rexlemestrocel-L (Mesoblast Inc., New York, NY, USA) is a human bone marrow-derived stromal precursor antigen (STRO)-3⁺ MPC product that is in development for the treatment of several inflammatory diseases with unmet medical needs, including chronic low back pain associated with disc degeneration (intradiscal application) [90] and chronic heart failure (transendocardial application) [91].

For studies of intravenous administration, rexlemestrocel-L was provided in vials containing 150 × 10⁶ cells cryopreserved in 4 ml medium consisting of 50% alpha-minimum essential medium, 42.5% ProFreeze™ NAO freezing medium and 7.5% DMSO, which is 0.33 g DMSO per vial [78]. In a clinical dose-escalation study for the treatment of diabetic kidney disease, 10 adult patients each were treated with one dose of 150 × 10⁶ or 300 × 10⁶ cells per patient (mean body weight approximately 100 kg) [78], which delivered 0.33 g and 0.66 g of DMSO, respectively, corresponding to average DMSO doses of 3.3 and 6.6 mg/kg. Prior to administration, the thawed cells were diluted in 100 ml of normal saline and infused over 45 min [78], resulting in DMSO concentrations of 0.28% (v/v) and 0.56% (v/v) and DMSO infusion rates of 0.007 and 0.015 g/min for the low and high cell dose, respectively. No infusion-related reactions were observed, and none of the reported adverse events were considered related to treatment [78]. In a dose-escalation study of rexlemestrocel-L for the treatment of type 2 diabetes, 15 adult patients each received one dose of 0.3 × 10⁶, 1 × 10⁶, or 2 × 10⁶ cells/kg [79], delivering 0.66, 2.2, or 4.4 mg/kg DMSO, respectively. No acute adverse events were associated with the infusion [79].

Orbcel-C™

Orbcel-C™ (Orbsen Therapeutics, Galway, Ireland) is a human umbilical cord-derived, CD362-enriched MSC product in development for the treatment of immune-mediated inflammatory diseases such as primary sclerosing cholangitis, rheumatoid arthritis, lupus nephritis, Crohn’s disease, and ARDS [80, 81, 92]. For two reported trials in ARDS [80, 81], Orbcel-C was supplied in three fixed doses of 100, 200, and 400 × 10⁶ cells in 10, 20 and 40 ml, respectively, of a solution consisting of equal volumes of Plasma-Lyte 148 supplemented with 2.5% HSA and CryoStor^® CS10 medium (containing 10% DMSO), resulting in a final DMSO concentration of 5% (v/v) [81]. The thawed product was diluted to a final volume of 200 ml [80, 81]; the reconstituted infusion solution then contained 0.25% (v/v), 0.5% (v/v), or 1% (v/v) DMSO, resulting in 0.55, 1.1, and 2.2 g of DMSO, respectively, per 200 ml cell infusion. Assuming a mean trial participant weight of 93 kg [81], patients received 5.9, 11.8, or 23.7 mg DMSO/kg per cell dose. Infusions were administered over 30–90 min [80, 81], resulting in DMSO infusion rates of 0.006–0.018, 0.012–0.036, and 0.024–0.073 g/min. All treated patients were premedicated with chlorphenamine [80, 81]. In a phase 1 trial in ARDS [80], 3 patients each received one infusion at one of the three dose levels of Orbcel-C. A total of 4 mild adverse events possibly related to cell treatment were reported in these 9 patients, including 2 events of mild infusion-related pyrexia (1 each in the low-dose and the middle-dose cohorts) [80]. In a subsequent phase 2 trial, 30 patients with ARDS received an infusion at the highest dose level, with 1 event of mild infusion-related pyrexia reported [81].

Orbcel-M™

Orbcel-M™ (Orbsen Therapeutics, Galway, Ireland) is a human bone marrow-derived, anti-CD362 antibody-selected MSC product in development for the treatment of diabetic kidney disease [82]. In an ongoing clinical trial, 16 patients received one infusion of 80 × 10⁶ cells cryopreserved in 40 ml CryoStor^® CS10 (containing 10% DMSO) (n = 12) or 40 ml of CryoStor^® CS10 without cells (placebo) (n = 4) [82], delivering DMSO doses of 4.4 g per patient. The infusions were administered over 10–20 min, resulting in DMSO infusion rates of 0.22–0.44 g/min. All patients were premedicated with oral acetaminophen and intravenous hydrocortisone and chlorphenamine. One placebo-treated patient experienced an episode of moderate bronchospasm shortly after the infusion, which resolved completely after approximately 50 min with appropriate treatment [82].

