Practical Barriers to Delivering Autologous Bone Marrow Stem Cell Therapy as an Adjunct to Liver Resection

Introduction

Autologous bone marrow-derived stem cell (BMSC) therapy has offered great promise in enhancing tissue regeneration following a range of acute and chronic diseases and could prove valuable in enhancing liver regeneration following liver resection.

Liver resection involves removal of the section of liver containing malignant disease, including primary and secondary tumors. Liver resection may on occasion be performed for benign disease where the lesion is symptomatic or where the diagnosis is uncertain. Although the liver is capable of regenerating following resection of up to 75% of its volume, liver failure may ensue if regeneration of the remnant liver volume is insufficient to meet functional requirements. Incidence of post-resectional liver failure (PLF) ranges between 0.7% and 9.1% and is the main factor implicated in postoperative mortality rates of up to 5% [1]. Liver cirrhosis is a major risk factor for poor surgical outcome following liver resection, with evidence of cirrhosis present in the majority of patients with hepatocellular carcinoma [2]. Even with previously healthy underlying liver parenchyma, many patients undergoing liver resection may have received chemotherapy in the preoperative period, further impairing liver function [3]. Reducing the incidence of PLF has traditionally focused on minimizing risk factors for PLF, such as intraoperative blood loss and optimizing management of co-morbidities. In some cases, the future liver remnant (FLR) volume can be increased, either by 2-stage hepatectomy or portal vein embolization, where the portal vein supplying blood to the portion of the liver containing the tumor is occluded resulting in hypertrophy of the FLR [1]. Severe liver dysfunction and PLF may still occur, however.

Autologous BMSC therapy offers the potential to provide an effective treatment option in facilitating liver regeneration after liver resection. The ability to administer autologous BMSCs offers a clear advantage over other nonautologous cell therapies, negating the requirement for immunosuppression and risks of sensitization. Prospects for enhanced cardiac regeneration post-myocardial infarction has received the greatest attention with recent 5-year outcome data following BMSC transplantation post-myocardial infarction demonstrating longstanding improvement in cardiac performance and mortality [4]. BMSC therapy has been shown to enhance hepatic regeneration in acute and chronic settings in both preclinical studies and initial clinical pilot investigations [5,6], but translation to clinical trials especially in relation to liver surgery has been restricted by a range of practical issues.

This article addresses the practical barriers currently restricting the development of clinical trials into bone marrow-derived cell therapies and hepatic regeneration following liver resection. The article reviews evidence of potential efficacy, trial design, methods of therapy delivery, and assessment of efficacy.

Bone Marrow Stem Cells and Hepatic Regeneration: Preclinical Evidence

Bone marrow contains hematopoietic stem cells (HSC), responsible for renewing circulating blood elements, and mesenchymal stem cells (MSC) that contribute to a wide range of mesenchymal tissues. Bone marrow stem cells (BMSC) have long been recognized as possessing the potential to support hepatic regeneration, initially following the discovery of donor-derived cells in livers of patients who had received bone marrow transplantation [7,8]. Subsequently, a distinct mobilization of BMSCs has been observed in humans following liver resection or liver transplantation [9 –11]. In rodents, green fluorescent protein-labeled BMSCs have been shown to migrate in significant numbers to the regenerating liver following liver resection [12].

