Hmgb3 Regulates the Balance between HSC Differentiation and Self Renewal.

Hmgb3 is an X-linked member of a family of sequence-independent chromatin-binding proteins that is preferentially expressed in hematopoietic stem cells (HSC; lin⁻, c-kit^HI, Sca-1^HI, IL-7Rα⁻). Hmgb3 deficient mice (Hmgb3^−/Y: KO) contain normal numbers of HSCs, capable of repopulation and self-renewal, but fewer common lymphoid (CLP: lin⁻, c-kit^LO, Sca-1^LO, IL-7Rα⁺) and myeloid (CMP: lin⁻, c-kit⁺, Sca-1⁻, IL-7Rα⁻) progenitors (Nemeth et al. Blood, 2005). We hypothesized this reduction may be due to impaired production of CLP and CMP by KO HSCs or by rapid differentiation of KO CLP and CMP. CMP were isolated from littermate control (WT) and KO mice. The initial CFU number was compared to CFU numbers from CMP cultured for 5 days in serum-free media containing SCF, Flt3L, IL-3, IL-6 and TPO. WT CMP (n = 12) increased 2.7 ± 1.8-fold in CFU-GM/1000 CMP between d0 and d5. KO CMP (n = 12) increased 4.9 ± 1.9-fold between d0 and d5 (p = .16, ns). We conclude that WT and KO CMP are functionally similar. To test whether reduced CMP levels in KO mice are due to decreased CMP production, we compared the effect of 5-fluorouracil (5-FU) on WT and KO HSC, CLP, and CMP populations. FACS analysis showed that lin− cells from untreated WT mice were 0.14% HSC, 0.10 % CLP and 1.2% CMP compared to 0.13% HSC, 0.06% CLP, and 0.84% CMP in KO mice. 2 days post-injection (p.i.) of 150 mg/kg 5-FU, lin− cells from WT mice were 0% HSC, 0.12% CLP, and 1.4% CMP compared to 0.01%, 0.09%, and 1.1% respectively in KO mice. 4 days p.i., WT mice had 0% HSC, 0.25% CLP, and 0.03% CMP, but KO mice showed a recovery of HSC (1.04 ± 0.34%; p < .001), CLP (0.68%) and CMP (2.8%). 8 days later, WT mice recovered comparable levels of HSC (1.06%) CLP (0.19%) and CMP (0.44%). We analyzed repopulating activity by transplanting 1 x 10⁴ lin⁻, Sca-1⁺, IL-7Rα⁻, c-kit^HI or c-kit^NEG cells (Ly 5.2) from untreated and 4-day p.i. WT or KO mice with 1 x 10⁶ congenic (Ly 5.1) bone marrow cells into Ly5.1/5.2 recipients. 12 weeks after transplant, untreated WT and KO c-kit^HI cells repopulated recipients (> 90% Ly 5.2) but not untreated c-kit^NEG cells. In agreement with previous results (Randall et al., Blood, 1997), recipients of c-kit^NEG cells from 4-day p.i. WT bone marrow contained 78.4 ± 18.0% Ly5.2 cells (n = 6), representing all of the repopulating ability as c-kit^HI cells from 4-day p.i. WT mice were absent. In contrast, recipients of either KO c-kit^NEG cells (86.3 ± 7.2% Ly 5.2; n = 7) or ckit^HI cells (82.8 ± 6.9% Ly5.2; n = 6) from 4-day p.i. KO mice showed long-term repopulation. Because c-kit cannot be used as an HSC marker in post 5-FU bone marrow, we isolated lin⁻, Sca-1⁺, IL-7Rα⁻, side-population (SP)⁺ cells to analyze the cell cycle status of WT and KO HSCs 2 and 4 days p.i. 14.0% of untreated WT and 13.3% of KO HSCs are in the S/G2/M phases (Nemeth et al., Blood, 2005). 2 days p.i. 32.3% of WT SP⁺ HSCs and 31.9% of KO SP⁺ HSCs were in the S/G2/M phases. 4 days p.i., 21.2 ± 0.2% of WT SP⁺ HSCs were cycling while only 12.4 ± 2.2% of KO SP⁺ HSCs were cycling (p < .01), leading us to hypothesize that KO HSCs have recovered and resumed steady-state cycling levels. To determine whether Wnt signaling regulated KO HSC self-renewal, KO mice were mated with mice containing a lacZ reporter gene for the Wnt pathway (TOPGAL). We observed a 2.8-fold increase in lacZ expression in steady-state HSCs isolated from TOPGAL x KO mice compared to TOPGAL controls. We conclude that Hmgb3 deficiency allows for rapid HSC self-renewal, corresponding with increased Wnt signaling, and that Hmgb3 regulates the balance between the production of CLP and CMP and self-renewal.