Combining all the above studies, the majority of the total 1173 patients analyzed, namely 1069 patients (448 children and 621 adults), received 1–24 infusions of MSC/MPC products or an equivalent placebo delivered as a cryopreserved formulation containing DMSO at the most commonly used concentration of 10% (v/v) or 10% (w/v) (Table 3). After thawing, no washing steps were performed to remove the DMSO, so the entire amount of DMSO in the required volume of product was administered. The amount of DMSO administered by infusion was therefore dependent on the cell dose and the cell density of the frozen product. At typical MSC doses of 2 × 10⁶ cells/kg and a cell density of 6.7 × 10⁶ cells/ml (remestemcel-L [83], Temcell^® [73, 74, 86]), only 30–33 mg DMSO/kg was administered [64–71, 73–75], which is 30 times lower than the dose of 1 g/kg generally considered acceptable for HSC transfusions [59–62]. At higher MSC doses (up to 8 × 10⁶ cells/kg [64, 65]) and/or lower cell densities (1–3 × 10⁶ cells/ml, MSC-FFM [88]), higher amounts of DMSO were administered (up to 400 mg/kg [76]; 2.5 times lower than the dose of 1 g/kg administered with HSC transfusions). In all these settings, the incidence of treatment-related adverse events was very low. Only isolated infusion-related reactions were reported, and often none at all (Table 3).

In addition to the amount of DMSO administered, the tolerability of a DMSO-containing MSC therapy product may also depend on the rate at which DMSO is infused into the recipient. This rate can be slowed by diluting the cell suspension during reconstitution and/or reducing the infusion rate of the final infusion solution. In the above studies, only approximately a quarter of the patients (260/1069) treated with an MSC product cryopreserved with 10% DMSO received the thawed product without any further dilution step (MSC-FFM [76, 77], Orbcel-M™ [82]), whereas in approximately half of those patients (511/1069), the 10% DMSO concentration was diluted during reconstitution to 3.75% in the final infusion solution (Remestemcel-L [64–66, 69–72], Temcell^® [73–75]) [for the fourth quarter of patients (n = 298), the final DMSO concentration could not be determined from the available information (Table 3)]. When defining the infusion rate, it should be taken into account that depending on the contact time between DMSO and the thawed cells, DMSO may adversely affect the viability and functionality of the cells after thawing [93]. For example, in a post-thaw stability study of bone marrow-derived MSCs cryopreserved in three different media containing 10% DMSO, post-thaw viability decreased by 5.5–7.9% and the percentage of non-apoptotic non-necrotic (Annexin 5-negative/propidium iodide-negative) cells decreased from 90–94% to 76–81% during 6 h of storage at room temperature in the cryomedium, with both effects reduced when the cell suspension was diluted after thawing to reduce the DMSO concentration to 5% or 3.3% in the final formulation [94]. Clearly, these observations can only give a rough orientation. Because cell stability can vary between different MSC types and cryopreservation protocols, the optimal infusion rate must be determined specifically for each MSC product in its final, reconstituted form on the basis of thorough stability testing by the manufacturer to ensure that each patient will receive viable and functional cells with maximum safety.

Importantly, adverse events observed with a DMSO-containing MSC therapy product cannot necessarily be attributed to DMSO. Especially in patients with systemic conditions such as aGvHD, it can be difficult to distinguish the side effects of the product from the symptoms of the underlying disease. Furthermore, adverse effects may also be due to other components of the product. Although MSCs are generally considered safe [39, 40], side effects from the cells themselves cannot be completely ruled out. Cell therapy products containing DMSO may also contain other components that have been associated with infusion reactions. In particular, almost all of the MSC products used in the above studies were reported to contain HSA (Remestemcel-L [83, 84], Temcell^® [73, 74], MSC-FFM [88], or Orbcel-C™ [80, 81]), which has been associated with infusion-related reactions [95–97] that may be more common than previously thought [97]. Therefore, strategies to reduce or replace such components could also help maximize patient safety during treatment with cell therapy products, provided that the beneficial effects they have on cell stability and functionality can be achieved through other means.

