The New Liver Epidemic (Nature Biotechnology)

NASH has replaced hepatitis C as the most important nonalcoholic liver disease. Biotech and pharma are racing to tap into this enormous market.

March 2019

The first approved drug for nonalcoholic steatohepatitis (NASH), a potentially fatal liver disease that affects at least 5 million people in the United States alone, should arrive in 2019. By March’s end, Intercept Pharmaceuticals and Gilead Pharmaceuticals will release phase 3 trial data for their respective NASH drugs, with regulatory filings likely to quickly follow. Genfit and Allergan are close behind. These companies are competing for a huge market opportunity. “You can get a very large sales numbers when you have that many patients,” says Alan Carr, a biotech analyst for Needham, who says it’s too early to put a dollar value on the potential market: “It will be one of the larger indications, in terms of sales, in healthcare.”

NASH is an epidemic. The disease, long thought unimportant by the medical community, is set to overtake hepatitis C as the largest single cause of liver transplantation. NASH afflicts 2–5% of North Americans, and its broader entity, nonalcoholic fatty liver disease (NAFLD), around 30%. NASH is fueled by the general rise in obesity and diabetes but distinct in its pathophysiology. New treatments are badly needed.

All of the phase 3 drugs (Table 1), plus several in phase 2, have demonstrated efficacy. Yet none works in more than half of NASH patients (on the basis of phase 2 data), with complete resolution in only a subset of responders. “We’ve gotten used to amazing successes in viral hepatitis, where we get nearly 100% response in hepatitis B or cure in hepatitis C. I don’t know that we’ll ever get that in metabolic liver disease,” says Yaron Rotman, a hepatologist at the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK). “We may have to settle for reasonable control of the disease, as opposed to perfect resolution.” Some companies are already positioning themselves for combination therapy, hoping to boost drug response and disease resolution rates. So there may be many winners in the current race for FDA approval. “There is room for several of these [drugs] to succeed,” says Carr.

Anatomy of a killer

NASH is a relative newcomer of a disease, first named in 1980 by Jurgen Ludwig at the Mayo Clinic in Rochester, Minnesota. It starts with steatosis, an excess of fat in the liver, and once hepatocyte injury and inflammation appear (in the absence of alcoholism), you have NASH. This often progresses to fibrosis and then to cirrhosis, with end-stage liver disease a fatal outcome, absent a liver transplant. (NASH patients are also at elevated risk for liver cancer.)

At first, many doctors did not recognize fatty liver disease as separate from obesity and diabetes. The diseases are intertwined, but 30% of diabetics do not have fatty livers and not all NASH patients have diabetes. The known genetic risk factors for NASH were all validated controlling for diabetes and body mass index, and so are specific to NASH. Finally, African-Americans are relatively protected from fatty liver disease, unlike diabetes. So NASH is a distinct disease.

But it’s not yet fully understood. The key questions are these: what causes fat to build up in the liver, and why does it cause liver injury, inflammation and fibrosis? The answer to the first question is straightforward: insulin resistance impairs signaling to the liver in the ‘fed state’, so the body mistakenly seeks to mobilize fat for energy. It directs fatty acids from fat tissue into the liver for conversion to triglycerides and also triggers de novo fatty acid synthesis there. Triglyceride export from liver and fatty acid oxidation in mitochondria there both lag, and so fatty acids and triglycerides accumulate in the liver—steatosis.

What causes liver injury, inflammation and disease progression is less settled. The established ‘two-hit’ model, which postulated that liver fat buildup made hepatocytes vulnerable to injury from second hit, be it oxidative stress, cytokines or bacterial products, is now discredited. “The two-hit model was a nice story, but probably somewhat simplistic,” says Rotman. “The current story is a focus on lipotoxicity. Either free fatty acids or nonesterified free fatty acids, or some of their toxic metabolites, are driving hepatocyte injury.” For example, Gerald Shulman at Yale has described one such metabolite, diacylglycerol, which is an intermediate in fatty acid biosynthesis that builds up in fatty livers, as a signaling molecule driving insulin resistance and NAFLD1.

