Pseudotime trajectory analysis reveals divergent rod photoreceptor states during dark adaptation

This study utilizes pseudotime trajectory analysis of single-cell RNA-seq data to reveal that rod photoreceptors diverge into two distinct lineages during dark adaptation: one driven by MYC-mediated anabolism and stress responses, and another characterized by altered RNA splicing and preserved LKB1-AMPK energy sensing.

The Gist

Imagine your eyes are like a bustling city that never sleeps. Inside this city, there are millions of tiny workers called rod photoreceptors. Their job is to help you see in the dark.

Here's the problem: When the lights go out, these workers have to work twice as hard to keep the city running. They burn a massive amount of energy (ATP) just to stay awake and functional. It's like a factory switching from "day shift" to a frantic "night shift" where the machines are running at maximum speed, generating heat and waste.

This paper is a detective story about what happens inside these workers when the lights go out. The researcher, Ryutaro Ishii, didn't build new mice or run new experiments. Instead, he acted like a digital archaeologist, digging through old data (like a library of genetic blueprints) to see how these cells change as they adapt to the dark.

Here is what he found, explained simply:

The Great Split: Two Different Ways to Cope

When the lights went out, the rod cells didn't all react the same way. Instead of moving as a single group, they split into two different teams (or lineages), like a family of workers taking two different paths to survive the night.

Team 1: The "Power Surge" Crew (Lineage 1)

• What they do: This team decides to crank the engine up to 11. They turn on a master switch called MYC, which tells the cell to build things rapidly and burn fuel.
• The Consequence: Because they are working so hard, they are overheating. They are producing a lot of "exhaust fumes" (Reactive Oxygen Species) and their internal machinery is getting clogged with unfinished products (Protein Stress).
• The Metaphor: Imagine a car engine revving so high that it starts smoking. It's powerful, but it's putting a lot of stress on the engine. If this stress goes on too long, the engine might break down (which is bad for your eyes).

Team 2: The "Pause and Reorganize" Crew (Lineage 2)

• What they do: This team takes a different approach. Instead of revving the engine, they seem to hit the brakes on their paperwork. They have a massive backlog of unfinished instructions (unspliced RNA).
• The Consequence: They are holding onto these instructions, perhaps to slow down production and save energy. They are also keeping a specific "energy sensor" (the LKB1-AMPK module) active, which is like a smart thermostat that tells the cell, "Hey, we are running low on fuel, let's conserve power."
• The Metaphor: Imagine a construction site where the foreman stops sending out new blueprints. The workers stop building new walls and instead focus on organizing the materials they already have. It's a strategy to survive a famine rather than a strategy to build a skyscraper.

The Hidden Connection: The "Text Message" System

The researcher noticed something fascinating about Team 2. The "unfinished instructions" they are holding onto seem to be controlled by a specific type of text message system (microRNAs).

Think of it this way:

• Team 1 is shouting orders over a loudspeaker (MYC signaling), ignoring the noise.
• Team 2 is using a secret text message app (miRNA) to quietly tell specific workers to stop working or to wait. This text message system seems to be the key to keeping their "energy thermostat" working correctly.

Why Does This Matter?

For a long time, scientists thought all rod cells just did the same thing when it got dark. This paper suggests that nature has two different survival strategies for the same job:

1. Go hard or go home: Burn energy fast, risk damage, but maybe get the job done quickly.
2. Slow and steady: Pause production, manage energy carefully, and protect the cell from burning out.

The big question now is: Are these two teams friends or enemies?

• Is Team 1 the "hero" that helps you see, but risks damaging your eye over time?
• Is Team 2 the "protector" that saves the eye from damage but might make your vision a bit slower?

The Bottom Line

This study is like finding a map that shows two different routes through a storm. One route is fast but dangerous; the other is slower but safer. The researcher hasn't driven the cars yet (that requires new experiments), but he has found the map.

Understanding which route your eye cells take could help us figure out why some people lose their night vision or develop eye diseases as they age. It suggests that maybe we can't just "fix" the eye; we might need to help the cells choose the right survival strategy.

Technical Summary

1. Problem Statement

Rod photoreceptors face a paradoxical metabolic challenge: they consume significantly more ATP in darkness than in light to maintain the "dark current." Despite this high energy demand, the molecular mechanisms coordinating the adaptation from light to dark, and how rods maintain cellular homeostasis under this stress, remain poorly understood. Traditional in vivo approaches using transgenic models are time-consuming and often lack the breadth for unbiased screening of multiple regulatory pathways. The authors sought to use data-driven approaches to dissect the transcriptional dynamics and identify key regulatory programs governing this adaptation.

2. Methodology

The study employed a comprehensive bioinformatics pipeline to re-analyze publicly available murine retinal single-cell RNA sequencing (scRNA-seq) data (GSA accession: CRA006518) collected under normal light and dark-adapted conditions.

