A drug against pulmonary fibrosis shifted six protein "aging clocks" to a younger date.

The experimental drug rentosertib, developed for the treatment of idiopathic pulmonary fibrosis, was associated not only with changes in lung function but also with an unexpected shift in the molecular signatures of aging. In a new analysis of data from a randomized Phase IIa trial, all six independently developed protein models of biological age showed a trend toward a decrease in estimated age in patients receiving the drug. The most consistent signal was observed after four weeks of treatment. The study was published on September 7, 2026, in Nature Biotechnology.

In the most significant group, receiving 30 mg of rentosertib twice daily, several models estimated a reduction in estimated biological age of approximately 3-4 years, with one model estimating a reduction of nearly six years. In the other group, receiving 60 mg once daily, four models trained to predict chronological age showed a reduction of 2.71-3.46 years after four weeks. However, these figures should not be taken literally as proof that the patient's body has truly "aged four years." Rather, they represent a change in a mathematical indicator calculated based on blood protein concentrations.

The study is particularly interesting because the authors used not just one aging model, but six: ProtAge, two versions of OrganAge, PAC, ipfP3GPT, and PAOPAC. They were developed by different research groups, utilize different sets of proteins, and were trained on different endpoints—some based on chronological age, others on mortality risk. Despite this methodological heterogeneity, the direction of changes after therapy was similar.

However, the authors themselves emphasize a key limitation: the protein clock is unable to distinguish between a true slowing of biological aging and an improvement in the condition of a seriously ill patient. Idiopathic pulmonary fibrosis itself alters inflammatory, metabolic, and extracellular proteins in the blood, so some of the "rejuvenation" may simply reflect a weakening of the disease. Additional analyses have provided arguments against this simple explanation, but the study cannot definitively resolve the issue.

Key indicator	Result
Disease	idiopathic pulmonary fibrosis
Preparation	rentosertib
Target	TNIK kinase
Initial clinical trial	Phase IIa, randomized, double-blind
Total participants in the original study	71
Participants in protein analysis	42
Middle age	67.1 years
Proteins studied	2841
Protein clock of aging	6
Points of analysis	initial, 2, 4 and 12 weeks
The most consistent effect	Week 4
The most consistent scheme	30 mg twice daily
Typical assessment of age change	about 3-4 years
Maximum rating for one model	about -6 years
Scheme with the best FVC	60 mg once a day
DOI	10.1038/s41587-026-03286-y

What is rentosertib and why is it linked to aging?

Rentosertib, previously designated ISM001-055, is a small molecule that inhibits the activity of TNIK kinase. This protein target is involved in signaling pathways that regulate fibrosis, cell division, inflammation, and extracellular matrix remodeling. The drug was initially developed not as an anti-aging treatment, but as a potential therapy for idiopathic pulmonary fibrosis, a progressive disease in which normal lung tissue is gradually replaced by scar tissue.

TNIK was selected using a computational drug discovery platform, and the structure of rentosertib itself was optimized using generative artificial intelligence methods. Previous studies have linked TNIK not only to fibrosis but also to several molecular hallmarks of aging. In laboratory models, blocking this kinase reduced the transition of fibroblasts to activated myofibroblasts and attenuated the accumulation of extracellular matrix.

Further interest arose following experiments on cellular senescence. Previous work demonstrated that pharmacological inhibition of TNIK can alter the properties of senescent cells and reduce the severity of their proinflammatory secretory phenotype. Such drugs are considered to alter their pathological secretory activity rather than destroy senescent cells. This led to the hypothesis that rentosertib may simultaneously affect fibrosis and some processes of biological aging.

The new study attempted to test this hypothesis using patient blood samples. It's important to distinguish between the drug's development and the current analysis: the clinical trial was conducted to evaluate pulmonary fibrosis therapy, and biological age was an exploratory molecular endpoint, not the primary objective of the trial. Therefore, the results should be viewed as an early signal that requires separate verification, rather than as clinical proof of geroprotective activity.

How the original clinical trial was conducted

The initial Phase IIa study of rentosertib was conducted in China from July 2023 to June 2024. Of the 128 patients with confirmed idiopathic pulmonary fibrosis, 71 over 40 years of age were included in the study. Participants were randomly assigned approximately equally among four groups.

