Hematopoietic stem and progenitor cell (HSPC) mobilization by G-CSF is reduced in CD26 knockout (−/−) mice or mice treated with compounds that inhibit CD26 enzyme activity. However, how CD26 regulates mobilization is not known. CD26 is a surface dipeptidyl peptidase expressed on hematopoietic and non-hematopoietic cells. The fact that CD26 is expressed on HSPC, has suggested that intrinsic HSPC CD26 is involved in their mobilization. We reasoned that if intrinsic HSPC CD26 is crucial for their mobilization, then CD26+ HSPC would mobilize more readily than CD26− HSPC. After G-CSF treatment, CD26+ SLAM LSK and LSK cells increased 2 and 3.7-fold in peripheral blood (PB) whereas CD26− SLAM LSK and LSK cells increased 3.9 and 8.7 fold respectively; with the ratio of CD26+/CD26− HSPC in the bone marrow (BM) remaining unchanged. In chimeric mice where hematopoietic cells were either wild-type or CD26−/− in origin and the stromal environment was wild-type, equivalent CFU-GM and LSK cells were mobilized by G-CSF. These studies suggest that CD26 expression on hematopoietic origin cells including HSPC is not required for the G-CSF induced mobilization of HSPC.
The BM stromal environment includes both osteolineage cells and endothelial cells (EC). To explore whether osteolineage cell CD26 is involved in G-CSF mobilization, we quantitated CD45−CD119−CD31−Sca-1+ Alcam− cells enriched for mesenchymal progenitor cells (MPC) and CD45−CD119−CD31− Sca-1- Alcam+ cells enriched for osteoblasts (OB) in the BM of mice treated with G-CSF or G-CSF plus the CD26 inhibitor diprotin A. Total MPC and OB number as well as the retention factors SDF-1, SCF-1 and osteoprotegerin (OPG) were reduced equally in the BM of G-CSF and G-CSF plus diprotin A treated mice, suggesting that osteolineage cell CD26 is not involved in mobilization. Since endothelial cells (EC) express CD26 and HSPC resides proximal to sinusoidal endothelial cells (SEC), we evaluated if SEC CD26 has a role in G-CSF induced HSPC mobilization. G-CSF treatment increased CD26 expression on CD45- Ter119− Sca-1− VEGFR3+ SEC by 2-fold. Because mobilized HSPC must transmigrate across SEC to enter the PB, we determined if EC CD26 regulates HSPC transmigration. Treatment of confluent EC monolayers with G-CSF in vitro, increased transendothelial migration of both wild-type and CD26−/−HSPC, which was significantly reduced by diprotin A, supporting the hypothesis that enhanced expression of CD26 on BM SEC by G-CSF is involved in HSPC mobilization.
Since SDF-1 is a BM retention factor for HSPC and CD26 can cleave SDF-1 in vitro, it is possible that SDF-1 degradation by CD26 plays a role in HSPC egress. However, direct evidence supporting proteolytic inactivation of SDF-1 by CD26 in response to G-CSF is not available. Using a novel bio-binding assay and mass spectroscopy, we found similar levels of active SDF-1 in the BM extracellular fluid and on SEC membranes from G-CSF treated wild-type and CD26−/− mice, suggesting that G-CSF mobilization is not linked to SDF-1degradation. We therefore searched protein sequence databases for known EC products containing a CD26 recognition site and implicated in HSPC trafficking. We found that the neurotransmitter neuropeptide Y (NPY), a sympathetic nerve and EC product contains an N-terminal Tyr-Pro sequence and can be cleaved to NPY3–36 by CD26. We hypothesized that if cleavage of NPY by EC CD26 is involved in transendothelial migration of HSPC, then reduced HSPC migration through EC monolayers in the presence of G-CSF plus diprotin A would be reversed in the presence of NPY3–36. As expected, NPY3–36 reversed the diprotin A mediated attenuation in HSPC migration through G-CSF treated EC. In vivo, the attenuated response to G-CSF mobilization in diprotin A treated or CD26−/− mice could be completely reversed by co-administration of NPY3–36, but not full length NPY. In G-CSF mobilized wild-type mice, we found that expression of gap junction proteins on BM SEC, including CD31 and VE-cadherin are reduced coincident with increased EC permeability, which is attenuated by co-administration of diptotin A and restored to normal in mice receiving G-CSF plus diprotin A and NPY3–36, but not full length NPY. In conclusion, we have found that CD26 regulates G-CSF induced mobilization that is independent of the SDF-1 paradigm; rather elevated SEC CD26 in response to G-CSF enhances transendothelial migration and mobilization of HSPC via cleavage of NPY.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.