Such signaling “can lead for instance to inflammation and mitochondrial dysfunction,” says Bart Staels, a metabolic disease researcher at the University of Lille in France. Certain patients, overburdened with disease or genetically predisposed, Staels explains, then develop NASH. “Patients who have steatosis that move to NASH progress to what I call often a state of metabolic inflexibility,” he says. “They’re unable to compensate.” Mitochondria malfunction, reactive oxygen species bathe cells, hepatocytes suffer injury, inflammation flares: in other words, NASH. And injury and inflammation activate pathways driving liver fibrosis in up to half of all patients.

Some effective treatment options already exist. In the phase 3 PIVENS trial, sponsored by the NIDDK, 43% of NASH patients improved histologically on vitamin E, and 34% responded to Takeda’s Actos (pioglitazone), versus 19% on placebo2. Vitamin E, an antioxidant, protects against metabolic stress (implicated in NASH), whereas Actos is a peroxisome proliferator-activated receptor-g (PPARg) agonist. PPARg, a nuclear receptor, promotes lipid storage in fat cells, so Actos presumably improves NASH by keeping fatty acids from being sent to the liver to accumulate there instead. (In the liver, PPARg also boosts insulin sensitivity.) The downside is that PPARg agonists like Actos cause weight gain. “Patients don’t like that, and pioglitazone in general had gotten some really bad press in terms of potential cardiac side effects,” says University of Michigan hepatologist Monica Konerman. “So patients in general are pretty wary.” As for vitamin E, the PIVENS trial excluded patients with diabetes or cirrhosis, so no one knows if such NASH patients—the majority—would benefit.

Another option, proven to be effective, is lifestyle intervention. A 2005 pilot study showed a 60% histological response rate after a year of nutritional counseling3, and a 2010 randomized trial achieved a 67% response rate following weight loss, compared with 20% in the placebo group4. “If we can get patients, especially patients with early-stage disease, to improve their eating habits and be more physically active, we can cure [NASH],” says Konerman. Even liver fibrosis can be reversed, she adds. But “lifestyle intervention is an extremely difficult thing to maintain,” says Rotman. Hence the search for a pill.

The bile acid solution

Origins for the current wave of NASH drugs date back to a 1981 faculty party in La Jolla, California. There Roberto Pellicciari, a visiting professor of organic chemistry from the University of Perugia in Italy, first met University of California, San Diego gastroenterologist Alan Hofmann. Hofmann was one of the world’s few experts on bile acids. It was long known that the liver makes bile acids from cholesterol and secretes them via the gallbladder into the gut, where they help the absorption of fat from food. But Hofmann suspected that bile acids played a wider role. “He said, ‘Roberto, no one wants to work on bile acids,’” recalls Pellicciari. “‘One day a receptor will be discovered, that’s my prediction. Then it will become a financial prospect.’” Pellicciari returned to Perugia convinced Hofmann was right. His lab developed several bile acid analogs, then left the field around 1993 as scientific interest in bile acids continued to languish.

All that changed in 1999, when three separate groups, including one from Glaxo Wellcome in Research Triangle Park, North Carolina, reported that the farnesoid X receptor (FXR) was a bile acid receptor5, fulfilling Hofmann’s prophecy. The discovery “was a real renaissance of work on bile acids as signaling molecules,” says Pellicciari. He called up Glaxo Wellcome (now GlaxoSmithKline; GSK) and set up a collaboration. Within just a few months, in December 1999, Pellicciari’s lab in Perugia synthesized obeticholic acid (Ocaliva), a synthetic bile acid derivative that bound FXR 100 times more potently than natural bile acids. That’s critical, says Pellicciari, because natural bile acids only activate FXR when they reach toxic levels, but Ocaliva does so at tolerable dosages.