Finally, it is important to note that for the approved products, premedication of the patient with corticosteroids and/or antihistamines is recommended or required [83, 86, 88]. Although several publications did not report whether patients were actually premedicated (Table 3), premedication appears to be common and may have contributed to the low incidence of adverse events associated with intravenous infusions of DMSO-containing MSC products.

Potential toxicity of DMSO from topical MSC therapy

For local application routes, the issue of potential toxicity of DMSO in MSC products is particularly relevant, as the suspension volumes feasible for local applications are typically too small to allow significant dilution of the cryopreservative [98]. Therefore, some authors have advised against direct local administration of thawed cell products containing DMSO, especially to poorly perfused tissue sites such as articular or discal cartilage and to neural tissues [7, 99].

In the field of skin wound healing, several promising attempts to treat chronic, refractory wounds of various etiologies by topical application of MSCs have been reported [100–107]; however, all of these studies appear to have been pilot studies or early-phase trials in which freshly isolated cells or cryopreserved cells that underwent DMSO-depleting washing steps after thawing were used. In contrast, to our knowledge, no studies have been published on the topical application of bedside-thawed MSC products in wound healing applications. Therefore, it was necessary to refer to reported topical applications of DMSO outside of MSC therapies. The risk assessment had to consider that toxicity from DMSO application to wounds can essentially occur in two ways: first, direct topical toxicity could cause unwanted local reactions and/or adversely affect wound healing, and second, systemic toxicity could result from transdermal absorption of the readily permeating DMSO molecule into the bloodstream.

Potential direct topical toxicity

For topical administration of cell doses of 1 × 10⁶–2 × 10⁶ cells per cm² wound area, which are generally considered effective in wound healing applications [102–105], even at high freezing densities of 1 × 10⁷ cells/ml, the thawed cell suspension must be administered without dilution, to achieve technically feasible application volumes of a few tens of milliliters (100 to 200 µl [104, 105]) per cm² wound area. This means that if a thawed cryopreserved cell product is applied directly, the concentration of DMSO applied to the wound surface will be the same as that of the cryomedium [typically 10% (v/v)].

In addition to the context of cell-based therapies, DMSO has been applied to skin wounds, either as an anti-inflammatory agent itself without other active ingredients or as a solubilizer for active ingredients such as antibiotics or anti-inflammatory agents. In humans, over 2,000 cases have been reported in the literature (Table 4). Repeated topical application of sprays or solutions containing DMSO has been used for the treatment of diabetic ulcers, venous ulcers, burns, infected wounds, suppurative and purulent wounds, skin lesions due to systemic sclerosis, and skin flaps [108–116]. The reported DMSO concentrations range from 2% (v/v) [111] to 60% (v/v) [114–116], with varying frequencies of application (up to several times daily) and durations of treatment. Local adverse effects of DMSO application, including warmth, erythema and irritation at the application site, have occasionally been reported, but have occurred only with frequently repeated applications of DMSO concentrations of 35% (v/v) and above [112–115], i.e., with at least 3.5 times higher concentrations than would be applied with MSC products [typically 10% (v/v)].

Table 4.

Adverse effects of DMSO reported from the use for topical treatment of wounds in humans

