1. IntroductionMesenchymal stem/stromal cells (MSC) have gained great interest as new medical treatments. Clinical development of MSC therapies is based on extensive studies in ani-mal models for human disorders and diseases demonstrating improved outcomes [1]. MSC can act by three major classes of mechanisms [1] differentiation into different types of cell lineages and integration into tissues, which have applications for regenerative medicine, [2] MSC direct contact with host cells to modulate functions of effector cells, and [3] secretion of factors including those that promote cell survival and growth, and cytokines that modulate inflammation and immune cell function [2,3]. Proof of concept for efficacy of MSC in the clinic has been demonstrated for Graft vs. Host Disease, which is believed to involve one or both latter two mechanisms by modulating cytokine storm and inhibiting inflammation [4,5]. However, the functional roles and fates of MSC differentiation and integration after injection into humans have not been elucidated. Understanding mechanisms of MSC action, which has been difficult in vivo even in animal models, will facilitate improved treatments for translational studies [2,3].Cytokine storm occurs in injuries and diseases that have persistent highly elevated levels of pro-inflammatory cytokines [6]. Cytokine storm often occurs in sepsis with inflammatory responses to pathogens [7], which can lead to multiple organ failure with a mortality rate of >25% [8]. The human survival rate has decreased to this level over the past several decades primarily due to improved diagnosis and more aggressive critical care, however, no new therapies have been developed over decades [9]. MSC are effective in treating sepsis in rodent models [10] by releasing anti-inflammatory cytokines including IL-1ra, IL-4, and IL-10, and prostaglandins, e.g., PGE2, which suppress inflammation and resolve cytokine storms [11].Cell encapsulation in alginate was initially developed to treat diabetes using islet cells to release insulin [12]. Analysis of encapsulated MSC (eMSC) in vitro has proven that secretion of cytokines and other factors can suppress secretion of pro-inflammatory cytokines from activated immune cells such as macrophages [11]. Although encapsulated islets are functional in vivo for short periods [13], recent modifications including the use of less adhesive alginates, which minimize foreign body reactions, yielded encapsulated islets that release insulin for as long as 9 months in non-human primates [14], providing preclinical feasibility for translation [15].eMSC promote functional recovery after myocardial infarction [16], hindlimb ischemia [17], and spinal cord injury (SCI) [18,19] in animals. In addition, encapsulated genetically engineered cells secrete bioactive proteins in vivo [20,21,22,23]. Encapsulation in alginate prevents migration of cells out from the capsules, thereby allowing effects of secreted factors to be analyzed without complications due to direct interactions of the encapsulated cells with host cells. In contrast to IV-injected MSC that disappear rapidly [24], encapsulation prolongs MSC survival in vivo for weeks to months [14].We showed previously that intrathecal injection of eMSC into the cauda equina one day after rat SCI mitigated inflammation and improved functional recovery in SCI [18]. By comparison to empty capsule controls, eMSC increased expression of CD206, a marker for anti-inflammatory M2 macrophages, at ~2 cm from the SCI site at thoracic segments 9–10 [18] and decreased expression of the pro-inflammatory isolectin IB4 expressed on activated microglia and macrophages one week after injection [19].

We report here that localized injection of encapsulated human MSC in a rodent model modulates host cytokine expression within the CNS in SCI. Encapsulation enables localized delivery of MSC, and sustained survival of MSC and secretion in vivo. Given these advantages of eMSC, we have developed a scalable semi-continuous system to generate encapsulated cells in quantities sufficient for clinical translation.