The first group received 30 mg of the drug once daily, the second received 30 mg twice daily, the third received 60 mg once daily, and the fourth received placebo. Treatment lasted 12 weeks, with standard pulmonary fibrosis therapy allowed. The groups included 18, 18, 18, and 17 patients, respectively.

The primary objective of Phase IIa was safety, not proof of efficacy, much less slowing aging. The incidence of adverse events occurring after treatment initiation ranged from 72.2% to 83.3% in the drug groups and 70.6% in the placebo group. Of the 71 patients, 55 completed the twelve-week study period, and 16 discontinued participation early.

Blood samples were collected before treatment and then after two, four, and twelve weeks. After the study's completion, participants were asked to consent to extensive protein analysis. Forty-three participants consented, but one was excluded from the current study due to a missing sample at week twelve. Ultimately, the biological age analysis was based on a total of 42 patients: 11 in the placebo group, 11 each in the 30 mg once-daily and twice-daily groups, and 9 in the 60 mg group.

Group	Initial research	Protein analysis
Placebo	17	11
Rentosertib 30 mg once a day	18	11
Rentosertib 30 mg 2 times a day	18	11
Rentosertib 60 mg once a day	18	9
Total	71	42

What is the protein clock of aging?

Biological clocks are statistical models that attempt to estimate the body's state based on molecular data. Instead of a birth date, the model receives hundreds or thousands of biological indicators and looks for combinations that are characteristic of people of different ages or with different mortality risks. If a 70-year-old's profile is more similar to the average 65-year-old, the model might assign them a biological age of 65.

In the current study, the researchers measured 2,841 proteins in serum using the Olink Explore high-throughput panel. Proteins are particularly interesting for such models because they are directly involved in inflammation, metabolism, intercellular signaling, vascular function, and tissue remodeling. In this sense, the protein profile can respond relatively quickly to drug intervention.

The authors used six models. Four were primarily trained to estimate chronological age, and two were trained to estimate mortality-related traits. Even at the baseline, these two classes of models differed significantly: the chronological age models correlated quite well with each other, with coefficients of 0.78-0.89, while the correlation between them and the mortality models was only 0.25-0.61.

This difference actually heightens interest in the overall direction of the results. If six nearly identical algorithms had yielded the same answer, the independence of such confirmation would be limited. Here, the models were quite different. But the downside is that the term "biological age" conceals different mathematical constructs that are not necessarily measuring the same fundamental property of the organism.

What kind of watch did you use?

Model	main idea
ProtAge	protein age assessment
OrganAge - Chronological Version	chronological age prediction
OrganAge - a version of mortality	mortality risk profile
PAC	protein model of aging/mortality risk
ipfP3GPT	protein model taking into account the pulmonary fibrosis profile
PAOPAC	additional protein model of age
General principle	the change in many proteins is translated into conventional "years"

All six models showed a move towards a younger profile.

The authors compared three treatment regimens with placebo at three time points—two, four, and twelve weeks. This yielded a total of 54 comparisons: six hours x three time points x three dose regimens. Of these, 21 comparisons met the authors' threshold for statistical significance after adjusting for multiple testing.

The signal was particularly concentrated at week four. At that time, 11 of 18 comparisons between treatment regimens and placebo were significant. A permutation analysis with 100,000 random patient reassignments showed that 21 significant results were extremely unlikely to have occurred by chance alone: p<0.0001.

The 30 mg twice-daily regimen proved to be the most consistent, yielding nine significant comparisons. For 60 mg once-daily, there were seven, and for 30 mg once-daily, there were five. Moreover, both chronological age-based and mortality-based models detected the signal for the twice-daily regimen.

Removing six participants who experienced grade 3 or higher adverse events from the analysis did not significantly change the picture. This reduces the likelihood that the change in protein age was simply a consequence of a severe toxic reaction. However, the small sample size still means that even robust statistical findings need to be replicated in a significantly larger group.

Comparison	Result
Total statistical comparisons	54
Significant after correction	21
Significant in week 4	11 of 18
30 mg 2 times a day	9 significant comparisons
60 mg 1 time/day	7
30 mg once a day	5
Getting this number of results by chance	p<0.0001

"Minus four years" does not mean a real rejuvenation of four years

The most eye-catching figure is a reduction in estimated age of approximately three to four years after four weeks of therapy. In one model, the most sensitive group saw a reduction of nearly six years. But such figures require careful translation into everyday language.