Type of study	Type/etiology of wounds	N	Galenic formulation	DMSO concentration (v/v)	Application frequency	Duration of treatment	Effect of DMSO vs. control	Adverse effects	Reference
Retrospective field report	Upper and/or lower extremity ulcerations, infected wounds, burns1)	13211)	Spray	n.s.	n.s.	As needed	n.a.	None	[108]
Retrospective field report	Venous ulcers, herpes zoster, herpes simplex, impetigo, burns, surgical wounds1)	991)	Spray	n.s.	As needed	As needed	n.a.	n.s.	[109]
Prospective, randomized controlled trial	Venous ulcers	50	Spray	n.s.	1× daily during the first week, then 1x weekly	Until healing, up to 12 weeks	Significant increase in ulcer healing rates	Local itching and erythema (n = 3); no systemic side effects	[110]
Prospective, randomized, controlled, double-blind trial	Digitals ulcers in systemic scleroderma	25	Soaking solution	2%	3× daily for 15 min	12 weeks	No difference in various wound healing parameters and pain	Urticaria in 1 patient	[111]
Prospective, controlled trial	Diabetic foot ulcers	20	Soaking solution	25%, raised to 50% if needed	3× daily	Until healing, up to 15 weeks	Significant increase in ulcer healing rates	25% DMSO: well tolerated; 50% DMSO: local irritation, burning sensation	[112]
Prospective, uncontrolled trial	Ulcers in systemic sclerosis	152)	Soaking solution	First 2 days 35%, then 46%	First 2 days 3× daily, then 4× daily	6 months to 1 year	n.a.	“a minimum of adverse reactions”; no ocular or other serious toxicity	[113]
Retrospective field report	Inflamed and suppurative wounds	510	Aqueous solution	40 – 60%	n.s.	n.s.	n.a.	Dermatitis (n = 8); Nausea/vomiting (n = 3)	[114]
Prospective, randomized, controlled consecutive case series	Mastectomy or ilioinguinal lymphadenectomy skin flaps	24	Spray	60%	First 2 days every 2 h, then every 4 h	10 days	Significant reduction in mean area of flap ischemia	Local erythema and warmth; garlic-like breath odor	[115]
Prospective, randomized, controlled consecutive case series	Mastectomy skin flaps	33	Spray	60%	First 2 days every 2 h, then every 4 h	7 days	Significant reduction in mean weight of necrotic tissue in flap edges	Garlic-like breath odor	[116]
Prospective, randomized, controlled, double-blind trial	Digital ulcers in systemic scleroderma	28	Soaking solution	70%	3× daily for 15 min	12 weeks	No difference in various wound healing parameters	Significant skin reactions in 29% of patients, garlic-like breath odor	[111]

Studies are ordered by increasing DMSO concentration of the product, with studies in which the DMSO concentration is not specified listed first. n.a., not applicable; n.s., not specified

¹Cases of dermatologic conditions not usually associated with open wounds were excluded

²With a total of 61 ulcers

Although topical DMSO has been used in many patients, its tolerability and safety have often not been systematically analyzed and reported in the literature. In addition, most of these patients have been treated outside of clinical trials or in uncontrolled studies, so there are no control group data available for most wound etiologies. Specifically, while controlled studies have provided evidence of beneficial effects of DMSO in venous ulcers, diabetic foot ulcers, and skin flaps [110, 112, 115, 116], for the other wound etiologies studied, only nonobjectifiable perceptions of beneficial effects, if any, of DMSO on wound healing have been reported. Therefore, based on these studies, potential adverse effects of DMSO on wound healing cannot be ruled out with certainty. To further complicate matters, the effective amount of DMSO applied to the wound surface cannot be deduced from any of these studies. Therefore, we also reviewed published controlled animal studies in which the size of the wound area, the amount and DMSO concentration of the formulation administered, and the frequency and duration of application were predetermined and reported [117–123] (Table 5).

Table 5.

Topical administration of DMSO in preclinical wound models in controlled studies

Wound model			N	DMSO treatment					Control	Effect of DMSO vs. control	Reference
Species	Wound type	Wound size		DMSO concentration (v/v)	Amount1) per application	DMSO dose per wound area2)	Application frequency	Duration of treatment
Rat	Full-thickness excisional wounds	4 cm2	5	10%	400 µl	11 mg/cm2	1× daily	21 days	Water	Greater wound size reduction on day 7 and 12, less wound size reduction on day 21, less inflammatory infiltration, epithelialization and collagen maturation on day 21, unchanged fibroblast and blood vessel density on day 21	[117]
	Full-thickness excisional wounds in diabetic animals	0.8 cm2	10	10%	100 µl	13.75 mg/cm2	1× daily	7 days	No treatment	No histopathological differences on day 7	[118]
	Skin ulcerations induced by intradermal doxorubicin extravasation	varying.	9	99%	3 drops	n.d.	3× daily	10 days	Saline	Significant reduction of necrotic areas and inflammatory cells	[119]
	Hind paw longitudinal incisional wounds	Skin 1 cm, fascia 1 cm, muscle 0.5 cm	6	100%	10 µl	n.a.	Once	n.a.	No treatment	Reduced pain behavior; no difference in healing at the macroscopic and microscopic levels	[120]
Dog	Second-degree burn wounds	4 cm2	53)	99%	n.s.	n.d.	Every 3 days	Until healing, up to 27 days	No treatment	Greater wound size reduction on day 21 and day 24, less clinical signs of inflammation	[121]
Pig	Partial-thickness wounds	0.7 cm2	64)	95%	100 µl (per wound)	149 mg/cm2	1× daily	7 days	Water or no treatment	No difference number and types of wound bacteria and in wound re-epithelialization over 7 days, undetectable DMSO and dimethyl sulfone serum levels over 6 days	[122]
Mouse	Full-thickness excisional wounds	0.8 cm2	12	0.007–0.14%5)	50 µl	0.0048–0.096 mg/cm2	1× daily	Until healing	Saline	Lower concentrations (0.035%)6) accelerated but higher concentrations (0.14%)7) delayed wound closure	[123]
	Full-thickness excisional wounds in diabetic animals	0.8 cm2	12	0.035%6)	50 µl	0.024 mg/cm2	1× daily	Until healing	Saline	Acceleration of wound closure	[123]