4. DiscussionThe registration at ClinicalTrials.gov of greater than one thousand and five hundred clinical trials testing MSC for various indications [33], most recently including COVID-19 [34], underscores the tremendous interest in therapeutics with these cells. The use of eMSC by lumbar puncture generated changes in immunomodulatory factors such as PGE2 suggesting that secreted factors suppress cytokine storm acutely in vivo and promote improved outcomes. Several factors have prevented eMSC technology from moving to the clinic including the failure of long-term eMSC survival due to foreign body reactions, for which there are solutions [14,35]. There are also limitations in scale-up to produce sufficient quantities of eMSC for treating patients and we describe herein a solution to the problem of scale-up using a novel design for encapsulation and recovery of capsules.There appears to be a loss of approximately half of eMSC within the first week in vivo and this survival level was maintained in vivo for at least 6 weeks. The surviving MSC after 6 weeks secreted a slightly higher level of the anti-inflammatory prostaglandin PGE2 [11] ex vivo than eMSC maintained only in vitro without LPS activation, indicating the cells maintained secretory activity in vivo. However, when activated with LPS ex vivo, the eMSC expressed highly elevated levels of PGE2, suggesting that exposure to the inflammatory environment in vivo primed them for subsequent responsiveness, demonstrating that eMSC is a bio-responsive system. This dramatic effect should be considered as preliminary insofar as very limited numbers of capsules were retrieved from two rats and then pooled for a single experiment. Additional studies are needed to determine the extent and timing of eMSC responsiveness in vivo. Similar experiments are not feasible with free MSC as they cannot be recovered from the body after injection [36].The action of the eMSC must be via secreted factors given that the cells are retained in the capsules for at least for 6 weeks. Considering that the capsules are injected below the end of the spinal cord, the improved recovery in the injury site and locomotion [19] indicate that eMSC act at a distance. The increased expression of rat IL-10 mRNA in the injury site is likely to result from injection of human eMSC at a distance from the SCI site in the cauda equina. This is in contrast to free MSC, which migrate extensively but survive for only a few days after injection [36]. We have also observed that intraperitoneal injection of eMSC in rodent models of sepsis reduced serum levels of TNFα, confirming that eMSC act a distance (unpublished observations). Thus, eMSC are a better designed system than free MSC for long-term survival, making long-term treatment in chronic diseases involving inflammation feasible.The systemic effects demonstrated by changes in cytokines in the sera after IP eMSC injection is another example of eMSC action at a distance. The effect of eMSC by IP to generate changes in cytokines in blood is novel. Among the many clinical trials testing MSC, multiple doses are often provided two days to one week after an initial dose [33]. It is likely that eMSC will be advantageous because the cells survive longer than with free MSC and may not need additional dosing for several weeks. MSC have been demonstrated to save lives in graft vs. host disease [37], ARDS [38], and COVID-19 [34], and eMSC may be more effective as a bio-responsive therapy.MSC are produced by many organizations and companies in very large quantities to inject hundreds of millions of cells per patient in clinical trials [33]. eMSC are effective in several animal models of disease and injury, but this technology has not been translated to the clinic so far in part because a method for scale-up has not been devised. We have invented a RaCCS that enables scale-up to yield 36 million cells in a preliminary experiment. Considering that 30,000 eMSC produced a similar response to 250,000 free MSC in rat SCI [19], encapsulation of 36 million eMSC should be equivalent to ~300 million eMSC, which is enough to treat at least two patients with a minimal effective dose of 150 million free MSC [33]. In any case, these estimates indicate that eMSC can be produced in sufficient doses for use in at least small clinical trials.It has been suggested that MSC apoptosis and efferocytosis plays a role in the anti-inflammatory action of MSC in graft vs. host disease [37]. This should occur in less than one week after injection since MSC are barely detectable thereafter [36]. It is possible that loss of ~1/3 of eMSC that we observed within the first week in the injured spinal cord may be due to apoptosis. However, dead cells were rarely observed by Live/Dead assay in capsules retrieved from the spinal cord at one or six weeks after delivery. Thus, long-term effects of eMSC are likely attributable to secreted factors.The ability of eMSC to reduce blood levels of pro-inflammatory cytokines within 5 h in the LPS-induced endotoxicity underscores the rapidity of the response to eMSC injected intraperitoneally, outside the bloodstream in the rat. However, this is not the best model for sepsis. In preliminary studies we found using a more appropriate model for sepsis, i.e., mouse cecal ligation and puncture, that human eMSC reduced levels of IL-6 and TNF-α in sera after 16 h treatments with eMSC (unpublished observation MG, SB, MK). The combined results suggest than eMSC may be useful for acute treatment of cytokine storms that occurs in many inflammatory disorders including COVID-19 [34].

The constant pressure mode of the encapsulator is advantageous over the constant flowrate mode because tubing connections can fail as the pressure rises especially in long runs to scale up production of capsules. When using constant pressure, one should first determine a high pressure that does not compromise cell viability and then determine needle length, needle outer-diameter, and needle inner-diameter as desired. As inner needle diameters increase, flow rates increase to the power of 4. Shortening the needle increases the flow rate linearly. Shorter and wider diameter needles have higher flow rates, yielding larger capsules with high cell yields in shorter run times. Optimizing these parameters may increase yields in encapsulations.

Capsules with diameters larger than 0.5 mm are difficult to inject through syringes unless their caliber is very large because they tend to aggregate. Two ways to keep the capsule size relatively low is to decrease the needle outer diameter and increase the applied voltage. Beveled needles have smaller outer diameters than blunt ones, thus producing smaller capsules without lowering the flow rates [19]. High electric fields do not decrease the viability of cells in microcapsules due to the Faraday cage effect and it has been reported that voltages as high as 30 kV do not decrease viability [39,40]. The downside of using high voltages is increased needle vibration as we observed with the 2-inch 27 G needle at 8.0 kV. This effect can be minimized using shorter needle lengths.The most widely used electrostatic cell encapsulator is produced by Nisco, Zurich, Switzerland (http://www.nisco.ch/var_v1.htm, accessed on 30 November 2022). In this system, cells suspended in alginate monomers are driven by a syringe pump at a constant flow rate to extrude droplets from a needle. The unpolymerized alginate droplets are driven under an electrostatic potential into a collecting vessel where they are crosslinked by 20–100 mM divalent cations in a solution that is mixed using a stir bar to prevent capsules from clumping. At the conclusion of the run, the apparatus is disassembled to retrieve the capsules from the collection bath for further processing including post-encapsulation treatments and washing into cell culture media. Other encapsulators from Buchi, Flawil, Switzerland (https://www.buchi.com/us-en/products/spray-drying-and-encapsulation/encapsulator-b-395-pro, accessed on 30 November 2022) and Inotech, Flawil, Switzerland (http://www.encap.ch/, accessed on 30 November 2022) have more complex designs for batch stirring and collection but they also are closed systems that do not allow sampling or capsule collection until the run is terminated.Given that relatively high concentrations of divalent cations used for crosslinking and unpolymerized alginate may be detrimental to cells, it is important to transfer capsules into physiological buffers as soon as possible after encapsulation. One also needs to consider that the duration of the encapsulation run can result in large differences in crosslinking times and exposure to divalent cations between formation of the first and the last capsule that are unavoidable with batch reactors. For this reason, protocols using batch reactors often add a post-crosslinking period to ensure sufficient cross-linking for capsule stability while reducing relative differences in total crosslinking times among the capsules. Nevertheless, capsules will be subject to different exposure times that introduce heterogeneity in the population. Although encapsulation details are often not reported, the shortest run times and hence minimal crosslinking times appear to range between 20–30 min using up to ~1 mL of alginate with ~1–20 million cells [16]. This yields several millions of encapsulated cells, which are sufficient for studies with rodents but not enough for large mammals. Scale-up is problematic with these closed system encapsulators because they require longer run times that will increase heterogeneity further.