Staff of the Yale ILD Center of Excellence, where many of the co-authors work, including Yale co-authors Naftali Kaminski and Yu Guoying. (Source: Naftali Kaminski)Lethal fibrosis in lungs of idiopathic pulmonary fibrosis (IPF) mouse models can be reversed, say Yale University researchers.

No drug on the market can do this. But the Yale crew pulled it off, in mice, by temporarily restoring (a mimic of) one of the body’s own anti-fibrosis agents, sharply reduced in IPF: microRNA-29. A five-year, $1.4-million-a-year National Institutes of Health (NIH) CADET grant was awarded just after the EMBO Molecular Medicine report was published.

The purpose: move to clinic fast.

“We are very excited,” said co-senior author Yale pulmonologist Naftali Kaminski, by email to Bioscience. “Initially most people used their inhibitor either as a preventive, or disease-modifying early in injury. We waited until fibrosis was established at day 10, after bleomycin [created fibrosis], then gave miR-29 [a microRNA-29 mimic]. We saw significant reversal of fibrosis compared to mock-treated animals.”

MiR-29, a master regulator, inhibits several fibrosis pathways at once. This paper, with others, may change IPF therapy, some feel.

“It challenges our perspective,” Kaminski said. “Lung fibrosis is not simple scar tissue. It is a dynamic process of lung remodeling with very active molecular networks— all my work over the years suggests it. Inhibiting key regulatory events may indeed reverse it.”

The mouse model used— the standard— is poor, Brigham & Women’s Hospital pulmonologist Gary Hunninghake (uninvolved in the study) wrote in an email to Bioscience. But, he said, the study describes a “promising approach to treat pulmonary fibrosis. There is evidence down-regulation of the miR-29 family might play an important role in the pathogenesis of pulmonary fibrosis. This paper demonstrates a molecular mimic of the miR-29 family could be more durable (than the human miR-29), gets into the lung, and may result in a reduction in collagen expression likely important to the pathogenesis of this disease.”

Said University of Minnesota pulmonologist Peter Bitterman, also uninvolved in the study, via email: “Against a background of hopelessness born from decades of negative clinical trials, for the first time there is enthusiasm. We have positive clinical trials for two investigational agents (pirfenidone and nintedanib) that slow IPF progression based on their ability to correct aberrant signaling in fibrogenic lung myofibroblasts. Adding to the enthusiasm, this report establishes that microRNA mimicry offers a promising new therapeutic direction in pulmonary fibrosis based on a completely different principle: correcting pathological expression of the genes that actually mediate fibrosis.”

“Highly promising,” he said.

Slowly unraveling the IPF mystery

IPF kills two to five years post-diagnosis. It is oft called a disease of “irreversible” scarring, making leather of lungs. But recent data suggest the fibrosis found in IPF is dynamic, a living knot that may be untied.

Cedars-Sinai Medical Center pulmonologist Cory Hogaboam decries the new study’s bleomycin-IPF mouse model as one of simplified fibrosis. But, he said: “I agree the earlier idea of fibrosis was that it is irreversible, not a dynamic process of build-up/break-down. The collagen matrix is not static, but full of turnover, dramatic change. I agree we are coming around to this.”

The field is also abuzz with the idea that IPF fibrosis is caused more by aberrant wound healing than chronic inflammation. This gelled when the triple-therapy arm of a trial targeting IPF inflammation— prednisone, azathioprine, and acetylcysteine— halted early, in 2012, as patients on drugs died sooner than controls. This summer acetylcysteine alone was found ineffective.

But also this summer, the NEJM described trials of two drugs, pirfenidone and nintedanib, which target aberrant wound healing over inflammation. They had side effects, and didn’t reverse fibrosis, but they did slow it to an unprecedented degree. “New hope,” said an NEJM editorial.

Co-discoverer of microRNA, Victor Ambros, who is on the Scientific Advisory Board of MiRagen, the company behind this miR-29 mimic. (Source: Wikimedia/ Jane Gitschier)miRagen

Coming from another direction, in 2008 UT Southwestern Medical Center cardiologist (and miR-29 mimic study co-senior author) Eva van Rooij reported, with pioneering mouse modeler geneticist Eric Olson, that miR-29 is down-regulated in cardiac fibrosis post heart attack. Later work found miR-29 down-regulated in liver and lung fibrosis.

Kaminski teamed with van Rooij, then with start-up miRagen, to try replacing miR-29 in lungs.

MiRagen formed in 2007 around Olson’s miR-208 mouse, the first miRNA knockout. Watching different miRNAs change post-heart attack, “we picked up miR-29,” said miRagen CEO William Marshall to Drug via interview. “Amazingly, we did bioinformatics to predict what genes miR-29 regulates and saw the predicted gene set was regulated by miR-29, and comprised of every important ECM factor with major implications in fibrosis. Our eureka moment. We filed the IP.”

An early surprise came upon injecting a miR-29b mimic into mouse tails. It migrated largely to the lung, staying two to four days. “Serendipity,” Marshall said.

The Yale/miRagen team injected the mimic on day three and 10, post bleomycin. By day 14, control mice reduced miR-29 levels comparable to IPF-patient levels. Treated mice saw increases. “Robust” inflammatory and fibrotic responses were “prevented,” the report said.

The group next sought to see if the miR-29 mimic affected “established fibrosis” on days 10, 14 and 17, post bleomycin. It “normalized” levels of fibrotic collagens in IPF mice, without harming controls. Collagen-associated hydroxyproline levels were “blunted.” Collagen growth was blocked in vitro human cells, too. "Importantly, the anti-fibrotic effect was observed even when the mimic was administered after fibrosis was already established," says Bitterman.

