TiF1-gamma plays an essential role in murine hematopoiesis and regulates transcriptional elongation of erythroid genes

Abstract

Transcriptional regulators play critical roles in the regulation of cell fate during hematopoiesis. Previous studies in zebrafish have identified an essential role for the transcriptional intermediary factor TIF1γ in erythropoiesis by regulating the transcription elongation of erythroid genes. To study if TIF1γ plays a similar role in murine erythropoiesis and to assess its function in other blood lineages, we generated mouse models with hematopoietic deletion of TIF1γ. Our results showed a block in erythroid maturation in the bone marrow following tif1γ deletion that was compensated with enhanced spleen erythropoiesis. Further analyses revealed a defect in transcription elongation of erythroid genes in the bone marrow. In addition, loss of TIF1γ resulted in defects in other blood compartments, including a profound loss of B cells, a dramatic expansion of granulocytes and decreased HSC function. TIF1γ exerts its functions in a cell-autonomous manner as revealed by competitive transplantation experiments. Our study therefore demonstrates that TIF1γ plays essential roles in multiple murine blood lineages and that its function in transcription elongation is evolutionally conserved.

Figure 1. Abnormal erythropoiesis in TIF1γ-deficient mice

Erythroid cell maturation was measured by FACS based on the expression of the cell surface markers, c-Kit, CD71 and Ter119(Mx-cre for A&B, Vav-cre for D&E). The c-Kit^hiCD71⁺ Ter119⁻ cells represent the earlyerythroidprogenitors(see the cytospin staining in supplemental figure 1C). Populations R1-R4 represent the progressive maturation of cells. Bar graphs represent the absolute number of cells in each population. Representative FACS plots were shown in A&B. (A, D) Bone marrow cells. (B, E) spleen cells. The results are shown as mean ± SD from 3 or 4 mice in each group (* q ≤ 0.05, **q≤ 0.005). (C) BFU-E colony assay using spleen cells from Mx-cre mice (n=3). The results are shown as mean ± SD. (F) Western blots comparing the protein level of phosphorylated Smad5 (top) and total Smad5 (middle) in R2 cells between vav-cre; tif1γ^fl/+ (con) mice and vav-cre; tif1γ^fl/fl (KO) mice. Western blot for b-actin was used as a loading control. The experiments were repeated twice using independent groups of mice. (G) Real-time RT-PCR analyses for selected erythroid genes in R2 stage of spleen erythroid cells from vav-cre; tif1γ^fl/fl (KO) mice and vav-cre; tif1γ^fl/+ (con) mice. Results are shown as fold changes (KO vs. con), normalized to the level of b-actin. Error bars represent mean ± SD from three independent experiments.

Figure 1. Abnormal erythropoiesis in TIF1γ-deficient mice

Erythroid cell maturation was measured by FACS based on the expression of the cell surface markers, c-Kit, CD71 and Ter119(Mx-cre for A&B, Vav-cre for D&E). The c-Kit^hiCD71⁺ Ter119⁻ cells represent the earlyerythroidprogenitors(see the cytospin staining in supplemental figure 1C). Populations R1-R4 represent the progressive maturation of cells. Bar graphs represent the absolute number of cells in each population. Representative FACS plots were shown in A&B. (A, D) Bone marrow cells. (B, E) spleen cells. The results are shown as mean ± SD from 3 or 4 mice in each group (* q ≤ 0.05, **q≤ 0.005). (C) BFU-E colony assay using spleen cells from Mx-cre mice (n=3). The results are shown as mean ± SD. (F) Western blots comparing the protein level of phosphorylated Smad5 (top) and total Smad5 (middle) in R2 cells between vav-cre; tif1γ^fl/+ (con) mice and vav-cre; tif1γ^fl/fl (KO) mice. Western blot for b-actin was used as a loading control. The experiments were repeated twice using independent groups of mice. (G) Real-time RT-PCR analyses for selected erythroid genes in R2 stage of spleen erythroid cells from vav-cre; tif1γ^fl/fl (KO) mice and vav-cre; tif1γ^fl/+ (con) mice. Results are shown as fold changes (KO vs. con), normalized to the level of b-actin. Error bars represent mean ± SD from three independent experiments.