Meanwhile, GSK was working out the biological function of FXR, using its own GSK FXR agonist as a biological probe. Gradually, it emerged that FXR activates fibroblast growth factor 19 (FGF19), a secreted growth factor that, once in the liver, suppresses cytochrome P450 7A1 (CYP7A1), the rate-limiting enzyme in bile acid synthesis6. That is beneficial because excess bile acids, which accumulate in NASH, can damage liver cells via their detergent action. Beyond reducing liver bile acids, FXR activation and FGF19 secretion have other positive effects: insulin sensitivity goes up, glucose production in the liver goes down, and fewer fatty acids get made because FGF19 downregulates the transcription factor SREBP-1c (sterol regulatory element binding protein 1c), a master regulator of fatty acid synthesis.

For an Ocaliva disease application, Pellicciari settled on primary biliary cholangitis (PBC), a relatively rare disease involving progressive destruction of the bile ducts. GSK soon bowed out. “They called me one day and said, ‘Roberto, you know, PBC is a small disease, we’re not interested,’” recalls Pellicciari. “So they gave back the patent to the University of Perugia.”

Pellicciari, to take Ocaliva forward, cofounded Intercept Pharmaceuticals in New York in 2002 with Mark Pruzanski, a Canadian doctor with extensive business experience who became CEO. Intercept licensed the Perugia patent and launched phase 2 trials in PBC in 2007. The FDA approved Ocaliva for PBC in 2016.

More consequentially, for Intercept and for the whole NASH field, in 2010 the NIDDK sponsored Ocaliva’s phase 2 FLINT trial in NASH. The trial was stopped early for efficacy in 2014, with 45% of Ocaliva-treated patients showing histological improvement, versus 21% on placebo7. Ocaliva also led to an improvement in fibrosis scores in 35% of patients—the first NASH drug to demonstrate a significant antifibrotic effect. Intercept’s stock nearly quadrupled in a single day.

Intercept’s stock spike set off a frenzy of investor interest in NASH, and pharma and biotech activity in the area exploded. A Needham report lists six other FXR agonists alone now in clinical development for the disease, and 63 drugs overall from 45 companies. Many others are in preclinical development.

Ocaliva remains the favorite in the race for first FDA approval, on the strength of the FLINT trial results showing both NASH resolution and fibrosis improvement. “Oca as an FXR agonist continues to be the only investigational NASH drug that has shown efficacy across the board,” said Pruzanski, still the company’s CEO, in an October conference call.

But Ocaliva isn’t the perfect NASH drug. In phase 2, it paradoxically caused an increase in HOMA-IR (Homeostatic Model Assessment of Insulin Resistance), a measure of hepatic insulin resistance, and it also raised low-density lipoprotein (LDL) and lowered high-density lipoprotein (HDL)—the wrong direction for serum lipids to go. (Because cholesterol is no longer metabolized to bile acids, presumably it’s shunted to the bloodstream.) And 23% of patients experienced itching, a common symptom of bile acid disorders, in some cases so severe that treatment had to be interrupted. The drug’s reputation also suffered when the FDA in 2017 reported 19 deaths of PBC patients on Ocaliva due to liver toxicity and 11 other cases of liver injury from the drug, typically after being dosed daily instead of twice-weekly or weekly as recommended. “That boiled down to physicians not following directions in the label,” says Carr, who says he isn’t concerned by the number of deaths. But there are also background worries about malignancy because FGF19 (downstream of FXR) binds the fibroblast growth factor receptor 4 (FGFR4), whose activation promotes cell proliferation and, in certain rodent models, cancer.

Intercept plans to minimize Ocaliva’s side effects by possibly titrating the dose gradually upwards in patients, to minimize itching, and by treating patients with statins to address LDL elevation. But whether the drug’s benefits exceed its problems enough to justify FDA approval awaits phase 3 results and Intercept’s subsequent New Drug Application (NDA) filing. The phase 3 REGENERATE trial is set to read out in the first quarter of this year.