n.a., not applicable; n.d., not derivable; n.s., not specified

¹Refers to the amount of the solution applied

²Calculated on the basis of reported wound size and volume and DMSO concentration of the solution used, given a density of DMSO = 1.1 g/ml

³With a total of 10 DMSO-treated wounds

⁴With a total of 200–250 DMSO-treated wounds. On each of the 6 pigs, 100–125 wounds were inflicted, one-third (33–42) of which each were either treated with DMSO or water or left untreated

⁵Reported as 1–20 mM

⁶Reported as 5 mM

⁷Reported as 20 mM

In two studies [117, 118], nondiabetic and diabetic rat skin wounds were treated with 10% (v/v) DMSO, the same concentration that would be administered if an undiluted MSC product cryopreserved in 10% (v/v) DMSO was used for wound treatment. In four further studies in rats, dogs and pigs [119–122], even almost or completely undiluted DMSO [95-100% (v/v)], i.e., 9.5 to 10 times the DMSO concentration in an MSC product preserved in standard cryomedia, was administered. In all these studies, topical DMSO had no or even beneficial effects on wound healing, such as accelerating wound closure and reducing inflammation and pain, even when the treatment was daily for one to several weeks (Table 5). In only one of these studies, a pro-healing effect observed after 7 and 12 days of daily application of a 10% (v/v) DMSO solution changed to a delay in wound healing after 21 days of daily application compared with the control treatment [117]. However, this should not be relevant for the safety assessment of DMSO contained in topical MSC products for wound treatment, as MSC products are usually applied to wounds only once or a few (up to four) times at longer intervals (e.g. after a lack of response to treatment has been detected) [100–107].

In addition to the concentration, the amount of DMSO may also be critical to the tolerability of topical DMSO-containing products. Assuming a typical DMSO concentration of 10% for cryopreserved MSC products and an application volume of 100 µl/cm² wound area [104], 11 mg of DMSO per cm² wound area would be applied. This DMSO dose is at the lower end of the range of DMSO doses from 11 to 149 mg/cm² that have been repeatedly applied to wounds in the above studies in rats [117, 118] and pigs [122] without adverse effects (Table 5).

An exception among the animal studies listed in Table 5 is the study in healthy and diabetic mice that used significantly lower (millimolar) concentrations of DMSO, resulting in significantly lower doses of DMSO (0.0048–0.096 mg/cm²) applied daily to skin wounds. In this study, while lower doses of DMSO accelerated wound healing, higher (but still well below other published studies) doses of DMSO delayed wound healing compared with saline-treated controls [123]. Unfortunately, the authors did not comment on this apparent discrepancy with the other (previously published) studies and did not address possible reasons, e.g., species-specific differences. In any case, adverse effects, if any, were observed only with repeated (daily) application of DMSO, whereas MSC products are usually applied to wounds only once or at most a few times at longer intervals [100–107].

