Blimp1 Controlled Cell Fate Decisions in Mouse Retinal Development
By Miranda Lyons-Cohen
Photoreceptors are the cells in the retina that detect light. Photoreceptors are in the outer most layer of the 3-layered retina. Bipolar cells (BCs) are “interneurons” that are located between the retinal ganglion cell (RGC) layer and the photoreceptors, relaying information between the two. BCs are like the peanut butter that holds the two slices of bread together for a PBJ sandwich. During retinal development, a population of undifferentiated cells with the possibility of becoming different types of photoreceptors, BCs, and Müller glia (support cells) express a transcription factor called Otx2. Both photoreceptors and BCs express Otx2, which is required for the growth of both of these cell types. Transcription factors are proteins that bind to a specific DNA sequence and control the transcription of that genetic information into certain functional proteins encoded by that piece of DNA. Transcription factors can be used as visible “markers” of a gene’s expression. A 2012 paper by Brzezinski et al. investigates how populations of Otx2 expressing cells choose between a photoreceptor or bipolar cell fate. They examined the role of another transcription factor, Blimp1, found in the retina determining that Blimp1 expression stabilizes developing photoreceptors versus bipolar cells and regulates the decision between bipolar and photoreceptor cell fates.
Blimp1 expression was first characterized in wild type (WT, normal) mouse retinas by using an antibody to detect the Blimp1 protein. This process is called immunolocalization. Blimp1 expression was initially characterized with Otx2 and Crx, revealing a pattern of timed expression in which Otx2 is expressed before Blimp1, which is expressed before Crx. This is logical, showing the cell population of either photoreceptors or bipolar cells (marked by Otx2) gave rise to a population of photoreceptors (Blimp1) which then gave rise to photoreceptor differentiation (Crx). Blimp1 expression was also characterized with NeuroD1 and Trβ2 indicating that Blimp1 is expressed in both rods and cones, the two types of photoreceptor cells.
These initial experiments suggested that Blimp1 played a role in photoreceptor development so the next step was to create a Blimp1 loss-of-function mutant mouse. They used a technique called conditional knockout in which a specific target gene is eliminated from one organ in the body, in this case the retina, rather that the entire genome as traditional gene knockout would entail. These blimp conditional knockout (CKO) mice did not contain the gene that codes for Blimp1 only in the retina thus allowing for close examination of the effects of Blimp1 on the eye. Also, Blimp1 traditional knockout mice died in utero so conditional knockout was not only a useful technique but also necessary. They also generated a heterozygous control that maintained its Blimp1 function. Heterozygous mice acted as a useful control because they still expressed Blimp1 but were crossed with one of the strains to eliminate Blimp1, making sure nothing else was different in the mutant mice other then loss of Blimp1 function. Examining these 3-4 week old CKO mice revealed startling differences from the heterozygous control. It was immediately clear that there had been a loss of photoreceptor cells and a gain of BCs in the CKO mice by ~50% as seen by staining for Chx10, Vsx1, and PKC (Table 1). The laminar organization, or the discrete cell layers, of both cell groups were disrupted. Interestingly, they also found ~50% more Müller glia cells which are long cells that span the entire length of the retina providing structural support as seen by the Sox9 marker (Table 1). Importantly, they did not observe any proliferating (dividing) or apoptotic (dying) cells in the CKO retinas confirming that the observed results were not due to cell death or growth and all cell population changes occurred during development.
WT and CKO mice were characterized with multiple markers, and “extra” bipolar and Müller glia were observed at the expense of photoreceptors in CKO mice. They next looked at early CKO retinas for changes in photoreceptor markers to demonstrate the role for Blimp1 in regulating photoreceptor fate. This was one of their most crucial experiments because it explicitly characterized the role of Blimp1. During embryonic development of CKO mice, small sections of Vsx1 were observed. By postnatal day 1, a number of cells were labeled with both Vsx1 and Chx10, indicative of bipolar cell specification. By postnatal day 3, CKO mice had a large number of these bipolar cells while control mice displayed no Vsx1 or Chx10 indicating that the loss of Blimp1 was somehow responsible for the bipolar cell fate. All the Vsx1 and Chx10 cells observed were also co-labeled with Otx2 consistent with the notion that it was the Otx2 cell population that was adopting the bipolar cell fate in mutant mice. So far, these data indicate that Blimp1 negatively regulates Chx10 and Vsx1 in the Otx2 population. The negative regulation mechanism in normal eyes would presumably lead to equal development of photoreceptors and BCs required for normal eye function, but without this regulation, there were “extra” BCs growing at the expense of photoreceptors.