Healing liver injury

So will Gilead’s first phase 3 trial of its selonsertib, an oral ATP-competitive inhibitor of apoptosis signal-regulating kinase-1 (ASK-1). Selonsertib targets a key mechanism of liver injury in NASH. In 1997, a Japanese group discovered ASK-1 as an activator of the p38 and JNK signaling pathways. Oxidative stress in NASH triggers ASK-1 signaling. This happens not just in hepatocytes (to induce cell death), but also in hepatic stellate cells, contractile cells that lay down collagen in response to liver injury, and in inflammatory cells, where ASK-1 boosts cytokine and chemokine production. “Activation of this [ASK-1] pathway really affects all three of those cell types and contributes to liver injury,” says Rob Myers, Gilead’s VP of clinical research.

But ASK-1 was a bold target choice because p38 inhibition has been a graveyard for pharma and biotech in diseases like rheumatoid arthritis, mainly due to intolerable toxicity. At least 22 p38 inhibitors have failed clinical trials in the years leading up to Gilead’s selonsertib program. Why should ASK-1 inhibitors do any better? “It’s really a function of the fact that p38 is regulated by multiple kinases, not just ASK-1,” says Myers. Selonsertib, Gilead found, completely inhibits the active form of p38 in liver, where there’s oxidative stress, but not in blood, where presumably some other kinase activates p38 in the absence of ASK-1. The theory is that p38 in liver is uniquely dependent on ASK-1 for its activation, whereas p38 elsewhere in the body remains active, thus limiting on-target side effects. Formal proof for this is lacking, but clinical evidence is reassuring.

Selonsertib, in a small randomized phase 2 trial8, was well-tolerated if not totally benign, with 5 of 62 patients overall suffering serious adverse events, versus none on the surrogate placebo arm. Gilead has tested selonsertib on close to 600 other patients in other indications, and “there was really no hallmark toxicity of the compound across the program,” says Myers. The drug appeared effective, with 13 of 30 patients (43%) in the high-dose group showing an improvement in NASH histology, compared with 2 of 10 on simtuzumab, a lysyl oxidase-like 2 (LOXL2) inhibitor. The trial did not include a placebo arm, but Gilead considers simtuzumab a de facto placebo because it has since shown no efficacy in NASH. These results suggest selonsertib has a positive effect on fibrosis, but a few patients progressed to cirrhosis. There was no improvement in liver fat content or insulin resistance (as expected, given the drug’s downstream mechanism of action).

But at this point Carr thinks Intercept’s Ocaliva leads selonsertib in the approval race. “There’s more to work with,” he says. “They [Intercept] have a completed placebo-controlled trial that demonstrates benefits for both inflammation and fibrosis. That’s an important distinguishing accomplishment. And there is a commercial track record—it is on the market for another indication.”

The contenders

Next in line in phase 3, after Ocaliva and selonsertib, are Genfit’s elafibranor and Allergan’s cenicriviroc. These trials should read out by year’s end. Elafibranor is a small-molecule agonist of the PPARa and PPARd nuclear receptors. Elafibranor by design does not activate PPARg, the target of Actos. That avoids the weight gain and other side effects that have plagued the entire PPARg activator drug class, known as thiazolidinediones. With elafibranor, “we have never seen a signal of PPARg activation, which has put so much stigma on the whole class of PPARs,” says Bart Staels, who co-founded Genfit with Florence Séjourné, Jean-Charles Fourchart and Jean-François Mouney in 1999.

PPARg mainly promotes lipid storage in adipose tissue, whereas PPARa regulates the fasting response in the liver. The latter is activated by fatty acids and drives fatty acid oxidation and elimination, which is helpful in NASH, where fatty acids that overload the liver generate their toxic metabolites. Stael’s group has shown that PPARa also has anti-inflammatory effects independent of its metabolic action9 and can reverse liver fibrosis in mice. Targeting PPARd is also beneficial because in the liver it stimulates glucose utilization and inhibits gluconeogenesis, which improves insulin resistance. “There is a complementary function of PPARa and d,” says Staels. Genfit originally tried elafibranor in metabolic syndrome, but discovered the drug’s liver protective effects in patients (and preclinical models) and shifted to NASH.

