TGFβ1 priming enhances CXCR3‐mediated mesenchymal stromal cell engraftment to the liver and enhances anti‐inflammatory efficacy

Abstract The immunomodulatory characteristics of mesenchymal stromal cells (MSC) confers them with potential therapeutic value in the treatment of inflammatory/immune‐mediated conditions. Previous studies have reported only modest beneficial effects in murine models of liver injury. In our study we explored the role of MSC priming to enhance their effectiveness. Herein we demonstrate that stimulation of human MSC with cytokine TGβ1 enhances their homing and engraftment to human and murine hepatic sinusoidal endothelium in vivo and in vitro, which was mediated by increased expression of CXCR3. Alongside improved hepatic homing there was also greater reduction in liver inflammation and necrosis, with no adverse effects, in the CCL4 murine model of liver injury treated with primed MSC. Priming of MSCs with TGFβ1 is a novel strategy to improve the anti‐inflammatory efficacy of MSCs.


| INTRODUC TI ON
Mesenchymal stromal cells (MSC) represent a promising therapeutic approach in many conditions, including inflammatory liver disease and graft versus host disease, 1 as a consequence of their potent immunomodulatory properties. 2 However, their efficacy in rodent and human models of liver injury has been variable, with some studies demonstrating benefit from MSC infusions [3][4][5] whilst others report that infusion of conditioned medium from MSC cultures was sufficient to confer efficacy. 6 Moreover, the mechanism of action by which MSC exert their effects in models of liver damage is poorly delineated with reports suggesting they may be mediated by a reduction in oxidative stress 3 and/or reduced lymphocytic ingress to the injured liver with a secretome analysis suggesting this latter effect may be chemokine dependent. 6 Whilst others have suggested that a component of MSC action may occur remotely without requirement for homing to the injured organ, 7,8 the relative lack of efficacy of MSC in models of liver injury has been attributed to low levels of MSC engraftment in the damaged liver. Using flow-based assays we and other groups have demonstrated that β1 integrin and CD44 are involved in the firm adhesion of MSC to hepatic sinusoidal and human umbilical endothelium. 9,10 Notably, chemokine receptors did not appear to contribute significantly to human MSC recruitment, 11 which was unexpected considering chemokine receptors play a significant role in leukocyte recruitment. 12 Moreover, studies using murine MSC adhesion to murine aortic endothelium 13 18 and also impact upon subsequent targeting in tissue. To mitigate for this, cell surface glycans on MSC have been chemically engineered into an E-selectin binding motif in order to encourage engraftment to endothelium that expresses high levels of E-selectin. 19 Similarly pre-loading of therapeutic MSC with paramagnetic nanoparticles has been utilized to allow specificity of delivery 20 ; however, these methods of enhancing MSC migration are unlikely to be acceptable for clinical practice for logistical, safety and cost reasons. Therefore, we explored the consequences of cytokine stimulation of MSC upon their hepatic engraftment and efficacy. We used cytokines known to increase inflammatory cell ingress and that are elevated in liver disease such as TNFα, IFNγ, TGFβ 1 , LPS, IL1β, IL4, IL6, IL8 and IL10. [21][22][23][24] Importantly, MSC have been reported to have receptors for these cytokines including TNFRI and IIR, 25 IFNγR, TLR4, 26 IL-1R, IL-4R, IL-6R, 27 IL8R (CXCR1) 25 and IL10R. 28 Herein we report that pre-stimulation of clinically relevant human MSC with TGFβ 1 enhances their binding/engraftment to hepatic sinusoidal endothelium ex vivo and in vivo in a CXCR3-dependent manner and results in greater potency to reduce liver damage in an acute model.  29 and Alcoholic Liver Disease [ALD]).

| Human liver tissue and cell culture
All samples were collected with local research ethics committee approval (reference number 06/Q2702/61) and informed, written patient consent. Freshly collected liver tissue was either snap frozen and sectioned to 10 μm for Stamper Woodruff adhesion assays or used for the isolation of hepatic sinusoidal endothelial cells (HSEC), biliary epithelial cells and hepatic myofibroblasts as previously described. 30 Where indicated, cultured primary cells were treated with 10 ng/mL TNFα and IFNγ (both Peprotech) for 24 h prior to use in adhesion assays.

| Adhesion and migration assays
Adhesion of MSC to cultured cell monolayers, human liver tissue sections or mouse liver sections (control and CCl 4 treated) was assessed using a modified Stamper Woodruff static adhesion assay.
To assess migration of control or TGFβ 1 -stimulated MSC we used a modified 48-well Boyden chamber as previously described. 31

