TGFbeta1 stimulates the over-production of white matter astrocytes from precursors of the "brain marrow" in a rodent model of neonatal encephalopathy

Abstract

Background: In children born prematurely and those surviving cerebral ischemia there are white matter abnormalities that correlate with neurological dysfunction. Since this injury occurs in the immature brain, when the majority of subventricular zone (SVZ) cells generate white matter oligodendrocytes, we sought to study the effect this injury has on gliogenesis from the SVZ. We hypothesized that there is aberrant glial cell generation from the SVZ after neonatal hypoxia ischemia (H/I) that contributes to an increased astrogliogenesis with concomitant oligodendroglial insufficiency. Mechanistically we hypothesized that an increase in specific locally produced cytokines during recovery from injury were modifying the differentiation of glial progenitors towards astrocytes at the expense of the more developmentally-appropriate oligodendrocytes.

Methodology/principal finding: For these studies we used the Vannucci H/I rat model where P6 rats are subjected to unilateral common carotid ligation followed by 75 min of systemic hypoxia. Retroviral lineage tracing studies combined with morphological and immunohistochemical analyses revealed the preferential generation of SVZ-derived white matter astrocytes instead of oligodendrocytes post hypoxia/ischemia. Microarray and QRT-PCR analyses of the damaged SVZ showed increased expression of several cytokines and receptors that are known to promote astrocyte differentiation, such as EGF, LIF and TGFbeta1 signaling components. Using gliospheres to model the neonatal SVZ, we evaluated the effects of these cytokines on signal transduction pathways regulating astrocyte generation, proliferation and differentiation. These studies demonstrated that combinations of EGF, LIF and TGFbeta1 reconstituted the increased astrogliogenesis. TGFbeta1-induced Smad 2/3 phosphorylation and the combination of EGF, LIF and TGFbeta1 synergistically increased STAT3 phosphorylation over single or double cytokine combinations. Pharmacologically inhibiting ALK5 signaling in vitro antagonized the TGFbeta1-induced increase in astrocyte generation and antagonizing ALK5 signaling in vivo similarly inhibited astrogliogenesis within the SVZ during recovery from H/I.

Conclusion/significance: Altogether, these data indicate that there is aberrant specification of glial precursors within the neonatal SVZ during recovery from neonatal H/I that is a consequence of altered cytokine signaling. Our studies further suggest that antagonizing the ALK5 receptor will restore the normal pattern of cell differentiation after injury to the immature brain.

Figure 1. Over-production of astrocytes in the injured hemisphere at 7 days of recovery as a consequence of neonatal H/I.

(A) Sections from animals sacrificed at 7 days of recovery were stained with antibodies against GFAP (green) and counterstained with DAPI (blue). (B,C) Sections were stained with Rip for oligodendrocytes and myelin in the ILH (C) compared to the CLH (B). (D,E) Sections were stained with antibodies against GFAP (red), S100b (green) and counterstained with DAPI (blue) (E) depicts the ILH; (D) depicts the CLH (D). (F,G) The ILH SVZ (G) was stained with antibodies against vimentin (green); (F) depicts the CLH. (H) Inset depicts BrdU+/Vim+ cells in the ILH SVZ from animals injected with BrdU (red) at day 3 of recovery and analyzed at 7 days of recovery. Scale bar represents 100 µm (A), 50 µm (B,C,F,G), 25 µm (D,E,H).

Figure 2. SVZ cells preferentially generate white matter astrocytes rather than oligodendrocytes after H/I.

Replication-incompetent retroviruses containing the alkaline phosphatase (DAP) reporter gene were stereotactically injected at P8 into the lateral ventricles (two days after H/I). For quantitative analysis at 10 days after DAP injection (A) cell types within the white matter (WM) were classified based upon morphological criteria as either astrocytes (arrow), oligodendrocytes (arrowhead) or indeterminate (B). Data are averaged from 8 hypoxic sham, 8 contralateral hemispheres (CL), and 7 ipsilateral hemispheres (IL) * indicates p<0.05 compared to CLH, ∧ indicates p<0.05 compared to hypoxic sham animals as determined by ANOVA with Fisher's PLSD post-hoc test.

Figure 3. SVZ cells preferentially generate GFAP+ astrocytes in white matter instead of GST π + oligodendrocytes.