This echoed a study finding miR-29, given via Sleeping Beauty (SB) DNA integration, blocked fibrosis— if did not roll it back. “Very similar,” said Kaminski. “In two different methods, using different bleomycin doses, miR-29 is effective against established fibrosis. Administering SB-based miR-29 gene delivery on day 14 (established fibrosis), you do not see the same increase in fibrosis on day 28 as in untreated animals.” Administering miR-29 mimic at day 10 and 17— harvesting at day 21— “we see less fibrosis compared to animals treated with control mimic, and quite similar to that of saline-treated animals (same age, after 21 days of saline).” Kaminski said.

He finds it “hard to say” why SB delivery did not reverse fibrosis. Methods differed. But the paper offers “strong evidence [that] increasing miR-29 can blunt fibrosis. How frequently do we see two completely independent interventions in the same pathway show very similar results?”

Some problems, much promise

Some note other miRNAs are up- or down-regulated in IPF. Kaminski responds: “Probably all the data on microRNAs differentially expressed in human IPF lung” come from his group’s work on let-7, miR-154, miR-21 and more. “We are excited about miR-29 because it seems a transcriptional regulator of extracellular matrix (ECM) proteins, and signaling molecules (IGF, PDGF), based on our work and others. Also fascinating is a recent paper showing miR-29 was the only microRNA family to regulate shift of fibroblasts into a ‘fibrotic’ phenotype when exposed to IPF lung ECM proteins.”

The field should embrace a “core pathways” paradigm, he said. “Data for miR-29 is very impressive on evidence for down-regulation in liver, lung, kidney, heart, peritoneum, skin in humans, and animal models. It is as good as any target, except maybe the AVB6 integrin studied extensively over 16 years.” As miR-29 targeting is new, “the evidence is very, very impressive.”

Still, there is “growing evidence the bleomycin mouse model has limited correlations to established chronic fibrotic disease in older patients,” said Hunninghake. Agreed Hogaboam: “Many regenerative mechanisms in mouse lung don’t appear present in elderly patient populations with irreversible scarring,” adding that he thought a humanized model should be considered. But Kaminski feels “the most important next step is a second model of fibrosis, and perhaps another route of administration.”

Hogaboam would like to see more tests for human fibroblast fibrosis markers like the collagens, smooth-muscle actin, and fibronectin, and profibrotic signals including CTGF, IL-13, and hypomethylated DNA (i.e. CpG). Kaminski said there is “significant literature” on miR-29 targets, “extensive, convincing, and beyond the paper’s focus. We presented the down-regulation of collagens by the mimic in primary fibroblasts from patients, and its dose-dependence at baseline, and in response to TGFb, as another way to illustrate relevance to human disease. But as so many others have shown multiple fibrotic targets for miR–29,” there was no need to repeat all, and CTGF and IL-13 are not part of fibrotic lung phenotyping. It will be “interesting” to see if miR–29 affects epigenetic regulators like hypomethylated DNA, but that can be pursued later.

Importantly, Hogaboam’s work shows that endogenous mechanisms needed to activate miR-29 can degrade in IPF. Biopsies to identify them may be key, he said. Kaminski agrees. Pre-dosing, “we should have evidence the pathway is aberrant in the patient. No IPF drugs have such biomarkers.” He asked an IPF meeting to develop them. In his CADET grant, he plans “significant efforts” to validate some.

Hogaboam said the immune system could reject RNA mimics. There is no evidence yet, said Kaminski. “But we will carefully assess.”

Replacing miRNAs— “putting nature’s pathway back in place”— is safer than inhibiting them, Marshall believes. Hunninghake and Bitterman warn of side effects. But Hunninghake also believes the data, alongside “evidence for the role of miR-29 in human disease, and very promising findings of miRNA targeted therapy in HCV patients, suggests this is a line of therapeutics worth pursuing.”

Said Bitterman: “The results establish proof of concept for miR-29 mimics as potential therapeutics to prevent, and reverse, lung fibrosis, and perhaps other forms of tissue fibrosis in settings where endogenous miR-29 is reduced.”

A new place

IPF is in a new place. Pirfenidone and nintedanib, nearing FDA-approval, “have a real effect,” Hogaboam said. (UPDATE: FDA approval was officially granted Oct. 15. See accompanying Drug Discovery & Development story.) Another CADET grant aims at speeding to trial an antibody to a protein downstream of miR-29, CHI3L1, ubiquitous in IPF. (See Bioscience.) Two more interesting drugs, Hogaboam said, are Gilead's Simtuzumab (targeting LoxL2) and Biogen Idec's STX-100 (targeting alpha-beta-6 integrin), both in Phase 2 trial. And high-resolution CT scans will bring earlier diagnosis, key as fibrosis is more dynamic— potentially reversible— earlier. Hogaboam said the question may be “if the patient has time to turn over matrix to normal function.”

Kaminski is optimistic. “I was part of this paradigm shift in the previous decade. When we did gene expression microarrays of IPF lungs, we did not see significant inflammation. And since then, the role of TGFb activation, and changes in epithelial cell and fibroblast phenotypes, matrix metalloproteases, and developmental pathways, have all been demonstrated independent of inflammation."

At last, he said, "we have strong data grounded in findings from human lungs. For somebody who, over 12 years, has been doing discovery science, it is very exciting to engage in implementation.”

Proof-of-principle trials of miR-29—for post-surgical cutaneous fibrosis, followed by scleroderma—may start in mid-2015, said Marshall.