In a randomized phase 2 trial10, elafibranor missed its primary endpoint, NASH resolution without worsening of fibrosis. But that result may be misleading because most patients had early-stage NASH. “If you have less severe disease, that leaves you less room to show improvement,” says Carr. And the trial used an outdated definition of NASH resolution. As now defined by regulatory agencies, and only counting advanced cases of NASH, elafibranor was superior to placebo by 19% to 9%. And the overall histological response rate in these patients was 48%, compared with 20% for placebo. The drug improved markers of insulin resistance and glucose homeostasis, and was well tolerated, albeit with an increase in serum creatinine typically associated with PPARa agonists.

But until the phase 3 trial reads out, elafibranor remains a question mark. “The drug was having some sort of effect,” says Carr. “But … we didn’t have the same kind of rigorous trial and outcome as what we saw with Intercept’s trial.”

Allergan’s cenicriviroc has a completely different mechanism of action. Unlike the other three drugs, cenicriviroc, an inhibitor of the chemokines CCR2 (C-C motif chemokine receptor 2) and CCR5, targets downstream events in NASH to hopefully prevent fibrosis from progressing to liver scarring. Allergan acquired cenicriviroc when it bought Tobira Therapeutics in 2016 for $600 million up front plus $1.1 billion in future considerations . Tobira originally developed cenicriviroc for HIV infection—CCR5 is (among other things) the receptor for entry of HIV-1 into immune cells—but a favorable effect on fibrosis risk scores and inflammation markers prompted a switch to NASH.

A biological rationale does exist. In NASH, signaling though CCR2 and CCR5 in immune cells orchestrates the recruitment and activation of monocytes and lymphocytes, respectively. This drives the process that causes hepatocyte injury to progress to inflammation and fibrosis. CCR5 is also expressed on stellate cells—bad actors in fibrosis.

Cenicriviroc’s randomized phase 2 trial11 did not meet its primary endpoint, NASH improvement with no worsening of fibrosis, because the drug didn’t significantly improve NASH histology. Tobira’s stock plunged 60% following the announcement. However, 20% of patients achieved an improvement in fibrosis without worsening of NASH, compared with 10% on placebo. On the basis of these data, Allergan, after acquiring Tobira, advanced cenicriviroc into phase 3 last year for NASH patients with fibrosis, setting fibrosis improvement as the primary endpoint.

One anomaly in phase 2 was that the drug reversed fibrosis without improving histological markers of inflammation, which is puzzling given cenicriviroc’s anti-inflammatory mechanism and the general presumption that liver inflammation drives fibrosis. That “peculiar outcome,” in Carr’s words, “is somewhat of a question mark around what will happen in phase 3.”

Go in early or late?

Beyond the first four targets—FXR, ASK-1, PPARa/d and CCR2/CCR5—many others are the subject of advanced clinical trials. There are at least 38 agents in phase 2 for NAFLD or NASH, according to Needham’s report. Many are FXR agonists (Box 1). Also, multiple companies are targeting an enzyme that genetic studies recently revealed is critical in NASH (Box 2). The big divide in industry is over whether it’s better to intervene early or late in the disease process. “We strongly believe that metabolic dysfunction caused by diabetes and obesity in the early stages are actually key drivers associated with NASH,” says Marcus Schindler, senior vice president, global drug discovery at Novo Nordisk. “While early biology is at play, it might be a good idea to intervene then.” Novo Nordisk is doing this in NASH. The company started with Victoza (liraglutide), a glucagon-like peptide-1 (GLP-1) analog first approved for diabetes 10 years ago. GLP-1 is an incretin, a peptide hormone released by the gut in response to a meal—a kind of satiety signal—with a variety of beneficial effects. These include boosting insulin secretion, and indirectly slowing liver glucose production, as well as improving pancreatic beta cell function. And patients lose weight. In NASH, “improving the overall metabolic state—diabetic—and weight loss activity will have a beneficial effect,” says Schindler. In a small randomized phase 2 trial12, Victoza treatment did work: it led to NASH resolution in 9 of 23 patients, compared with 2 of 22 on placebo. But fibrosis worsened in two Victoza patients. Novo Nordisk is instead advancing its next-generation GLP-1 analog, Ozempic (semaglutide, FDA approved in 2017 for diabetes), now in phase 2 for NASH. Dosing is weekly injection, and an oral formulation is under development. “Semaglutide is a more powerful molecule,” says Schindler. “We need to hit the system hard.”