| Hepatic engraftment of MSC
All animal procedures were conducted in accordance with UK laws with the approval of the Home Office and local ethics committees (PPL 40/3201). Carbon tetrachloride (CCl 4 ; Sigma Aldrich) diluted 1/4 in mineral oil (Sigma) was administered by intraperitoneal injections (1 mL/kg, twice weekly for 8 weeks or acutely as a single injection) into 9-week-old C57Bl/6 wild type male mice.
Where indicated, MSC were pre-incubated with blocking antibodies raised against chemokine receptors (anti human CXCR3, CCR5 or CXCR4 at 20 μg/mL, all from R+D systems) for 15 min at 37°C, washed and re-suspended in PBS 0.1% BSA. To study engraftment of MSC into liver and non-hepatic organs, MSC (control or 5 ng/mL TGFβ 1 -stimulated) were labelled with Direct red (DiR 5 μM; Invitrogen) or CFSE according to manufacturer's instructions. Cells 1 × 10 6 were either injected into the hepatic portal or tail vein of mice that had been acutely injured with CCl 4 (1 mL/ kg IP, 72 h). Organs were harvested 72 h later and imaged using an IVIS Spectrum Imaging System (Perkin Elmer). Fluorescent and photographic images of individual organs were analysed using Living Image software. Full details of all experimental protocols are available in Appendix S1.    To test adhesion and engraftment of MSC in injured liver in vivo CFSE-labelled MSC were infused into control or acutely CCl 4 -injured C57 Bl/6 mice via the portal vein. MSC were infused either unstimulated, or stimulated with TGFβ 1 , IL4 or IL10. We observed increased engraftment of TGFβ 1 -stimulated MSC in injured mouse livers (2.29 ± 0.08 fold increase; p < 0.001) compared to unstimulated MSC ( Figure 2B or C), whereas IL4 and IL10-stimulation had no impact on engraftment.

| TGFβ 1 stimulation of MSC increases chemokine receptor expression, promotes redistributes redistribution chemokine receptors to the cell surface from the cytoplasm and enhances migration to their cognate ligands
Since TGFβ 1 -stimulation of MSC ( Figure 3A and CXCR3 blockade reduced binding to injured liver sections back to basal levels ( Figure 4A). We then infused CFSE-labelled MSC into CCl 4 -injured mice via the portal vein, and observed increased engraftment of TGFβ 1 -stimulated MSC in mouse livers. Whilst blocking CXCR3 on unstimulated MSC had no effect on their engraftment in injured mouse livers, there was a marked effect on TGFβ 1 -stimulated MSC with engraftment reducing from a 2.32 ± 0.22 fold increase from baseline to a 0.63 ± 0.11 fold reduction (p < 0.001; Figure 4B).
In contrast, blockade of CXCR4 and CCR5 blockade on MSC had no effect on engraftment of either control or stimulated MSC in injured

| TGFβ 1 -stimulated MSC reduce liver inflammation, necrosis and liver serum aminotransferase levels in a mouse model of liver damage
To determine the impact of MSC infusion on the pathogenesis of  Figure 6A). Similarly, the injury-

| Infusion of unstimulated and TGFβ1stimulated MSC after acute CCl 4 injury results in a reduction in M1-like/M2-like ratio of hepatic macrophages
The

| DISCUSS ION
We have demonstrated that TGFβ 1 stimulation of MSC more than doubles their homing to the acutely injured liver and is associated with a resultant further reduction in inflammation and hepatic dam- Recent studies suggest that allogeneic MSC, although hypoimmunogenic, are not intrinsically immune privileged and that allogeneic MSC induce a memory T-cell response resulting in rejection. 47 Although human MSC are even more likely to generate an immune response after infusion into mice, we did not see an increase in  can reduce oxidative stress 3 and CD45 infiltration. 6 We saw a reduced CD45 + infiltrate after administration of MSC, by immunohistochemistry and flow cytometric analysis of digested murine liver, which correlated with reduced tissue necrosis and ALT in serum.
Our data suggest that TGFβ 1 stimulation also enhances the ability of MSC to suppress T cell proliferation or recruitment, and thus this may be a factor in the superior efficacy seen with primed cells.
However, further work is required to establish whether the efficacy seen with TGFβ 1 -dependent priming of MSC is predominantly driven by enhanced immunomodulatory action of MSC or their increased hepatic homing.
Furthermore, our data indicate that hepatic macrophage profile changes significantly following administration of MSC, with or without, TGFβ 1 stimulation. Our data indicate that MSC infusion is associated with a reduction in differences in the proportion of macrophage subsets expressed as a ratio of Ly-6C hi /Ly-6C lo (M1-like to M2-like) macrophages. The differential expression of Ly-6C has been used to identify monocyte subsets in rodent models of liver injury where Ly-6C hi monocytes exhibit pro inflammatory phenotype (M1) and Ly-6C − monocytes exhibit the restorative phenotype (M2). 56 As recognized by the literature, surface marker expression of macrophages is likely to be more complex and dynamic and thus even more extensive panels (CD163, CD206, CD68 and TLR4) do not completely characterize the full phenotype of macrophages in vivo 57 and our panel is acceptable with these caveats. 58   MSC. However, there is also evidence that lung-resident monocytes can also phagocytose trapped MSC and differentiate to regulatory macrophages which can then migrate to distant sites. 61 Given we did indeed see a background level of MSC entrapment in the lungs ( Figure 5), it is also possible that cells trafficking from this site could contribute to the pool of M2 macrophages we identified in our injured livers. Thus, in our model, the hepato-protective effects of TGFβ 1 primed MSC may be linked to a direct suppression of T cell activation and recruitment, and enhanced macrophage recruitment and differentiation within the liver, thus shifting the hepatic microenvironment towards a more reparative situation. Further study of the phenotype of hepatic myeloid cell subsets would be of value.
In conclusion, we have demonstrated that priming of MSC with

This work was supported by University Hospital Birmingham
Charities. PNN is supported by the NIHR Birmingham Biomedical Research Centre based at University Hospitals Birmingham and the University of Birmingham. The views expressed are those of the author and not necessarily those of the NHS, the NIHR or the Department of Health.

CO N FLI C T O F I NTE R E S T S TATE M E NT
There are no relevant disclosures.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.