Controversy has existed as to the possible mechanism by which BMSCs may contribute to liver regeneration. Hypotheses have included transdifferentiation to hepatocytes, cell fusion creating hepatocyte cell hybrids, or through paracrine effects [13,14]. While the transdifferentiation of BMSCs into hepatocytes has been reported [15], many more recent studies have not found this to occur at a clinically relevant magnitude [14]. Furthermore, studies demonstrating functional rescue of a metabolic liver model [16] have later been shown to occur as a result of cell fusion between macrophages and hepatocytes and not transdifferentiation [17]. The model that produced this rescue of the metabolic liver disease (FAH deficiency) produces a very strong selection pressure and it is questionable as to whether this would be relevant or reproducible clinically. With regard to MSCs, they can clearly be differentiated to a hepatocyte-like phenotype in vitro [18], but how functional these cells are in vivo is still debatable. For example, upon transplantation these MSC-derived hepatocyte-like cells can produce some hepatocyte proteins, for example albumin, but typically have not been shown to have the full complementation of metabolic machinery to be characterized as fully hepatocytic [19]. Another issue is that engrafted MSCs can differentiate into scar-forming myofibroblasts and so phenotypic stability as hepatocyte-like cells needs to be confirmed in long-term preclinical transplant models [20]. Given these findings, BMSCs are now not thought to play a significant role in the direct generation of a clinically relevant quantity of hepatocytes, but instead are thought to contribute to hepatic regeneration mainly through paracrine effects, in particular enhancing angiogenesis. Just as neoplastic growth is dependent on the development of new blood vessels, the regenerating liver must be sustained as its mass rapidly increases. Sinusoidal endothelial cells (SECs) are thought to play a central role in coordinating this process, with the majority of BMSCs migrating to the liver committing to this cell type [12,21,22].

The exact role of bone marrow-derived MSCs and HSCs in hepatic regeneration remains unclear, although mechanisms of benefit have been demonstrated for both cell lineages. Mobilized HSCs in rodent models have been shown to reduce mortality in acute liver injury and accelerate recovery in chronic liver injury by inducing proliferation of resident hepatocytes through endogenous mechanisms [23]. In rodent toxin-induced acute hepatic injury, HSCs have been shown to facilitate vascular remodeling, primarily via collateralization, targeting higher activity toward severely damaged areas [24]. Bone marrow-derived MSCs are able to inhibit the proliferative and fibrogenic function of activated stellate cells, inducing stellate cell apoptosis [25]. Through synthesis of growth factors and cytokines, administration of bone marrow-derived MSCs can enhance repopularization of necrotized tissue by endogenous cells, up-regulate anti-apoptotic and cell survival signals and may also play a role in protection against oxidative insults [26]. Molecules produced by bone marrow-derived MSCs can induce a selective emigration of leukocytes from the liver, reducing fibrosis in a rodent model of fulminant hepatic failure [27]. Commitment of HSCs and MSCs to other cell types including macrophage or myofibroblast cell lineage has also been demonstrated in the liver, with cell lineage contribution likely dependent on the nature of injury [28,29]. The injury–recovery interaction is certainly complex with specific macrophage subpopulations shown to possess distinct roles in both injury and recovery phases of hepatic injury, promoting fibrogenesis during injury and promoting matrix degradation during recovery [30]. These preclinical studies demonstrate a range of mechanisms by which both HSCs and MSCs may contribute to hepatic regeneration. While the predominant function of these different cell lineages in hepatic regeneration remains unclear, both HSCs and MSCs may have a direct influence on resident hepatocyte proliferation and modulate the hepatic microenvironment.

There is now substantial evidence demonstrating that BMSCs can influence the pathological consequences of acute and chronic liver failure, although not necessarily favorably. di Bonzo et al. [31] demonstrated the propensity of MSCs to form myofibroblast-like cells in areas of hepatic injury, with others suggesting that bone marrow-derived myofibroblasts may play a significant role in human liver fibrosis [28]. These findings perhaps limit the future role of human MSCs in considering therapies to support hepatic regeneration, certainly prior to conclusive evidence that MSCs could be preprogrammed to provide a predictable response [32]. HSCs have been used for many years in the treatment of hematological malignancy and as such their behavior and safety profile following human administration is well established. The focus for translational studies has turned to HSCs that have emerged as the most promising adult stem cell source.

Bone Marrow Stem Cells and Hepatic Regeneration: Clinical Evidence

Clinical research into BMSCs and hepatic regeneration lags somewhat behind that for cardiac disease; however, several phase I studies have been conducted and the National Institute of Health (US) database of randomized clinical trials lists 4 trials currently recruiting patients with chronic hepatic disease [33]. Published studies have been small and essentially establish safety and feasibility of direct hepatic administration of BMSCs, rather than demonstrating efficacy, although benefits have been reported.

