(I-324) Developing RNA-based sensors for cell-specific detection and manipulation using ADAR to improve liver organoids

Organoids are promising models in studying the functions and processes in native organs, particularly assembly during development [1]. However, current methods of organoid generation are still not ideal, and the resulting organoids fail to accurately represent native organs perfectly. The cell populations of these organoids are not sufficiently mature or lack accurate gene expression [2]. Thus, there needs to be a way to detect and manipulate undesired cell populations. The ability to sense specific cell types would enable engineering of a sense and response to specifically deliver a payload to target populations. RNA has been a target for cell sensing since every cell has its unique RNA profile [3].

We investigated a recently developed system for RNA sensing with adenosine deaminase acting on RNA (ADAR). These enzymes sense double stranded RNA (dsRNA) and edit adenosine to inosine, a nucleotide structurally similar and recognized as guanosine during translation, in base mismatches within dsRNA [4]. This editing mechanism can be utilized with engineered sensors to control translation of the sensors’ protein payloads downstream of an included stop codon [5]. Such protein payloads may be transcription factors for further maturation or caspases to induce apoptosis. Our study aimed to test the efficiency of the ADAR editing system and inducible apoptosis with our sensors containing Caspase9 downstream of the stop codon, design as made by Jiang et al [5]. To confirm successful editing by ADAR, dimerizing agent, AP1903, was added to induce apoptosis by activation of Caspase9 (Figure 1).

Our cell-specific construct consisted of genes mCherry and Caspase9, separated by a p2a sequence. The mRNA encoded contains a complementary region, a SERPINA1 binding site with a stop codon, for Alpha-1 Antitrypsin (AAT), a protein produced by the cell line used, HepG2. mCherry expression is only present upon dsRNA formation by the cell-specific sensor RNA. The plasmid encoding the ADAR enzyme, ADAR1p150, was ordered from AddGene.

The HepG2 cell line was cultured using 10% FBS in DMEM. Sensor constructs, 400ng, and ADAR enzyme, 100ng, were transfected into respective HepG2 cell groups using polyethyleneimine (PEI). Eight wells of a 48-well plate were seeded for testing. Four groups were assigned and compared after transfection and addition of dimerizing agent, if designated to receive it (Figure 2). Dimerization agent was added one day after transfection to groups assigned to receive it.

Two sets of live images of the cells were taken with an EVOS microscope. One set of images was taken one day after transfection just prior to addition of dimerizing agent. Two days after the addition of dimerizing agent, and three days after transfection, the second set of images was taken. In both sets of images, cells expressing mCherry were counted with Fiji software to compare between groups and to determine the fold change.

From live imaging, there was an observable decrease in mCherry expression in groups receiving dimerizing agent compared to groups that did not receive the agent (Figure 3). Acting as a control group, the average number of mCherry+ cells that received only sensor was 11,566. The remaining groups had cell counts much lower than the sensor only group, an average of 7067 or below, as well as fold changes below 0.75, indicating at least a 25% decrease in cells (Figure 4). 

There was a lower count in mCherry+ cells that received only sensor and dimerizing agent compared to cells that received only sensor and enzyme. This may be the stop codon within the sensor not completely gating translation of Caspase9, thus resulting in some Caspase9 expressed without elimination of the stop codon by the ADAR enzyme. Therefore, there could have been enough Caspase9 expressed where dimerization could occur without the help of an agent, dimerization aided further with the addition of the agent. This would also explain how the cells receiving sensor and ADAR enzyme still had a lower count than the control group.

Despite the background observed, the group receiving sensor, ADAR enzyme, and dimerizing agent had the lowest number of cells, indicating that the system is efficient in induced apoptosis. With more studies in determining the best way to completely block translation of protein payloads, controlled expression of proteins can be improved for cell fate manipulation.

In future experiments involving this synthetic sensor, other cell lines can be used to verify cell specificity like HEK293T. These cells do not express AAT, so it would be expected to not observe mCherry expression after transfection with the sensor. Additionally, further quantification of apoptosis can be done by marking Caspase3, a protein activated by Caspase9. This would provide more data to determine the efficiency of our tested ADAR system. Furthermore, other proteins can be used as the payload for different fates like maturation or differentiation.

I would like to acknowledge and thank lab members and mentors, Mohammad Naser Taheri and Yuda Xiang, for their patience, guidance, and help in this project as well as to my principal investigator and mentor, Dr. Mo Ebrahimkhani, for allowing this research in his laboratory. Funding was provided by the Swanson School of Engineering and the Office of the Provost at the University of Pittsburgh to Jennie Bostich, and by an R01 from the National Institute of Biomedical Imaging and Bioengineering (EB028532) to Dr. Ebrahimkhani.

1. 
   

   Velazquez et al. Trends Biotechnol 36, 415-429, 2018.

   
   


2. 
   

   Harrison et al. Front Med 8, 2021.

   
   


3. Qian et al. Nature 610, 713-721, 2022.


4. Kaseniit et al. Nat Biotechnol 41, 482-487, 2023.


5. 
   

   Jiang et al. Nat Biotechnol 41, 698-707, 2023.