“If Novo moves forward with an oral formulation, physicians are very interested,” says Needham’s Carr, who has interviewed many liver specialists. “Some physicians were thinking that even with a GLP-1 not being approved for NASH, they may use it off-label, as a frontline option.” The rationale, says Carr, is that reducing weight up front will improve liver pathology. “To deal with NASH,” he says, “it’s best to deal with the source of metabolic disease, right? Obesity.”

Others argue that targeting late-stage NASH, especially fibrosis, is best. The calculus is simple: reverse fibrosis and you prevent cirrhosis and liver failure. “We are very focused on the patients who have advanced fibrosis … because we know those are the patients who have complications of the disease,” says Gilead’s Myers. “When you look at patients with NASH, fibrosis is the most important determinant of outcome.” But reversing fibrosis with a drug so far has been harder to achieve than NASH improvement (reducing hepatocyte injury and inflammation). Nevertheless, more than a third of patients taking Gilead’s selonsertib did, in phase 2, show fibrosis reduction8.

But companies ultimately hope to tackle all NASH’s features—fatty liver, cell death, inflammation and the fibrosis that follows—by combining drugs. “There will probably be a variety of treatments and combinations,” says Schindler. Gilead is the leader so far, testing its three in-house drugs together in phase 2. But each company is positioning its lead drug as the basis for future combination therapy. “We are laser focused on establishing OCA [Ocaliva] as a backbone treatment for … NASH patients with fibrosis, to which other compounds with different mechanisms might eventually be added,” said Mark Pruzanski in Intercept’s October conference call. Staels notes Genfit’s elafibranor has backbone potential because of its “excellent safety profile, better in a way than OCA.”

But can combinations cut into the roughly 60% of patients who do not respond to the current drug candidates, or are these patients refractory to all treatments, making combinations useless? “Once we have acceptable data from the phase 3 studies, once we have medications on the market, the next step will be to figure out who are the nonresponders and are they universally nonresponders,” says Rotman. He hopes they’re not, and that combining drugs with different mechanisms of action will have an additive if not synergistic effect. But that’s not guaranteed.

Approaching the finish line

In the meantime, liver specialists expect to see at least one FDA drug approval by late 2019, and maybe two, says the University of Michigan’s Konerman. Europe, though, is a bigger question mark. That’s because the European Medicines Agency (EMA) may require both NASH resolution and fibrosis improvement in phase 3, not merely one or the other, to merit approval. The EMA issued a ‘reflection paper’ on the endpoint question in November, but had not issued a full guidance, pending further discussion. All current phase 3 trials have a primary endpoint of improving one (NASH or fibrosis) without worsening the other. Improving both would be very hard to achieve in such a multifaceted disease. “If you look at the compounds where we have clinical data right now, and then try to set both NASH resolution and improvement in fibrosis [as combined endpoint], it’s a really high bar,” says Lene Melchiorsen, Novo Nordisk’s project vice president for NASH.

Regardless of the EMA’s ruling, 2019 should be the year of NASH, culminating decades of effort. “We took a disease that was unrecognized, then became recognized but not thought to be important. People were saying, ‘Ah, it’s a little bit of fat in the liver,’” says Rotman. “Now I think we’re moving forward very nicely.” But drug approval will create its own set of problems. Once a drug is available, Konerman expects a flood of new patients, which will strain the health system’s capacity. “How can we see all of these patients, and … how are we going to get these people their medications?” she asks. “Because they’re probably going to be ridiculously expensive.” And Rotman cautions that drugs won’t replace diet and activity as the foundation of treatment. “In real life, it’s easier for the doctor to give a pill and for the patient to take a pill,” he says. “But the backbone will have to be lifestyle.”