• Data Preprocessing: Raw FASTQ files were processed using Cell Ranger (10x Genomics). Ambient RNA was removed (SoupX), and doublets were filtered (DoubletFinder). Spliced and unspliced count matrices were generated using velocyto.
• Dimensionality Reduction & Clustering: Standard Seurat workflows (PCA, UMAP) were used to cluster rod photoreceptors.
• Trajectory Inference: Pseudotime analysis was performed using Slingshot. The trajectory was rooted at the "Light" state (Cluster 5) to model the transition to "Dark" states, revealing a bifurcation into two distinct lineages.
• Functional & Motif Analysis:
  • Gene Set Scoring: UCell was used to compute module scores for Hallmark gene sets (MSigDB) and pathway activity (PROGENy).
  • RNA Processing: The unspliced RNA fraction was calculated. Intronic sequences of genes with elevated unspliced signals in specific lineages were extracted for de novo motif discovery (STREME) and matched against known RNA-binding protein (RBP) motifs (Tomtom/CISBP-RNA).
  • Regulatory Network Analysis: Upstream regulator analysis utilized Enrichr (ChEA, TRRUST, ENCODE) and ChIP-Atlas to identify transcription factors binding to specific loci (e.g., Strada, Cab39l). Protein-protein interaction (PPI) networks were mapped using STRINGdb.
• Statistical Framework: Comparisons between lineages utilized Wilcoxon rank-sum tests and Area Under the ROC Curve (AUC) to quantify effect sizes.

3. Key Contributions & Results

The analysis revealed that dark-adapting rods do not follow a uniform path but diverge into two distinct transcriptional lineages:

Lineage 1: The High-Anabolic/Stress State

• Characteristics: This lineage is defined by a MYC-driven anabolic program.
• Metabolic Signatures: It exhibits significantly elevated scores for oxidative phosphorylation, glycolysis, and mTORC1 signaling (AUC = 0.88).
• Stress Response: Despite high metabolic output, this lineage shows strong signatures of the Unfolded Protein Response (UPR)/ER stress and Reactive Oxygen Species (ROS) response.
• Implication: The persistence of anabolic programs (mTORC1) alongside stress signatures suggests these cells may struggle to attenuate energy-consuming processes under stress, potentially leading to proteotoxic and oxidative damage.

Lineage 2: The RNA-Processing/Adaptive State

• Characteristics: This lineage is characterized by altered RNA processing, specifically a markedly higher fraction of unspliced RNA (AUC = 0.07 favoring Lineage 2).
• Target Genes: The accumulation of unspliced transcripts disproportionately affects long, vision-related genes (e.g., Sag, Prph2, Pde6b).
• Mechanism: De novo intronic motif analysis identified enriched binding sites for splicing-associated RBPs (e.g., SRSF7, RBM45, PTBP1). PPI networks linked these candidates to core spliceosome components and ATP-dependent helicases.
• Energy Sensing: Unlike Lineage 1, Lineage 2 relatively preserves the STRADA/MO25β (CAB39L) module. This complex is essential for activating the LKB1-AMPK energy-sensing pathway.
• Regulation: Upstream regulator analysis of Strada and Cab39l implicated miRNA-mediated networks and TGF-β signaling as potential tuners of this energy-sensing module.

Divergence Mechanism

The study proposes a model where the divergence is driven by a trade-off between anabolic drive and energy conservation:

• Lineage 1 maintains high mTORC1 activity (anabolism) but fails to fully engage the LKB1-AMPK checkpoint, leading to stress accumulation.
• Lineage 2 appears to engage post-transcriptional regulation (via splicing modulation and miRNA networks) to potentially reduce translation output and preserve the LKB1-AMPK module, allowing for better metabolic adaptation.

4. Significance

• Novel Insight into Dark Adaptation: The study challenges the view of dark adaptation as a uniform process, revealing a bifurcation into distinct cellular states with different metabolic and post-transcriptional strategies.
• Link Between Metabolism and RNA Processing: It establishes a potential mechanistic link between energy sensing (AMPK/mTORC1) and RNA splicing kinetics, suggesting that the regulation of intron retention is a critical component of metabolic adaptation in photoreceptors.
• Hypothesis Generation for Retinal Degeneration: The findings suggest that Lineage 1 cells, characterized by high anabolic drive and unresolved stress, may be more susceptible to degeneration. Conversely, the preservation of the LKB1-AMPK module in Lineage 2 may represent a protective adaptive state.
• Future Directions: The authors emphasize that while these are correlative transcriptomic findings, they provide a strong framework for future experimental validation. Key next steps include biochemical assays of phospho-AMPK/mTOR targets, direct measurement of spliceosomal assembly, and investigating the role of miRNAs in tuning these energy-sensing pathways.

In summary, this paper utilizes advanced pseudotime trajectory analysis to uncover a complex, bifurcated regulatory landscape in rod photoreceptors during dark adaptation, highlighting the interplay between MYC-driven anabolism, RNA splicing regulation, and the LKB1-AMPK energy-sensing axis.