While clinical studies post-liver resection are yet to be conducted, a comparable human model is provided by pre-liver resection portal vein embolization. The technique is used to increase the volume of healthy liver (tumor-free) that will remain following liver resection in cases where the remaining healthy liver volume is thought to be inadequate. Portal vein embolization places a similar acute regenerative demand on the liver with the process of hepatic regeneration after portal vein embolization similar to that following liver resection [34]. Fürst et al. [35] compared portal vein embolization plus BMSC administration (n = 6) and portal vein embolization alone (n = 7). Groups were matched for volume of tumor-free liver, which was thought in both groups to be inadequate to meet functional requirements if the patient was to proceed straight to resection. The portal vein embolization group plus BMSC group showed significantly greater gain of FLR volume than portal vein embolization alone (77%, compared to 39%; P = 0.039) with significantly greater daily growth rate (9.5 mL/day, compared to 4.1 mL/day; P = 0.03).

Chronic liver disease has received greater attention with 8 initial phase I trials investigating the application of BMSCs to support hepatic regeneration. These studies provide evidence of the feasibility and safety of direct hepatic administration of BMSCs (Table 1). The etiology of chronic liver failure across the study groups was diverse, including alcoholic alone (n = 19), hepatitis C virus alone (n = 12), hepatitis B virus alone (n = 7), alcoholic plus hepatitis C virus (n = 2), alcoholic plus hepatitis B virus (n = 1), cryptogenic cirrhosis (n = 8), autoimmune hepatitis (n = 2), chronic cholestatic liver disease (n = 1), primary biliary cirrhosis (n = 1), and unknown etiology (n = 2). Diverse methods of BMSC isolation and selection ranging from highly selective to less discriminatory reflect current uncertainties and difficulties in selecting appropriate cell subtype. Although statements of efficacy cannot be made from these small uncontrolled trials, results are promising with modest improvement in measures of hepatic function noted by the majority of studies (Table 1). As primary liver carcinoma often occurs on a background of liver cirrhosis, studies in chronic liver disease may have relevance to BMSC administration following liver resection. Even patients with previously healthy liver parenchyma may have a degree of chronic hepatic impairment owing to preoperative chemotherapy [3]. In cirrhosis the anatomical disruption to normal hepatic architecture can be severe, with BMSCs potentially playing a role in the modulation of this abnormal environment [32]. However, more animal data is required to clarify the extent of any potential beneficial effect. Clearly, a significant improvement in regeneration would be required to justify an invasive approach.

While no authors have investigated the administration of BMSCs following liver resection as yet, the studies discussed provide initial evidence supporting the role of BMSCs in hepatic regeneration. The range of protocols used raises numerous questions as to the appropriate BMSC isolation and administration techniques, the fate of infused cells, and trial design.

Isolating Bone Marrow-Derived Stem Cells

BMSCs can be harvested either by direct bone marrow aspiration, or collection from peripheral blood following mobilization with granulocyte colony-stimulating factor (GCSF). It has been suggested that GCSF administration alone may itself provide a sufficient increase in circulating BMSCs to promote growth in regenerating tissues [46,47]. However, a major concern with GCSF administration in patients undergoing liver resection for malignancy is the potential for GCSF to promote tumor growth at both intra- and extrahepatic sites [48]. In addition, cases of spontaneous splenic rupture have been reported following GCSF administration even in healthy individuals [49] adding a further concern to this method of cell harvest. Direct bone marrow aspiration would avoid these problems and to minimize discomfort could be performed under general anesthetic immediately prior to hepatic resection. It should be noted that some patients with a background of chronic liver failure may have suppressed bone marrow function and as such may be unsuitable for a major operation let-alone preoperative bone marrow harvest in addition. If despite impaired bone marrow function, the patient is deemed suitable to undergo liver resection, BMSC harvest and administration may still be possible although this should be reviewed on an individual patient basis.

