Limitations of inferring antiviral efficacy of interfering particles from observational natural histories

This paper critiques Hariharan et al.'s 2025 study on naturally occurring defective HIV genomes, arguing that their observational data cannot validly assess the therapeutic potential of interfering particles, that their reported reproductive numbers are internally inconsistent with viral dynamics, and that established alternative mechanisms better explain the observed phenomena.

The Gist

The Big Picture: A "Friendly" Critique

Imagine two scientists, Hariharan and his team, found a group of "broken" HIV viruses living inside people. These broken viruses can't reproduce on their own, but they hijack the healthy viruses to copy themselves. Hariharan's team thought, "Maybe these broken viruses are actually helping the body fight the infection by crowding out the bad ones."

However, Neha Khetan, Gustavo Vasen, Davey Smith, and Leor Weinberger (the authors of this new paper) are saying: "Hold on a minute. You can't tell if these broken viruses are actually helping just by watching what happens naturally. And the math you used to prove they are 'super-competitive' doesn't add up."

They aren't saying the broken viruses are dangerous (in fact, they seem safe!). They are just saying we can't claim they are a cure based on this specific study.

---

Analogy 1: The "Broken Car" in Traffic

The Scenario:
Imagine you are stuck in a traffic jam. You see a car with a flat tire (the "broken" HIV) driving slowly in the middle lane.

Hariharan's View:
"Look! That flat-tire car is blocking the fast sports cars (the healthy HIV). It's slowing down the traffic jam! Maybe if we put more flat-tire cars on the road, we can stop the traffic jam entirely."

The Authors' Counter-Argument:
"Wait. You are looking at a traffic jam that is already happening because the traffic lights are broken (the patients are on medication, but it's not working perfectly).

1. Observation isn't Proof: Just because you see a flat-tire car in traffic doesn't mean the flat tire caused the traffic to slow down. Maybe the traffic was already slow, and the flat-tire car just got stuck there. You can't judge the effectiveness of a 'flat-tire strategy' just by watching a random traffic jam.
2. The Medication Factor: These patients are taking strong medicine (ART) that is supposed to stop all cars from moving. If the medicine is working, both the sports cars and the flat-tire cars should be stopped. If the flat-tire cars are still moving, it's likely because the medicine isn't working well enough, not because the flat-tire cars are magically powerful.

The Lesson: You can't test a new traffic solution by just watching a broken intersection. You need to design a specific experiment to see if the solution works.

---

Analogy 2: The "Impossible Math" (The R0 Problem)

The paper spends a lot of time on a number called R0. Think of R0 as a "Reproduction Score."

• If a virus has an R0 of 2, it means one virus makes two new viruses. The infection grows.
• If a virus has an R0 of 0.5, it means one virus makes half a virus (it dies out). The infection disappears.

The Contradiction:
Hariharan's team measured the "Reproduction Score" of the healthy HIV in these patients and found it was below 1 (it should be dying out).

• The Problem: If the score is below 1, the virus should vanish completely, like a fire that runs out of oxygen.
• The Reality: The patients still have a steady amount of virus in their blood (a "viral set point").

The Analogy:
It's like a scientist saying, "I measured the water pressure in this pipe, and it's too low to push water through. But look, the pipe is full of water!"
The authors say: "The math doesn't work. If the pressure is that low, the water shouldn't be there. Therefore, your measurement of the pressure is probably wrong, or your method of measuring it is flawed."

Because the math is broken, we can't trust their conclusion that the broken viruses are strong enough to be a cure.

---

Analogy 3: The "Ghost in the Machine" (Alternative Explanations)

Hariharan's team thinks the broken viruses are actively "hijacking" healthy viruses to spread around (like a gang stealing a car to drive it).

The authors suggest a simpler explanation: The broken viruses are just "ghosts" leaking out of a hidden basement.

• The Basement (The Reservoir): HIV hides in a "basement" (latent cells) where it sleeps.
• The Leak: Sometimes, the basement door opens slightly (due to stress or other signals), and a few "ghosts" (broken virus particles) leak out.
• The Confusion: Hariharan's team sees these ghosts and thinks, "Wow, they are spreading everywhere! They must be super-active!"
• The Reality: The authors say, "No, they aren't spreading. They are just leaking out of the basement and then dying. They aren't a 'gang' taking over the city; they are just a few people walking out of a house."

The authors point out that we don't need to invent a complex "hijacking" story to explain what we see. A simple "leak" explains it perfectly.

---

The Good News (The Silver Lining)

The authors end on a very positive note. They say:
"Even though we can't call this a cure yet, Hariharan's study is still a huge success."

Why?
For a long time, scientists were worried that putting "broken" viruses into humans would be dangerous or toxic.

• The Verdict: These broken viruses have been living in people for years and haven't hurt them.
• The Takeaway: Nature has shown us that "broken" HIV is safe. This gives scientists the green light to keep trying to engineer better broken viruses (with more damage to them) that might actually work as a cure in the future.

Summary

1. Don't judge a cure by watching a natural disaster. You need a controlled experiment.
2. The math is broken. You can't have a virus dying out (R0 < 1) and still be present in large numbers at the same time.
3. There's a simpler explanation. The broken viruses might just be "leaking" from hiding spots, not actively fighting the good virus.
4. But it's safe! The most important finding is that these broken viruses are harmless to humans, which is a huge step forward for future therapies.

Technical Summary

1. Problem Statement

The paper addresses a critical methodological and interpretative flaw in a recent study by Hariharan et al. (2025), which reported the persistence and conditional replication of naturally occurring defective HIV genomes in humans.