The protein clock doesn't measure remaining lifespan and can't verify whether a patient's heart, brain, or immune system have literally become the same as those of someone four years younger. It measures how similar the current set of protein concentrations is to the profile the algorithm previously associated with younger age.

Therefore, a change from 70 to a conventional 66 "biological years" cannot be interpreted as a restoration of four years of life. It simply means that the mathematical function, based on a specific set of proteins, has changed its prediction by four units. Proving a true slowing of aging would require long-term data on functional capacity, diseases, disabilities, and mortality.

This is especially important in patients with severe age-related diseases. If a drug reduces inflammation or fibrosis, dozens of proteins can revert to a healthy profile. The aging model can interpret this change as "rejuvenation," although it actually records an improvement in the specific disease. The authors explicitly cite the inability to completely separate these processes as the main limitation of their study.

The maximum "age" effect and maximum improvement in lung function occurred with different regimens.

One of the most interesting arguments against the simple explanation of improved fibrosis was the difference between the dose regimens. In the initial clinical trial, the greatest improvement in forced vital capacity was observed with 60 mg once daily. After 12 weeks, the average change was +98.4 ml, compared to -20.3 ml in the placebo group.

However, the most consistent protein age signal was observed not with this regimen, but with 30 mg twice daily, even though the total daily dose was the same—60 mg. By week four, five of the six models showed a consistent negative age shift in this group.

The researchers then directly tested the extent to which changes in lung function explained changes in protein age. Depending on the clock used, changes in FVC explained only 1-18% of the variation, and the median coefficient of determination was only 0.06, or about 6%.

This is an important, but not definitive, argument. FVC is only one indicator of the severity of pulmonary fibrosis and does not reflect all the inflammatory and systemic changes of the disease. Therefore, a low correlation with FVC indicates that "rejuvenation" is not simply a mathematical reflection of an improvement in a single respiratory parameter, but does not prove an independent geroprotective effect.

Indicator after 12 weeks	Placebo	30 mg 1 time	30 mg 2 times	60 mg 1 time
Change in FVC	-20.3 ml	-27.0 ml	+19.7 ml	+98.4 ml
The strongest signal of the aging clock	No	weaker	the most coordinated	moderately consistent
Conclusion	-	-	maximum protein signal	maximum pulmonary signal

A comparison with 55,000 people showed the opposite direction of age-related changes.

To further test whether protein changes are truly the opposite of normal aging, the authors used data from 55,319 elderly participants in the UK Biobank. Of the 2,832 proteins measured in both the biobank and the clinical study, 758 showed significant changes with age in the large population sample.

Proteins altered by rentosertib were 1.74 times more frequently represented among age-related proteins than expected by chance. The probability of this enrichment was p<0.001. In other words, the drug particularly frequently affected those molecules whose concentrations typically change with aging.

For the 30 mg twice daily regimen, the direction of changes after treatment was negatively correlated with the direction of normal age-related change: Spearman's rho was -0.30, p<0.01. If a particular protein typically increased with age, it was more likely to decrease after treatment, and vice versa.

Interestingly, no such association was found with 60 mg once daily: r=−0.097, p=0.37, despite a more pronounced improvement in FVC. This again supports the hypothesis that the protein signal of the twice-daily regimen is not entirely determined by the reduction in pulmonary fibrosis. However, comparison with the biobank remains computational and does not replace direct testing of the drug for aging.

Rentosertib altered the concentrations of hundreds of proteins

The authors went beyond the final clock readings and analyzed each of the 2,841 proteins measured separately. Over 12 weeks, the concentration trajectories of 326 proteins showed significant differences between the treatment groups, while only two met these criteria in the placebo group.

Most of the changes were dependent on the treatment regimen: 237 proteins changed with only one dose. The most widespread response was again observed at 30 mg twice daily, with 142 proteins showing dynamics unique to this regimen. In contrast, at 30 mg once daily, a significant temporal change was found for only one protein: IDO1.

Among the proteins that decreased were COL1A1, MMP10, and FAP—molecules associated with fibrosis and extracellular matrix remodeling. PDGFB, which is involved in proliferative signaling, also decreased. Proteins involved in metabolism and cellular stress resistance, including NAMPT, SOD2, and ALDH1A1, were also altered.

Moreover, 89 proteins changed in the same direction across at least two treatment groups. This is particularly interesting because it suggests a common pharmacological response to TNIK inhibition that is not entirely dependent on a specific dosing regimen. However, some of these molecules are simultaneously involved in both fibrosis and aging, so it is again impossible to draw a clear distinction between the two processes.

