In a recent paper published in Science, Turnbull et al.¹ show that a modest fever-range rise in core temperature (~2 °C) directly restricts early influenza A virus replication in vivo, but that avian-origin PB1 (recapitulated by G180E/S394P) confers thermotolerance that overcomes this protection. These results establish fever as a physiological antiviral barrier and suggest that PB1 thermotolerance could inform surveillance of zoonotic strains and guide interventions that exploit polymerase vulnerabilities.

Fever is among the most characteristic features of influenza; however, its direct antiviral role remains difficult to prove in vivo.^1,2 A core challenge is disentangling temperature effects from accompanying immune signals. Turnbull et al.¹ isolated the temperature component, demonstrating that a modest increase in core body temperature (~2 °C), within typical febrile ranges, can markedly suppress influenza A virus (IAV) replication early during infection and shift disease outcomes from severe to mild. This protection is strain-dependent, as viruses harboring an avian-like polymerase basic protein 1 (PB1) polymerase subunit replicate under febrile conditions and cause severe disease, establishing fever-range hyperthermia as a frontline restriction mechanism and identifying PB1-mediated thermotolerance as a viral countermeasure.

This study builds on a long-recognized ecological contrast. Avian IAVs replicate in birds at relatively high physiological temperatures (~40–42 °C), while human-adapted seasonal viruses prefer the cooler upper respiratory tract (~33 °C) and often exhibit impaired replication in mammalian cells at ~40 °C. This temperature gap has fueled the hypothesis that fever acts as a host-imposed thermal barrier, but conventional experimental approaches, such as pyrogens or antipyretics, simultaneously alter inflammatory signaling and cellular physiology.

Turnbull et al.¹ circumvented these confounders using a “simulated fever” paradigm: rather than pharmacologically manipulating inflammation, they raised ambient housing temperature to sustain mouse core temperatures in the fever range, isolating temperature effects with minimal immune intervention. This design creates a stringent test: if elevated temperature alone is antiviral, replication and disease should diverge under controlled hyperthermia before inflammatory differences arise.

In vitro profiling identified temperature sensitivity as a distinguishing phenotype between viruses of human and avian origins. A human-origin laboratory strain, PR8 (A/Puerto Rico/8/1934), is strongly attenuated at 40 °C vs. 37 °C in human lung cells, while an avian-origin strain, Mallard (A/Mallard/Netherlands/10‑Cam/1999), replicates efficiently at both temperatures. Multiple human seasonal isolates are broadly inhibited at febrile temperatures, whereas diverse avian viruses remain largely unaffected, showing that fever-range replication capacity is not an idiosyncratic trait but a recurring axis of host adaptation.

The genetic basis of febrile-temperature resistance maps strongly to the viral polymerase, particularly PB1. Reassortant viruses show that avian PB1 confers robust replication at 40 °C, aligning with influenza evolution; pandemic lineages often acquired avian-origin PB1, suggesting polymerase thermotolerance may influence disease severity in mammals by weakening fever-dependent restriction.^3,4

Mechanistically, the study pinpoints PB1 features that “avianize” a temperature-sensitive human-origin backbone. By constructing PR8–Mallard PB1 chimeras and progressively refining the responsible region(s), the authors identified two substitutions (G180E and S394P) sufficient to confer high-temperature replication. Consistent with a polymerase-centric defect, PR8’s 40 °C limitation tracks with impaired viral genome replication, rescued by avian PB1 or by the PR8 PB1 double mutant. Together, these experiments support a model in which fever-range temperatures directly stress polymerase, and PB1 sequence dictates whether polymerase activity collapses or persists.