Following BMSC harvest, appropriate cell types must be isolated using suitable stem cell markers. A range of HSC cell subtypes enriched for stem cells have been identified, most consistently possessing the classic HSC cell surface antigens CD34, CD133, and c-kit [50]. Flow cytometry and magnetic-activated cell sorting have been used effectively for cell selection. It is important to note that these cell surface markers do not represent homogeneous cell populations, rather a heterogeneous mix of immature hematopoietic and endothelial stem cells with continually and reversibly changing phenotype depending on the state of activation [50,51]. The dynamic expression of cell surface antigens may limit reliable stem cell identification, necessitating development of alternative techniques, perhaps based on characteristics of stem cell function. Selecting stem cells based on functional characteristics should facilitate delivery of a more reliable HSC population. To this end specific metabolic markers are being pursued, with aldehyde dehydrogenase a promising target given its expression in early immature cells [52]. Exploiting the quiescent nature of stem cells may prove an additional identification tool, where fluorescent-activated cell sorting purification strategies aim to select stem cells by identifying G0-related intracellular proteins such as p27^Kip1 and p21^waf1 [51,53]. Identifying molecules implicated in stem cell trafficking, such as VEGFR-2, CD184, and CD26 may enable further purification of HSC populations [51].

Administering Bone Marrow-Derived Stem Cells

Direct administration of BMSCs to the liver can be performed either through portal vein or hepatic artery infusion. In animal models, the onset of liver regeneration is first noticed in cells surrounding the portal vein of liver lobules [54]. High first pass engraftment following portal vein infusion of BMSCs has been reported in rodents with the majority of peripherally administered stem cells found in the vicinity of portal tract areas [55]. These findings suggest that the portal tract may be a major BMSC entry route into the liver indicating portal vein administration as an appropriate starting point.

Much experience has been gained in the delivery of other therapies into the portal vein, such as chemoembolization. Hepatic surgery presents a challenge with the risk of an additional invasive procedure in the early postoperative period and risk of renal failure caused by contrast agents used to identify hepatic vasculature. However, chemotherapy has been delivered to the liver via the portal vein in several trials during the postoperative period without complication [56]. The AXIS trial included 3,583 patients and assessed the efficacy of portal vein chemotherapy following resection of colorectal malignancy. Here, a catheter was inserted into the portal vein via a small mesenteric vein at the time of surgery and chemotherapy infused early in the postoperative period [57]. No complications directly relating to this technique were reported. This method of intraoperative portal vein cannulation could prove an effective method of delivering BMSCs via the portal vein in the early postoperative period, minimizing risks of additional intervention.

The majority of clinical investigations administering BMSCs in hepatic disease have used the hepatic artery as entry route to the liver, with just 2 studies using the portal vein (Table 1). The justification for choice of route was not discussed in any of the studies although a number of practical reasons for greater use of the hepatic artery administration route may be suggested. The 2 studies employing portal vein cannulation used either transileocolic or percutaneous transhepatic approaches to the portal vein. The transileocolic portal venous approach requires general anesthesia and laparotomy to identify and cannulate a venous tributary. The percutaneous transhepatic approach can be performed under local anesthesia but is a technically demanding procedure, requiring expertise of an interventional radiologist dedicated to hepatobiliary procedures [58]. Catheterization of the hepatic artery may be performed either using a transfemoral or transradial artery approach with advancement to the hepatic artery under fluoroscopic control [59]. Hepatic artery cannulation is less technically demanding than portal vein cannulation and can be performed by most general interventional radiologists. In addition, institutions may be familiar with hepatic artery cannulation in providing targeted chemotherapy to liver tumors, so favoring this technique. Out with these practical issues, the most efficacious BMSC administration route has yet to be demonstrated. Further investigation of this issue is awaited. Incidentally, the hepatic artery is preferred for targeted chemotherapy administration as tumor blood supply is derived principally from the hepatic artery [60].

The Fate of Infused Cells

A major barrier to delivering cell therapies is the current inability to effectively track transplanted or infused cells. The ability to noninvasively monitor the homing of stem cells to target tissues using an imaging modality that has near single cell resolution is crucial to determining engraftment efficiency and functional capability.