Potential systemic toxicity resulting from transdermal absorption

As a small amphiphilic molecule, DMSO easily penetrates the skin and can be absorbed from the skin into the blood. In humans, DMSO was detectable in the blood as early as 5 min after the cutaneous application of a 90% DMSO solution, with maximum serum levels occurring after 4–6 h [45], and the daily absorption of DMSO applied as an 80% DMSO gel was between 25% and 40% of the total DMSO dose [44]. Thus, in addition to its local effects, topical DMSO may also have systemic effects. In fact, in two historical human toxicity studies conducted in the late 1960s, daily dermal application of an 80% DMSO gel to deliver daily supratherapeutic doses of 1 g DMSO/kg for 14 or 90 days was associated with higher incidences of headache, sedation, nausea, and dizziness than in controls, whereas, even after 90 days of daily treatment, no effects of DMSO were observed on blood chemistry (except for transient eosinophilia possibly due to histamine release in the skin), hematologic and urine analyses, physical and ophthalmologic examinations, pulmonary function, and electrocardiogram and electroencephalogram recordings [124]. When lower, therapeutic doses of DMSO were applied dermally, as was done to treat musculoskeletal injuries and inflammation, the above systemic symptoms were much less common: With dermal application of 2–5 ml of a 90% DMSO solution, equivalent to 2–5 g of DMSO/patient, 1–4 times a day (average maximum daily dose of 10 ml, equivalent to 9.9 g of DMSO/patient/day) for several weeks or until symptoms were resolved, systemic adverse events such as nausea and headache were reported in only 1.6% of 4,180 treated patients, although all patients experienced the typical garlic-like breath odor caused by the DMSO metabolite dimethyl sulfide [125].

Unfortunately, in contrast to the above studies on the systemic resorption and toxicity of DMSO applied to intact skin, there are no published data on the systemic resorption and toxicity of DMSO after application to skin wounds. For this reason, the assessment of the risk of systemic DMSO toxicity from DMSO-containing MSC products used for the topical treatment of skin wounds was based on a worst-case scenario involving the modeling assumptions described below.

To anticipate the potential systemic toxicity of DMSO that may result from the topical application of DMSO-containing MSC products, two main factors need to be considered: (1) the amount of DMSO applied and (2) the extent to which the DMSO is actually absorbed into the bloodstream. When MSC products are used for wound healing, the amount of DMSO administered to the patient depends on the size of the wound. A systematic review of large observational studies of venous leg ulcers, one of the main target indications for MSC-based wound treatment, from all regions of the world revealed that the average wound size ranged from approximately 20 to 40 cm² [126]. For wounds of this size, with a DMSO concentration of 10% in the MSC product and an application volume of 100 µl/cm² wound area, 0.22–0.44 g of DMSO would be applied, and for very large ulcers of 100 cm², 1.1 g of DMSO would be applied. Thus, even with such an exceptionally large wound size, the amount of DMSO applied would be approximately one-tenth of the average maximum daily dose of 9.9 g of DMSO/patient/day applied topically for therapeutic purposes in 4,180 patients in the study described above [125].

The second variable determining the potential systemic toxicity of DMSO administered with topical MSC products is the extent to which the DMSO is absorbed into the bloodstream. This is much more difficult to predict as, to our knowledge, studies of transdermal absorption of DMSO in humans are available only after application to intact skin and only for higher (80–90%) concentrations of DMSO [44, 45], but not after application to wounds or for lower (10%) concentrations of DMSO. This gives rise to two caveats: On the one hand, the above observations from studies in which DMSO was applied to intact skin have limited applicability to the application of DMSO to damaged skin. Basically, the stratum corneum represents the main barrier to the penetration of exogenous substances through the skin [127]. Thus, although DMSO can cross this barrier relatively well [128], the transdermal absorption of DMSO through wound beds may be greater than that through uninjured skin with an intact stratum corneum. On the other hand, the transdermal absorption rate of dermally applied DMSO depends on the DMSO concentration in the applied product, decreasing at lower DMSO concentrations: while DMSO penetrates the skin most efficiently at concentrations of 80% and above, at concentrations of 67% and less, DMSO molecules become hydrated, which greatly reduces dermal penetration [129]. Taken together, in the absence of appropriate studies, the order of magnitude of the actual absorption rate of DMSO when a 10% DMSO-containing MSC product is applied to wounds remains unclear. The 25–40% absorption rate observed in humans was determined with intact skin and a significantly higher DMSO concentration (80%) [44], so the absorption rate of DMSO from MSC products applied to open wounds may be both significantly greater or significantly lower. Therefore, in the interest of maximum safety, the assessment of the risk of indirect systemic toxicity from DMSO delivered with topical MSC products was based on a worst-case scenario in which a 100% absorption rate for DMSO applied to wounds was assumed.