The most decisive advancement is the in vivo demonstration that fever-range hyperthermia can independently exert antiviral effects, and that PB1 thermotolerance can counteract this protection. Under standard conditions, both the parental PR8 and the temperature-resistant PB1 mutant cause severe disease and weight loss. Under simulated fever (where core body temperature was maintained at least ~2 °C higher), mice were protected from wild-type PR8: signs of disease were milder, weight loss reduced, and none of the animals reached humane endpoints. In contrast, the PB1 thermotolerant mutant largely overcame this protection, driving severe outcomes despite fever-range host temperatures (Fig. 1a).

The alternative text for this image may have been generated using AI.

Fever as a thermal barrier and PB1-mediated thermotolerance in influenza A viruses. a In a simulated fever setting (∼2 °C increase in core temperature), infection with the human-origin PR8 strain is attenuated, providing protection from severe disease. In contrast, avian-like PB1 in the PR8 background (PB1 G180E/S394P) increases polymerase thermotolerance and leads to severe disease under the same simulated fever condition. b Species-specific ANP32 cofactors modulate PB1-dependent polymerase temperature sensitivity at 40 °C: avian ANP32A is more permissive for PR8 polymerase function, whereas mammalian ANP32A/B unmask higher temperature sensitivity (restricted polymerase activity). (Created with BioRender.com)

Early viral load measurements support temperature’s direct effect: simulated fever was associated with ~10-fold reduction in lung viral load for wild-type PR8 at 24 h post-infection, whereas PB1 mutant titers decreased only about two-fold under the same conditions. No early pro-inflammatory surge attributable solely to warm housing was observed at this time point, indicating that temperature-mediated replication restriction, rather than enhanced inflammatory signaling, drives early viral load reduction. Suppression at 24 h post-infection is particularly important because even small shifts in viral kinetics at this stage can substantially reshape downstream pathology.

Host context further modulates the thermal phenotype. PR8 is strongly temperature-sensitive in human cells but replicates efficiently at 40 °C in chicken DF-1 cells, suggesting that host cofactors buffer PB1-encoded thermal fragility. Focusing on ANP32, PR8 became highly temperature-sensitive in DF-1 cells complemented with human ANP32A or ANP32B, but not with chicken ANP32A,⁵ whereas a Mallard-PB1 PR8 reassortant remained temperature-resistant across ANP32 conditions. Together, these data suggest that the polymerase–ANP32 axis is a key host–viral interface determining whether fever-range temperatures restrict replication (Fig. 1b). However, the precise molecular mechanism by which avian ANP32A modulates PB1-dependent temperature sensitivity and supports influenza polymerase activity at fever-range temperatures remains unclear.

This study reframes fever-range temperature as a physiological restriction factor with clear translational implications. First, fever may represent more than an epiphenomenon of immune activation; it can constitute a host state that directly suppresses influenza replication early enough to alter disease trajectory. This reframing naturally raises clinically relevant questions about when symptomatic fever suppression is beneficial for comfort and safety versus when it might inadvertently blunt a host advantage. These questions likely depend on the patient’s context and the viral genotype. Second, the PB1 sequence features associated with thermotolerance may complement existing surveillance priorities by identifying viruses with enhanced capacity to replicate under febrile stress. Third, by highlighting a genotype-dependent failure mode of polymerase under heat and a strong dependance on ANP32 cofactors, this study underscores a therapeutically relevant principle: polymerase fitness is constrained by host cofactor networks, suggesting that interventions that further destabilize polymerase function could, in principle, synergize with physiological constraints such as fever-range temperature.

As with any controlled model, the limitations define important next steps. Simulated fever does not fully recapitulate the endocrine, metabolic, and tissue-level gradients of endogenous febrile responses, and the local airway temperatures encountered by replicating viruses remain difficult to quantify precisely. PR8 is also a laboratory-adapted strain selected for experimental tractability, and broader validation across contemporary seasonal isolates, zoonotic spillover candidates, and human airway models is essential. Nonetheless, the central observation is compelling: a modest ~2 °C rise in core temperature can materially suppress influenza replication in vivo, and avian-like PB1 provides a plausible molecular route by which certain viruses can breach this thermal barrier.