Two µL of pNIT replication-incompetent retroviruses containing the GFP reporter gene were stereotactically injected at P8 into the lateral ventricles (two days after H/I). Cryostat sections (15 µm) were prepared after sacrifice at P27. Sections were stained for GFP (A,F), GST π (B,G) and GFAP (C,H) followed by DAPI counterstain (D,I). Merged color overlays are presented in panels E and J. Cell types within the corpus callosum (WM) were counted as GFAP +, GSTπ +, GSTπ weakly + and double negative (indeterminate). Panels A–E illustrate a representative GFAP+ cell, and panels F–J illustrate a GSTπ+ cell. Panel K depicts quantitative data averaged from 3 IL hemispheres and 6 CL hemispheres. * indicates p<0.05 compared to CLH as determined by student's t-test. Scale bar represents 20 µm.

Figure 4. Combinations of EGF, LIF and TGFß1 effectively increase astrocyte generation in gliosphere cells.

Gliospheres were grown in ProN supplemented with B104 conditioned medium and 10 ng/ml FGF-2. They were dissociated and plated onto culture dishes for 6 h in differentiation media with 2% serum and then switched to differentiation media supplemented with either 5 ng/mL EGF (E), LIF (L), TGFß1 (T) or combinations for 48 h. (A) Densitometric analyses of protein levels on Western Blots were performed for GFAP, Zebrin II and GS as indices of astrocyte differentiation. (B) Values represent the mean fold change over controls ± SEM from 3 independent experiments. * and # denote a significant increase and decrease respectively from control (P<0.05).

Figure 5. Differential effect of EGF, LIF and TGFß1 on Smad 2/3 and STAT3 phosphorylation in gliospheres.

Gliospheres were grown in ProN supplemented with B104 conditioned medium and 10 ng/ml FGF-2. They were dissociated and plated onto culture dishes and then treated with 5 ng/mL EGF (E), LIF (L), TGFß1 (T) either alone or combinations in differentiation media for 30 min. (A) Total cell lysates were collected and Western blot analysis was performed to determine levels of total Smad2/3, STAT3, pSmad 2/3 on Ser465/467 and pSTAT3 on Ser727 and Tyr705 residues. (B) Values represent fold change in pSmad2/3 over Smad2/3. (C) Values represent fold change in pSTAT3^Tyr705 over STAT3, ± SEM from 3 independent experiments. * Denote significant increase from control, # significant increase from treatment with a single cytokine or combination of two (P<0.05).

Figure 6. TGFß1 induced increased astrocyte proliferation is mediated through the ALK5 receptor.

Gliospheres were plated onto chamber slides in differentiation media supplemented with 5 ng/mL of EGF (E), LIF (L), and TGFß1 (T) in different combinations for 96 h and were pulsed with BrdU for the last 3 h in culture. The percentage of (A) GFAP+ and (B) GFAP+/BrdU+ cells is depicted. (C,D) Representative images of gliospheres differentiated in EGF (C) or combination of EGF and TGFß1 (D) stained for GFAP (green) and red (BrdU). (E,F) Gliospheres were similarly treated with EGF or a combination of EGF and TGFß1 for 96 hrs and the effect of these cytokines on oligodendrocyte generation were determined by O4 staining. Representative images of gliospheres differentiated in EGF (E) or combination of EGF and TGFß1 (F) stained for GFAP (green) and red (O4) are shown. (G) Gliospheres were propagated in EGF or EGF+ TGFß1 in the presence or absence of 10 µM SB431542, a potent ALK5 inhibitor, pulsed with BrdU and stained for GFAP and GFAP/BrdU (G). Results are representative of 3 independent experiments. * indicate statistically significant differences (p<0.05) for each condition compared to medium alone (M) controls using Student's t-test. The scale bar represents 35 µm for panels C,D and 50 µm for panels E,F.

Figure 7. Increased astrogliosis at 4 days of recovery observed in the injured SVZ after H/I is decreased by ALK5 inhibition in vivo.

(A,B) Representative images of CL (A) and IL (B) hemispheres from age-matched control animals (twice daily i.p. SB505124 injections) and stained with antibodies against GFAP (green) and counterstained with DAPI (blue). (C,D) Images of CL (C) and IL hemispheres (D) of H/I vehicle-treated animals. (E,F) Representative images of CL (E) and IL (F) of SB505124-treated animal at 4 days after injury. (G) Densitometric analyses of GFAP intensity within the SVZ in H/I and H/I + SB treated animals at 4 days after injury. Untouched controls treated with SB compound were also analyzed and no significance was found between IL and CL hemisphere data (data not shown). * and ** denote p<0.05 using ANOVA analysis with Fisher's PLSD post-hoc test comparing vehicle to H/I and H/I to H/I + SB respectively. Scale bar represents 50 µm.