Near the end of retinal development, at postnatal day 7, the heterozygous controls had distinct layers with appropriate ratios of Otx2 photoreceptors and bipolar cells within each cell layers. However the CKO mice had numerous Otx2 bipolar cells and very few photoreceptors. This caused an abnormally thin retina in the mutant mice. The numbers of bipolar cells and photoreceptors were quantified in control and CKO mice finding that the decrease in photoreceptors was proportional to the increase in bipolar cells yet the Otx2 labeled populations were the same. This shows that there was a one to one shift in differentiation between photoreceptors and bipolar cells and that the new cells were not coming from another population. By postnatal day 10, the CKO mice had a very thin retina composed of few photoreceptors compared to the control. A decrease in total number of Otx2 cells between postnatal days 7 and 10 was observed, suggesting that many of the “extra” bipolar cells were dying in the retina between days 7 and 10. Caspase3 was used as an apoptotic, or dying, cell marker. Between post natal days 1 and 3 no cell death was observed however after post natal day 7 there was a ~3 fold increase in the amount of cell death in CKO mice. However by postnatal day 10, the amount of cell death was equivalent between the heterozygous control and the CKO mice. Although it appeared as if the “new” bipolar cells were undergoing apoptosis caused by a change in cell fate based on lack of Blimp1, the cells actually correlated with the timing of bipolar cell maturation and the normal cell death of this layer because equal amount of cell death was seen overall in both CKO and control mice.
A final experiment confirmed their results by over-expressing instead of eliminating Blimp1. To test whether Blimp1 is what actually inhibits bipolar cell development, they engineered expression plasmids to drive green fluorescent protein (GFP) as a control or Blimp1-GFP in the retina. An expression plasmid is a small DNA molecule that is used to introduce a specific gene into a target cell; in this case it introduced GFP into Blimp1 expressing cells. For the Blimp1-GFP plasmid, the expression of Blimp1 would essentially be driving the expression of GFP causing an increase in visible brightness correlated with an increase in Blimp1 expression. The control was GFP not driven by Blimp1 expression. These constructs were injected into the eye at postnatal day 1. After 4 days, the transfected cells were examined to determine their bipolar or photoreceptor fate. The transfected cells expressed Otx2 in the control as well as the Blimp1-GFP cells. They found that Blimp1-GFP cells adopted photoreceptor fate 26% more readily and bipolar fate 24% less readily than the control. Blimp1 overexpression confirmed that Blimp1 is responsible for fate distribution of the Otx2 population.
Why are any photoreceptors seen in the mutant retinas? If Blimp1 was the only transcription factor controlling photoreceptor cell fate, than elimination of Blimp1 should lead to elimination of all photoreceptors. The possibility that there was another factor compensating for Blimp1 activity was suggested. They also posited that Chx10 is needed to repress photoreceptor cell fate while Blimp1 is needed to repress bipolar cell fate. This balancing act in the Otx2 population between photoreceptor and bipolar cell fate could help to generate necessary cell diversification in the developing retina. A possible explanation for their observation that the Müller glia cell population also increases in CKO mice is that Müller glia are required to maintain bipolar cell function. This hypothesis is supported by their observation that both Müller glia and bipolar cell numbers are increased by 50% in CKO mice. A 50% increase in bipolar cells might support a 50% increase in Müller glia. However, this does not explain the direct fate shift in some Otx2 cells to Müller glia. The idea that some Müller glia derive directly from the Otx2 cell population requires a lineage-tracing experiment and further investigation.
Brzezinski et al. were able to demonstrate that Blimp1 is expressed in developing photoreceptors in mammals. They found that genetic deletion of this transcription factor converted most of the potential photoreceptor cells into bipolar cells and a few Müller glial cells. As the mutant retinas matured, many of the fate-shifted cells died, however there was still ~50% more bipolar cells and Müller glia than the control. In a converse experiment, they were able to show that over-expression of Blimp1 promoted photoreceptor cell fate over bipolar cells. Together, all of these data support the hypothesis that Blimp1 acts downstream of Otx2 to regulate cell fate decisions in retinal cells.
Click here to access the research article on which this piece is based.