Trouble in the niche? Send in a statin

In this issue of Blood, Shi et al report on functional deficits of marrow endothelial progenitor cells (EPCs) in patients with poor graft function (PGF) after allotransplant which could be improved in vitro with atorvastatin exposure.¹ 

PGF is estimated to occur in about 5% to 27% of allografts and contributes to morbidity and mortality.²  PGF is to be distinguished from graft failure, which is often due to a low transplanted nucleated cell dose or to alloreactive immune responses mediated by residual host immunity. Graft failure is therefore seen most commonly in cases of HLA disparity between donor and host.³  In contrast to graft failure, PGF is characterized by full donor chimerism, and it can also be primary or secondary. It is often associated with postallograft effects of viral infections, of conditioning regimens, of drugs toxic to marrow, or of graft-versus-host disease (GVHD). In some cases, it has been associated with inflammatory mediators such as interferon-γ or tumor necrosis factor-α, which could also impact cells of the marrow microenvironment as well as hematopoietic stem and progenitor cells.² 

Shi et al have expanded upon their previous work⁴  which demonstrated reduced numbers of marrow EPCs in PGF cases in an attempt to define functional changes in EPCs and possible ways to overcome quantitative and functional deficits. Twenty-six cases of PGF with 100% donor chimerism were identified from 578 allogeneic transplants performed at a single center, and matched controls with good graft function (GGF) were selected from the same recipient cohort using case-control sampling with matching for pertinent variables. Strengths of this work are that GGF and PGF were carefully defined, well-matched contemporary controls were used, and patients with relapse or severe acute or chronic GVHD were excluded.

EPCs were isolated from each of these cohorts from light-density marrow cells cultured in medium supportive of endothelial cells for 7 days and identified through CD34, CD133, and CD309 (vascular endothelial growth factor receptor-2 [VEGFR-2]) expression. EPCs from patients with PGF had fewer cells with expression of Dil-acetylated low-density lipoprotein and Ulex europaeus agglutinin-1, and migration and tube formation capabilities were reduced. Intracellular reactive oxygen species (ROS) expression and apoptosis were higher in those EPCs from PGF vs GGF patients.

Given the functional alterations noted in EPCs from subjects with PGF, Shi et al took cues from prior work demonstrating that atorvastatin, a 3-hydroxy-3-methylglutaryl coenzyme A (HMGCoA) reductase inhibitor is able to improve function of EPCs in other vascular diseases⁵  and that p38 MAPK, one of the family of the mitogen-activated serine/threonine protein kinases, can regulate EPC dysfunction and can be modulated by HMGCoA reductase inhibitors such as statins.⁶  They found higher expression of phospho-p38 MAPK and its downstream target phospho-cyclic adenosine monophosphate–responsive element-binding protein (phospho-CREB) in EPCs from PGF vs GGF patients, and that atorvastatin, p38 inhibitors, and ROS inhibitors were able to improve the number and function of EPCs in PGF subjects. Furthermore, atorvastatin exposure improved colony-forming unit outgrowth from normal CD34⁺ cells after coculture with EPCs from PGF subjects although this increase did not reach statistical significance, and it reduced phospho-p38 MAPK and phospho-CREB in EPCs from PGF patients (see figure).

Although these in vitro observations offer insight into microenvironment changes associated with well-characterized PGF cases, they are no doubt just a beginning. It is likely that other niche cells with hematopoietic support function such as osteoblasts, adipocytes, or mesenchymal stem cells might also be functionally altered.⁷  The EPCs examined here were grown in culture for 7 days, so it is possible that they may not be reflective of freshly isolated or in situ EPCs. There are data from both human and murine systems that demonstrate a role for EPCs in support of hematopoietic stem cells (HSCs), but unlike the case in murine marrow where EPCs may be perivascular or sinusoidal,⁸  in human marrow, it is less certain where these cells reside or how they interact with hematopoietic and other niche cell types. Although atorvastatin improved the support ability of EPCs in terms of clonogenic progenitor output, this was not statistically significant, and more needs to be done to determine the effects of functionally aberrant EPCs on hematopoietic stem and progenitor cells as well as other immune cells of the niche.

The p38 MAPK signaling pathway can be stimulated by a variety of external signals including growth factors, environmental stress, and inflammatory cytokines, and its activation can generate a multitude of biological effects.⁹  Modulation of its activity in EPCs and other cells of the marrow microenvironment by atorvastatin might therefore have multiple effects on the immune response and on cell survival and differentiation. Furthermore, it interplays with other signaling pathways, and a broader examination of what other genes and pathways are differentially expressed between GGF and PGF EPCs would be important to discern through RNA sequencing or other nonbiased methods.

The current article is not able to answer whether the dysfunction in EPCs is due to direct toxicity from conditioning regimens, immunological effects, or to an inflammatory state established in the marrow due to HSC stress,²  undiagnosed viral infections, or subclinical GVHD. Donor-specific antibodies directed against CD34⁺/VEGFR-2⁺ EPCs have been reported, so this might also contribute to EPC dysfunction.¹⁰ 

Although graft failure can sometimes be treated with a stem cell boost or second transplant, in cases of PGF where donor chimerism is complete, other treatment modalities are needed. Hematopoietic growth factor administration is often ineffective, and immune-modulatory therapies have also met with limited success thus far. The observation of Shi et al that the EPC component of the niche is dysfunctional and can be modulated through inhibition of the p38 MAPK pathway via atorvastatin is a beginning step toward understanding this multifaceted problem and one that offers a potential therapeutic approach for future examination.