• The Conflict: Hariharan et al. concluded that their findings raise concerns about the therapeutic potential of Therapeutic Interfering Particles (TIPs), suggesting that naturally arising defective genomes failed to suppress HIV.
• The Authors' Critique: Khetan et al. argue that this conclusion is invalid because:
  1. Observational Fallacy: Therapeutic efficacy cannot be adjudicated from observational natural history (post-hoc data) without interventional trials.
  2. Confounding Variables: The study subjects were on Antiretroviral Therapy (ART), which suppresses both wild-type and defective viruses, making it impossible to isolate the specific effect of the defective particles.
  3. Mathematical Inconsistency: The reported basic reproductive numbers ($R_0$) in the Hariharan study are mathematically incompatible with the observed persistent viral loads.

2. Methodology

The authors employed a combination of mathematical modeling, logical deduction, and re-analysis of existing data to critique the Hariharan study.

• Mathematical Modeling:
  • Utilized an established within-host virus dynamics model (an extension of the "Basic Model" of virus dynamics) incorporating populations for target cells, HIV-infected cells, HIV virions, DIP-transduced cells, dually infected cells, and DIP virions.
  • Simulation: Performed 2,000 numerical simulations using a stratified Monte Carlo sampling approach (Latin Hypercube Sampling) across eight key parameters (e.g., T-cell production rate, infection rate, burst size, decay rates).
  • Phase Space Analysis: Calculated $R_0$ (wild-type) and $R_{DIP}$ (defective) to map the conditions required for sustained viral set points.
• Data Re-evaluation:
  • Critiqued the assay methodology used by Hariharan et al. for calculating $R_0$, specifically focusing on how the denominator was inferred (excluding transcriptionally silent cells).
  • Analyzed the timeline of ART regimens and the detection of defective genomes in the Hariharan study to test the hypothesis of "superinfection-mediated mobilization."
• Mechanistic Comparison:
  • Compared the biological structure of the naturally occurring defective genomes (Tat-proximal deletions) against engineered TIPs (which contain extensive deletions and inactivating mutations).

3. Key Contributions & Results

A. The Observational Limitation

The authors emphasize that the Hariharan study was conducted in individuals on ART. Since defective HIV genomes rely on wild-type enzymes (like reverse transcriptase) and are susceptible to ART, they should be suppressed at least as effectively as the wild-type virus. Therefore, the lack of therapeutic benefit observed cannot be attributed to the failure of the defective particles themselves, but rather to the suppression by ART. Attributing the lack of efficacy to the particles is a logical fallacy similar to blaming ART for failing to suppress viremia in patients with suboptimal adherence.

B. Mathematical Inconsistency ($R_0 < 1$ vs. Persistent Viremia)

• The Finding: Hariharan et al. reported an effective $R_0$ for wild-type HIV below unity ($R_0 < 1$) under their assay conditions.
• The Contradiction: By definition, if $R_0 < 1$, the virus must go extinct. However, the patients (P1 and P2) exhibited sustained, persistent viral set points.
• Model Proof: The authors' simulations (Fig. 1) demonstrate that no combination of $R_0$ and $R_{DIP}$ can produce a sustained viral set point if the wild-type $R_0$ is less than 1.
• Conclusion: The reported $R_0$ values are internally inconsistent with the observed biology. The authors suggest the defective $R_0$ may be overestimated due to assay artifacts (e.g., excluding silent infected cells from the denominator), while the wild-type $R_0$ is underestimated.

C. Alternative Mechanisms for Observations

The authors propose that the observations in the Hariharan study can be explained by well-established reservoir dynamics without invoking TIP-like interference:

• Latency Reversal: Clonally expanded Tat-deficient proviruses can undergo latency reversal via NF-κB activation, producing Viral-Like Particles (VLPs) even without Tat, similar to 3rd generation lentiviral vectors.
• APOBEC-Driven Diversity: The observed sequence heterogeneity can be explained by $\Delta$Vif-associated APOBEC-mediated hypermutation during VLP transduction, mimicking active replication without requiring sustained superinfection.
• Timeline Discrepancy: The defective genomes were only detected months/years after the initiation of "optimal" ART, not during the "suboptimal" ART period where superinfection was hypothesized to occur.

D. Biological Distinction from Engineered TIPs

The authors highlight a fundamental difference between the natural defects and engineered TIPs:

• Natural Defects (Hariharan): Contain small deletions (e.g., Tat-proximal) and can still express some viral proteins or reconstitute VLPs under specific conditions.
• Engineered TIPs (Pitchai et al.): Contain massive deletions and mutations that abrogate open reading frames throughout the genome. They cannot express any HIV proteins and strictly require superinfection to replicate, making them true "interfering" agents that cannot reconstitute the virus on their own.

4. Significance

• Scientific Rigor: The paper serves as a crucial correction to prevent the misinterpretation of observational data as evidence against the therapeutic potential of TIPs. It reinforces the principle that causality in therapeutic efficacy requires prospective interventional trials, not retrospective observational analysis.
• Safety Validation: The authors explicitly state that the Hariharan study does provide valuable evidence that defective HIV genomes are biologically tolerable in humans, addressing a major regulatory safety concern for TIP development.
• Future Directions: The paper clarifies that the lack of efficacy in the natural setting does not invalidate the TIP strategy. Instead, it suggests that successful TIPs require more extensive genome engineering (as seen in previous non-human primate studies) to ensure they cannot reconstitute functional virus and to maximize their interference capacity.
• Methodological Warning: It warns against using assay-dependent $R_0$ estimates that contradict fundamental viral dynamics principles (like the extinction threshold) as a basis for clinical decision-making.

In summary, Khetan et al. argue that while the Hariharan study confirms the safety of defective genomes, its conclusions regarding TIP efficacy are mathematically flawed and methodologically unsound. The data supports the tolerability of the approach but does not refute the potential efficacy of properly engineered TIPs.