In this scenario of complete DMSO absorption, 11 mg of DMSO/cm² wound area would enter the bloodstream at a DMSO concentration of 10% in the MSC product and an application volume of 100 µl/cm² wound area. Under these conditions, a patient with a 100 cm² wound would absorb a maximum of 1.1 g of DMSO. For a (rather low) body weight of 60 kg, this would correspond to a systemic exposure of 18 mg/kg. Even in this case, the exposure would be 55 times lower than that from an intravenous dose of 1 g/kg, which is generally considered acceptable to be administered with HSC transfusions [59–62].

Conclusions

Although extensive meta-analyses of clinical trials of MSC therapies have generally confirmed the overall safety of MSC therapies [39, 40], the potential toxicity of DMSO to patients treated with MSCs remains a topic of debate [7]. Unfortunately, the widespread use of DMSO as a cryoprotectant for MSCs for therapeutic purposes is offset by a notable underreporting of information on the composition of the cryopreservation media and reconstitution of the product in clinical trial reports [63]. Data from intravenous application of MSC products for which information on DMSO content and instructions for product reconstitution are available show that the dose of DMSO delivered with intravenous administration of therapeutic doses (2–8 × 10⁶ cells/kg) of MSC products in which the MSCs were cryopreserved in 10% DMSO at 1–6.7 × 10⁶ cells/ml is well below (2.5–30 times lower than) the limit of 1 g DMSO/kg body weight that is generally considered acceptable for HSC transplantation. In these settings, after adequate premedication with corticosteroids and/or antihistamines, only isolated infusion-related reactions, if any, have been reported in more than 1000 patients, many of whom have received multiple infusions. This positive safety data indicates that the current use of DMSO as cryoprotectant for intravenous MSC therapy products is generally safe.

In contrast to the systemic application of DMSO-containing MSC products, there are no published data on the topical application of DMSO delivered with MSC products onto skin wounds. Published experience with the application of DMSO to skin wounds outside the context of cell-based therapies suggests that at concentrations of DMSO that would be applied with undiluted DMSO-cryopreserved MSC products (typically 10%), no significant local adverse effects are expected. Because DMSO can penetrate the skin and enter the bloodstream, the possibility of systemic side effects must also be considered when using DMSO-containing MSC products topically. However, while the extent to which DMSO may be absorbed from MSC products applied topically to wounds is unknown, even in the worst-case scenario, which assumes 100% absorption of the DMSO content of an MSC product applied to a large (100 cm²) wound in a lightweight (60 kg) adult, the systemic exposure would be approximately 55 times lower than that from an intravenous dose of 1 g/kg, which is generally considered acceptable for administration with HSC transfusions.

While these conclusions are limited by the lack of data from prospective studies on the adverse effects of DMSO administered with MSC products, the available data do not suggest any significant safety issues with the DMSO contained in intravenous or topical MSC products cryopreserved according to current standard protocols, provided that premedication with corticosteroids and/or antihistamines is given for intravenous administration. In summary, although the development of DMSO-free cryopreservation protocols that ensure equally good cell viability and functionality is welcome, the current gold standard is unlikely to pose a safety barrier to the systemic and topical use of cryopreserved, bed-thawed MSC products. Nevertheless, closely monitored and well-reported clinical trials are needed to confirm this conclusion.

Abbreviations

aGvHD

Acute graft versus-host disease

ARDS

Acute respiratory distress syndrome

COVID-19

Coronavirus disease 2019

DMSO

Dimethyl sulfoxide

HSA

Human serum albumin

HSC

Hematopoietic stem cell

MPC

Mesenchymal precursor cell

MSC

Mesenchymal stromal cell

Author contributions

EN-R conceived and designed the review, performed the literature search, analyzed and interpreted the data, and drafted the manuscript. MAK supervised the study and provided critical revisions. All authors read and approved the final manuscript.

Funding

Not applicable.

Data availability

Not applicable.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

EN-R is an employee, and MAK is the COO of RHEACELL GmbH & Co. KG, Heidelberg, Germany.