Protein analysis	Result
Proteins measured	2841
Changed during the rent-certification	326
Changed to placebo	2
Unique to only one scheme	237
They changed identically in at least two schemes	89
Unique changes at 30 mg twice daily	142
Significant changes at 30 mg once daily	1

The drug suppressed the protein signature of cellular aging.

One of the most powerful mechanistic results was the analysis of proteins associated with cellular senescence. Senescent cells cease normal division but do not necessarily die. Instead, they can secrete a variety of inflammatory cytokines, growth factors, and enzymes that alter surrounding tissue. This secretory phenotype is considered one of the potential mechanisms of age-related pathology.

In the placebo group, the known SenMayo protein signature, which contains markers of senescent cells, increased over the course of the study. The normalized enrichment index was 1.48 with an adjusted p<0.01. In the treatment groups, the opposite trend was observed: most sets of cellular senescence proteins shifted toward lower levels.

A particularly broad effect was again observed at a dose of 30 mg twice daily: the core of the SenMayo signature included 22 reduced markers of cellular aging. Seven proteins were repetitive across all treatment groups: EREG, ESM1, IGFBP4, ITGA2, MMP10, MMP13, and SPP1.

The authors call this profile a senomorphic effect: the drug may not destroy senescent cells, but rather reduce their harmful molecular activity. This is consistent with earlier laboratory experiments with rentosertib, but the number of senescent cells was not directly measured in patients. Therefore, the serum signature is currently an indirect indicator.

The insulin-like growth factor and metabolic pathways were also altered.

An analysis of biological pathways revealed changes significantly broader than fibrosis alone. In total, the authors discovered dozens of regulated networks, with the 30 mg twice-daily regimen affecting the broadest set—approximately 40 biological pathways at the statistical significance threshold used in the analysis.

Both more active regimens suppressed extracellular matrix remodeling pathways, as expected for an antifibrotic drug. However, they simultaneously altered the MET, Gαs, and cell death receptor signaling pathways, as well as the transport and bioavailability of insulin-like growth factors.

The researchers paid special attention to insulin-like growth factor-binding proteins. IGFBP-1, IGFBP-4, IGFBP-6, IGFBP-7, IGFBPL1, and CCN1 were reduced in all treatment groups. Of these, CCN1, IGFBP4, and IGFBPL1 demonstrated the most consistent reductions across all three regimens.

This is of interest to gerontologists, as the insulin and insulin-like growth factor system is linked to the regulation of growth, metabolism, and lifespan in many species. However, the targeting of individual proteins in this system does not automatically lead to the conclusion that a "longevity program" has been activated in humans. Rather, the result points to a molecular pathway worthy of further investigation.

What biological processes were changed?

Process	Observed direction
Extracellular matrix	decreased remodeling
Fibrosis	reduction of a number of markers
Cellular aging	SenMayo signature reduction
Insulin-like growth factor system	pronounced restructuring
Fatty acid metabolism	increased activity of individual pathways at 30 mg 2 times/day
Glutathione metabolism	changes in the two-time regimen
Cholesterol metabolism	changes
Wnt signaling	especially at 60 mg once a day
Immune processes	changed in both active circuits

Why did the effect peak after four weeks and then stop increasing?

One unusual result was the time course. Most watches showed a maximal reduction in estimated age after four weeks, but by the twelfth week, the effect had ceased to increase. The number of statistically significant comparisons decreased.

At first glance, this could be interpreted as a waning of the drug's effect. However, a direct comparison of the fourth and twelfth weeks showed no significant age reversal for any combination of regimen and timing. Therefore, the authors use the term "plateau" rather than "loss of effect."

Even more interesting were the results for individual proteins. At 30 mg twice daily and 60 mg once daily, only 5-9% of protein changes were transient. Most continued to persist or develop until the twelfth week. Proteins with stable dynamics were particularly abundant in the aging clock itself.

A peculiar paradox emerges: proteins continue to change, but the final age indicator stops declining. This could indicate the achievement of a new physiological equilibrium, the body's adaptation, or simply a limitation of mathematical models that begin to compensate for each other as individual proteins continue to change. The study itself does not allow one to choose between these explanations.