Tagging stem cells with contrast agents effectively turns the cells into probes for different imaging modalities. Preclinical studies have used reporter genes such as firefly luciferase or green fluorescent protein to label BMSCs [61,62]. These reporter genes can then be detected by sensitive imaging devices such as the optical coupled device, single-photon emission computed tomography (CT), or positron emission tomography (PET). While these labels prove sensitive methods of stem cell monitoring, they are unlikely to have a role in clinical studies given potential complications following genetic manipulation of cells, their immunogenic nature, and the possibility of fluorescence uptake into other cells following cell death [63].

Clinically applicable imaging modalities include magnetic resonance imaging (MRI), PET, and CT. MRI provides the most sensitive imaging modality currently available in this setting with superparamagnetic iron oxide particles perhaps the most promising marker [63,64]. Such compounds consist of an iron oxide core coated with dextran or siloxanes encapsulated by a polymer [65]. However, these markers are diluted with cell division and there is the potential for transfer to non-stem cells [66]. The resulting hypointense signal may be difficult to distinguish from surrounding air, hemorrhage, necrosis, and macrophages. Labels used in PET scanning require genetic modification of the cell allowing effective observation of stem cells and their progeny. However, genetic labeling risks altering cell properties leaving this technique potentially unsuitable for clinical trials [66]. CT imaging is readily available but the high concentrations of high-density/high anatomic number materials required to create a signal above background signals mean X-ray-based modalities are unlikely to feature in stem cell tracking [63].

There is clearly a considerable gap between what is possible in cell tracking in animal models and what is possible in human studies. Currently, MRI with superparamagnetic iron oxide particles appears to offer the best stem cell tracking opportunity for clinical investigation and while tracking progeny is limited, the technique should enable identification of engraftment site. The technique is not yet licensed for tracking stem cells in clinical trials but is licensed for non-stem cell applications in the USA.

Trial Design and Assessing Efficacy

While many concerns regarding stem cells from other sources, such as the embryo, may restrict translation to clinical trials, the use of non-modified autologous cells should facilitate approval from regulatory authorities. There is now a growing base of phase I trial evidence demonstrating the feasibility and suggesting the safety of autologous BMSC transplantation in hepatic disease. More comprehensive research in other fields such as cardiology provides further evidence of the safety of BMSC administration and potential efficacy. Many trial design issues, relevant to both liver resection and chronic liver disease, are yet to be addressed however.

Early clinical investigations have relied on indicators suggestive of hepatic regeneration or reduced fibrosis such as clinical chemistry and volumetric assessment as primary outcome measures (Table 1). While providing some insight into the potential efficacy of stem cell interventions such trials do not provide any indication as to the role of BMSCs in hepatic regeneration. Looking to the future, effective cell tracking is likely to be the most important innovation in assessing contribution of cell therapies to hepatic regeneration and appropriate techniques are eagerly awaited.

Dose response to stem cell therapies has yet to be addressed. As noted in studies in chronic liver disease, a transient response may necessitate repeated infusions, equally, over zealous administration risks complications, such as portal hypertension or fibrosis. Clinical trials to date have employed single infusions of variable quantities of selective or purified BMSCs, in line with their feasibility assessment of the technique. Future studies will no doubt assess quantity of BMSCs required to promote adequate hepatic regeneration while appreciating potential complications. Simultaneous infusion of factors to promote BMSC uptake or homing to the liver may improve efficiency but could also promote tumor growth and should be avoided in the context of liver resection.

Appropriately powered, randomized controlled trials are indicated to further elucidate the potential benefit of BMSC application following hepatic resection. However, given the invasive nature of procedures required for BMSC isolation and administration such trials may be ethically questionable at present. Indeed, such human trials must be preceded by studies in directly relevant preclinical models to further ascertain the potential wider effects of BMSC therapy following liver resection in humans. Ultimately, given the invasive nature of BMSC harvest and administration benefits must be considerable to warrant inclusion in the future management of patients following liver resection.

Conclusion

Autologous BMSC therapies may provide a targeted therapy to enhance hepatic regeneration following liver resection, potentially minimizing the risks of this procedure. Techniques involving BMSC harvest, purification, and administration have been demonstrated for other indications and should be readily transferable to deliver BMSC therapies after liver resection. Cautious clinical studies are indicated to further elucidate the potential benefit in hepatic surgery.