A novel role for the nervous system in HSPS mobilization

Hematopoietic stem and progenitor cell (HSPC) transplantation is a vital tool in the treatment of hematopoietic and other malignancies. HSPC reside in the bone marrow (BM), and can be isolated directly from BM or from peripheral blood after a process termed mobilization: molecular cues are used to coax the HSPC out of their niche and into the circulation. A number of factors are known to induce mobilization, including a variety of cytokines and chemokines. The most commonly used clinical regimen is multi-day treatment with granulocyte-colony stimulating factor (G-CSF). However, there is a wide range of efficacy among different individuals, and the mechanism by which G-CSF induces mobilization is not well understood.

Cell adhesion molecules, proteases, chemokines, and cytokines have all been shown to play roles in regulating HSPC homing to, and egress from, the BM niche.1,2 Specifically, a key role for the CXCL12/CXCR4 axis has been established by a number of studies.3-6 CXCL12, also known as SDF-1, acts as a chemoattractant for HSPCs and is downregulated in the BM environment in response to G-CSF administration.3,7 In addition, studies have demonstrated that proteolytic activity induced by G-CSF treatment degrades CXCL12, and this degradation correlates with HSPC egress from BM.3,4 Cell adhesion molecules including integrins, VCAM-1, and selectins have all been shown to influence HSPC mobilization and/or adhesion to BM components.1,2 P/E-selectin double mutant mice exhibit defective homing to BM and increased HSPC numbers in peripheral blood. Inhibitors of selectins, including sulfated glycans such as fucoidan, a polysaccharide isolated from brown seaweed, also promote mobilization of HSPC.8,9 While fucoidan isn’t found in mammalian cells, a functionally similar molecule, sulfatide (sulfated galactosylceramide), is produced by human myeloid cells.

Interest in whether sulfatide might play a role in normal HSPC trafficking led Katayama et al. to examine mobilization in mice lacking UDP-galactose: ceramide galactosyltransferase (Cgt), a gene required for sulfatide production. Consistent with the authors’ hypothesis, Cgt-/- mice fail to undergo significant mobilization of HSPC in response to G-CSF and fucoidan. However, the reason for this result is surprising.

Although proteolysis was normal in Cgt-/- BM, differences in CXCL12 protein and RNA levels in bone tissue were detected in Cgt-/- mice. The authors attributed the failure of HSPC mobilization to the persistence of high levels of CXCL12 despite treatment with G-CSF. Osteoblast activity was also altered in Cgt-/- mice: cells were flatter and had shorter cell protrusions into bone tissue. Interestingly, G-CSF treatment was noted to result in similar changes in osteoblasts of wild-type mice.

The novel result came when the authors considered a potential role for the nervous system in the Cgt -/- mobilization phenotype. Besides sulfatide, Cgt is essential for production of galactocerebrosides, major components of the myelin sheaths that surround nerve fibers and facilitate nerve impulse conduction. Indeed, Cgt-/- mice die early in post-natal life due to severe tremor and ataxia.10,11 A series of experiments provided support for the involvement of the sympathetic nervous system (SNS) in control of HSPC mobilization. First, it was demonstrated that mice deficient in SNS function due to knockout of dopamine β hydroxylase (Dβh), the enzyme required for the production of norepinephrine, show a dramatic reduction in HSPC mobilization by G-CSF. Furthermore, osteoblast morphology and CXCL12 levels were not affected by G-CSF administration in the Dβh-/- mice, implying that G-CSF is unable to exert its normal effects in the absence of adrenergic signals from the SNS. Finally, treatment of wild-type mice with a β adrenergic antagonist, the beta blocker propranolol, also reduced HSPC mobilization, while the β2 adrenergic agonist clenbuterol partially rescued G-CSF induced mobilization in Dβh-/- mice.

Figure 1. Model depicting the role of the sympathetic nervous system (SNS) in HSPC mobilization from bone marrow. Prior to mobilization by G-CSF (left side) osteoblasts display normal morphology and produce CXCL12, which retains HSPC in the bone marrow niche. G-CSF (right side) decreases CXCL12 levels in bone marrow by causing release of proteases by neutrophils and decreased synthesis of CXCL12 by osteoblasts. The effect of G-CSF on osteoblasts is mediated in part by adrenergic neurons, based on the finding that knockout mice with inadequate myelination (Cgt-/- mice), or lack of norepinephrine (Dbh-/- mice), fail to respond to G-CSF.

Importantly, adrenergic signaling alone is not sufficient to induce HSPC egress from the BM: treatment of wild-type mice with clenbuterol did not cause mobilization. The authors suggest that G-CSF causes HSPC mobilization by inhibiting osteoblast production of CXCL12 through both SNS-dependent and independent means (Figure 1). Their results raise the interesting possibility that differences in sympathetic tone may explain variable responses to G-CSF, and suggest a potential clinical avenue for optimization of HSPC mobilization from patients and donors.


  1. Lapidot, T. & I. Petit (2002) Exp. Hematol. 30:973.
  2. Papayannopoulou, T. (2004) Blood 103:1580.
  3. Petit, I. et al. (2002) Nat. Immunol. 3:687.
  4. Wright, D.E. et al. (2002) J. Exp. Med. 195:1145.
  5. Levesque, J.-P. et al. (2003) J. Clin. Invest. 111:187.
  6. Broxmeyer, H.E. et al. (2005) J. Exp. Med. 207:1307.
  7. Katayama, Y. et al. (2006) Cell 124:407.
  8. Sweeney, E.A. et al. (2000) Proc. Natl. Acad. Sci. USA 97:6544.
  9. Frenette, P.S. & L. Weiss (2000) Blood 96:2460.
  10. Bosio, A. et al. (1996) Proc. Natl. Acad. Sci. USA 93:13280.
  11. Coetzee, T. et al. (1996) Cell 86:209.