Figure 2. TIF1γ deficiency leads to loss of B cells

(A & B) Bone marrow cells from Mx-cre; tif1γ^fl/+ (con) and Mx-cre; tif1γ^fl/fl (KO) mice were analyzed by FACS at 3-week post pI-pC injection. (A) The absolute numbers of cells in each population are shown. The results are shown as mean ± SD from 4 mice in each group (* q ≤ 0.01, ** q ≤ 0.001). (B) Representative FACS plots using B cell markers with the percentage of each cell population. (C) Real-time RT-PCR analyses for selected genes in pro-B and pre-B cells from vav-cre; tif1γ^fl/fl (KO) mice and vav-cre; tif1γ^fl/+ (con) mice. Results are shown as fold changes (KO vs. con), normalized to the level of bactin. Error bars represent mean ± SD from three independent experiments.

Figure 3. Accumulation of granulocytes in TIF1γ-KO mouse

(A) Representative FACS plots using myeloid cell markers with the percentage of each cell population. (B) Splenomegaly of 4-month old vav-cre; tif1γ^fl/fl (KO) adult mice (E) Accumulated myeloid cells in vav-cre; tif1γ^fl/fl adult mice in peripheral blood (PB, MGG staining), spleen and liver (H&E staining).

Figure 4. TIF1γ deficiency promotes myelopoiesis while inhibiting erythropoiesis

(A) Increased number of GMPs in vav-cre; tif1γ^fl/fl mice at 3-month old. The absolute numbers of cells are shown as mean ± SD from 4 mice in each group (** q ≤ 0.001). (B) Heat map showing regulation of genes representative of the erythroid signature and the myeloid signature from GMPs derived from vav-cre; tif1γ^fl/+ (con) and vav-cre; tif1γ^fl/fl (KO) mice. (C) Increased HSC levels in vav-cre; tif1γ^fl/fl mice at 3-month old. The absolute numbers of cells are shown as mean ± SD from 4 mice in each group (* q ≤ 0.05).

Figure 5. TIF1γ has cell-autonomous functions in the regulation of hematopoiesis

(A) Scheme of the competitive transplantation assay (B) Percentage of CD45.2⁺ cells in the peripheral blood of recipients one-week after pI-pC injection. The results are shown as mean ± SD from 9 mice in each group (** q ≤ 0.001). (C, D & E) FACS analyses to measure the numbers of CD45.2⁺ cells in each cell population of the recipient bone marrow at 18-week post pI-pC injection. The results are shown as mean ± SD from 4 mice in each group(* q ≤ 0.01, ** q ≤ 0.001).

Figure 5. TIF1γ regulates the transcription elongation of erythroid genes

(A) Quantitative (real-time) RT-PCR comparing the transcription levels of selected genes in R2 stage of erythroid cells between vav-cre; tif1γ^fl/fl (KO) mice and vav-cre; tif1γ^fl/+ (con) mice. Top: A schematic diagram showing the position of primers used in real-time RT- PCR analyses. Primers to detect the 5’ ends of transcripts are located within 120bp from transcription start site, and primers for the 3’ ends of transcripts are in the 3’ coding region or 3’UTR. Bottom: real-time RT-PCR analyses to compare the 5’ transcripts (gray) and 3’ transcripts (blue). Results are shown as fold changes (KO vs. con), normalized to the ratio of the 5’ transcript level of β-actin (ActB). Error bars represent mean ± SD from three independent experiments. Also see Supplemental Figure 4. (B) Western blots comparing the protein level of serine 2-phophoralated Pol II (top), serine 5-phophoralated Pol II (middle) and total Pol II in R2 cells between vav-cre; tif1γ^fl/+ (con) mice and vav-cre; tif1γ^fl/fl (KO) mice. Cells were pooled from 2 mice in each group. Western blots for the TATA binding protein (TBP, bottom) was used as a loading control. The experiments were repeated twice using independent groups of mice.