"The age of arteries has decreased by 16 years" - why this figure shouldn't be taken literally

The authors additionally applied organ-specific versions of one of the models. The model related to the arterial system yielded particularly strong results: at all time points in the treatment groups, it showed an estimated age that was 6.95 to 16.57 years lower than the placebo. Separate signals were also detected for the brain, pancreas, stomach, and immune system.

This may sound like an incredible "rejuvenation of arteries by 16 years," but such a translation would be scientifically incorrect. The organ-specific clock analyzes a set of circulating proteins statistically associated with the health of a specific organ. The researchers did not perform arterial biopsies or directly measure their structural or functional age.

Furthermore, the organ-specific results represent the most exploratory part of the analysis and are based on an extremely small group. The more models and organs tested, the higher the likelihood of unstable estimates, even with statistical correction. Therefore, the range of -6.95 to -16.57 years is best interpreted as a strong molecular signal rather than a quantitative clinical effect.

Nevertheless, the result may suggest directions for future research. If future trials confirm the protein signal and simultaneously demonstrate improvements in arterial stiffness, endothelial function, or other independent vascular parameters, the interpretation will become significantly more convincing. There is currently no such confirmation.

Why research hasn't yet proven that human aging is slowing down

The first and most obvious limitation is the study's limited sample size—only 42 participants. Furthermore, these were not 42 healthy individuals, but rather patients with an average age of 67.1 years and severe chronic lung disease. Therefore, the results cannot be automatically generalized to healthy individuals of the same age.

The second limitation is the short duration. The treatment lasted only 12 weeks. Aging is a process that develops over decades, and the blood's molecular profile can change significantly within days or weeks. It's impossible to determine from three-month protein changes whether a person will experience a slower decline in physical function or live longer.

Third, all participants in the protein analysis were of Asian descent and were treated in China. Although the underlying biological mechanisms are universal, protein levels and disease risk may be influenced by genetics, diet, underlying conditions, and environment. Replication in other populations is necessary.

The fourth and most fundamental limitation is the overlap between fibrosis and aging at the level of the same proteins. For example, LTBP2, which significantly influenced several clocks, is also associated with fibrotic processes. Therefore, part of the "age-related" shift may be a therapeutic improvement in the disease, mistakenly interpreted by the algorithm as rejuvenation. The authors explicitly acknowledge that protein clocks alone cannot solve this problem.

Limitation	Why is it important?
Only 42 people	ratings may be unstable
All have pulmonary fibrosis	the disease itself changes the protein age
12-week observation	too little to assess longevity
Average age 67.1 years	cannot be transferred to other age groups
All participants are of Asian descent	limited tolerability
Biological clock - surrogate	are not equal to real rejuvenation
Lots of computational analysis	dependence on models increases
There are no epigenetic or other independent clocks.	there is no cross-platform confirmation
There are no healthy participants	it is impossible to separate the treatment of disease from the impact on aging

Another important limitation is that the study is partly tied to the drug's developer.

Several authors of the new paper work for Insilico Medicine, the company that developed rentosertib. Alex Zhavoronkov is the company's founder and CEO, and Fedor Galkin, Shan Chen, Feng Ren, Alex Aliper, Mikhail Durymanov, Denis Sidorenko, and Hui Cui are listed as employees.

A commercial conflict of interest in itself does not invalidate the results. The study was peer-reviewed by Nature Biotechnology, the data were analyzed using several independently developed clocks, and the authors published the computational tools and data for verification. However, the developer's connection to the study requires particularly careful interpretation of the lofty claims about "reversing biological age."

It's telling that the scientific article itself is worded much more cautiously than some press releases. The authors speak of a reduction in predicted biological age and "potential geroprotective effects," rather than proven rejuvenation or life extension. They explicitly cite the small sample size, short follow-up period, predominance of computational methods, and the inability to distinguish between anti-fibrotic and anti-aging effects.

This caveat is especially important for aging research, where the attractive units of "years of biological age" can easily be turned into overly strong consumer inferences. For clinical gerontology, what ultimately matters is not the number of hours per se, but rather healthy lifespan, physical function, disease, disability, and mortality.

Why this work is still important for aging research

Despite its limitations, the study offers an interesting methodological insight. Typically, a new drug is tested against a specific disease, but its potential impact on aging is studied years or even decades later. Metformin and rapamycin are good examples of drugs whose geroprotective properties attracted interest long after their initial use.

The authors propose a different approach: collecting molecular data during the early stages of trials of age-associated diseases, allowing for the simultaneous assessment of the underlying disease and the biological signs of aging. This way, the potential geroprotective properties of a new drug could be identified even before its approval.

Protein analysis is particularly convenient for this task, as it requires only a blood sample and can be repeated multiple times during the study. In the current study, this allowed us to observe changes after two and four weeks and then compare them with the twelfth week.

But for this approach to become a true tool for developing anti-aging drugs, the clock itself must be reliably linked to clinical outcomes. It must be proven that a drug that reduces protein age actually reduces the risk of age-related diseases or preserves bodily functions. This is precisely the missing link in the chain of evidence.

What to check next

The first necessary step is to replicate the analysis in a much larger clinical trial. Rentosertib has already entered later stages of development for pulmonary fibrosis, which will potentially allow testing of protein indicators in a significantly larger number of patients. However, to achieve this, it would be advisable to include gerontological endpoints in the protocol upfront.

The second step is to use multiple independent biological layers. In addition to proteins, it's possible to measure the epigenetic clock, metabolites, lipids, blood cell transcriptome, and markers of immune aging. If completely different methods simultaneously show the same direction, the likelihood of artifact is reduced.

The third step is to incorporate direct functional indicators. Physical performance, muscle strength, vascular function, cognitive parameters, and clinical signs of frailty can be assessed. This will help us understand whether the shift in the molecular clock corresponds to a real improvement in bodily function.

Finally, to prove its geroprotective effect, the drug will need to be studied beyond severe pulmonary fibrosis. The authors themselves note that definitively separating the anti-fibrotic and anti-aging effects will require more direct studies, including other populations and tissue-based methods. Until then, rentosertib should be considered a candidate with a potential gerontological signal, not an anti-aging drug.

The main conclusion

A new study in Nature Biotechnology shows that rentosertib can significantly alter the blood protein profile of patients with idiopathic pulmonary fibrosis within a few weeks. All six independently developed aging models showed a shift toward a younger profile, with the most consistent effect observed after four weeks at a dose of 30 mg twice daily.

In the most sensitive models, the change was approximately 3-4 years, and in one, approximately six years. At 60 mg once daily, four chronological models showed a reduction of 2.71-3.46 years. However, this refers to the estimated protein age, not a proven return to a state of being several years younger.

The mechanistic data make the results more compelling than a simple mathematical estimate: rentosertib altered 326 proteins, counteracted age-related protein pathways from the UK Biobank, reduced cellular aging signatures, and reorganized insulin-like growth factor, metabolic, and extracellular matrix pathways. However, the improvement in FVC explained, on average, only about 6% of the variation in biological age change.

But the evidence remains early. The study included only 42 patients, lasted 12 weeks, and cannot definitively separate fibrosis reduction from true slowing of aging. Therefore, the correct conclusion is not that rentosertib "rejuvenated people by four years," but rather that the same drug induced a coordinated shift in several independent protein markers of aging and provided a strong enough signal to specifically test its geroprotective effect in future studies.

News source

Zhavoronkov A., Galkin F., Chen S., Ren F., Aliper A., Durymanov M., Sidorenko D., Cui H., Han J.-DJ, Xu H., Liu X., Xu Z., Kuppe C., Argentieri MA, Ying K., Goeminne LJE, Moqri M., Tyshkovskiy A., Gladyshev VN Integration of proteomic aging clocks in a phase 2a clinical trial supports simultaneous geroprotective assessment. Nature Biotechnology. Published online September 7, 2026. DOI: 10.1038/s41587-026-03286-y.

This is a secondary exploratory protein analysis of a previously conducted randomized, double-blind, placebo-controlled Phase IIa clinical trial of rentosertib. The analysis included 42 participants, in whom 2,841 serum proteins were analyzed at four time points and six biological age models were applied.

The main publication of the clinical trial of rentosertib itself:

Xu Z., Ren F., Wang P. et al. A generative AI-discovered TNIK inhibitor for idiopathic pulmonary fibrosis: a randomized phase 2a trial. Nature Medicine. 2025;31:2602-2610.DOI: 10.1038/s41591-025-03743-2. This study included 71 patients; the maximum lung function signal was observed with 60 mg rentosertib once daily: the mean FVC increased by 98.4 mL, whereas in the placebo group it decreased by 20.3 mL.