Box 1 | After Ocaliva

Intercept’s Ocaliva is the early favorite to be the first NASH drug to market. Meanwhile, other companies are seeking to better Ocaliva with their own FXR agonists. Gilead’s GS-9674, a small molecule that Gilead acquired from Phenex Pharmaceuticals AG in 2015, is a nonsteroidal FXR agonist. It acts mainly in the intestine, according to Gilead’s Rob Myers, as opposed to Ocaliva’s more systemic effects. In data presented at the American Association for the Study of Liver Diseases annual meeting in November13, GS-9674 treatment did not elevate LDL or cause insulin resistance compared with placebo. GS-9674 treatment, compared with Ocaliva, leads to more transient FGF19 elevation, and “this … more physiologic elevation accounts for an absence of a lipid [LDL] signal for this compound,” says Gilead’s Myers. But 14% of patients at the high dose did experience itching, one of Ocaliva’s side effects. As for cancer risk from activating FXR, says Myers, “a strategy of having sustained levels of a compound may be dangerous, particularly in patients who have NASH and advanced fibrosis [with] an underlying risk of liver cancer.” GS-9674’s more transient effects might be safer.

But is it effective enough? Future development of GS-9674 awaits the outcome of another phase 2 trial, which tests three Gilead compounds singly and in combination, using paired liver biopsies to evaluate efficacy. (The first trial used imaging and blood-based measurements, which are useful for fibrosis and for lipid measurements, but can’t reveal NASH resolution.) Enrollment was complete in November. “The histologic benefits we see with the different regimens will really guide us as to which regimen to take forward into phase 3,” says Myers.

Novartis’s non-bile-acid FXR agonist tropifexor also in phase 2, is further behind. Novartis has reported that the drug is well tolerated in healthy volunteers and is positioning it as the backbone for combination therapies in NASH. Last year, Novartis and Allergan jointly launched a phase 2 trial of tropifexor and Allergan’s CCR2/CCR5 inhibitor cenicriviroc, and in November Novartis partnered with Pfizer to test tropifexor in combination with three different experimental Pfizer NASH compounds.

Several other FXR agonists are in clinical development for NASH. NGM Biopharmaceuticals is taking a downstream approach, targeting FGF19 signaling instead of FXR. In designing its first-in-class drug, the company set out to eliminate the theoretical cancer risk of activating FGFR4—a crucial precaution because ectopic FGF19 expression in transgenic mice causes liver cancer. Its NGM282 is an engineered FGF19 variant that acts as a ‘biased ligand’ to selectively activate certain pathways downstream of FGFR4; for example, to suppress CYP7A1, but not tumorigenic signaling14

The company recently reported interim phase 2 trial results in NASH15. The drug reduced hepatic fat—biopsies to determine NASH resolution weren’t done—and fibrosis scores improved. The drug was well tolerated, with no itching reported. But NGM282, like Intercept’s Ocaliva, did raise LDL.

NGM was preparing an initial public offering as of mid-December and declined an interview request, citing the preoffering “quiet period.” The company seeks to raise $75 million. “We expect that these funds will allow us to complete our ongoing phase 2 and planned phase 2b clinical trials and begin preparation for phase 3 clinical trials of NGM282 for NASH,” the company wrote in its prospectus. As for the cholesterol problem, the filing stated, “We have demonstrated the ability of concomitant statin use to mitigate the serum LDL cholesterol elevations driven by NGM282 activity.”

Box 2 | Genetics reveal new targets

NASH is a genetically complex disease, but some known genes are major contributors to disease risk. In 2008, two genome-wide association studies independently reported an association between a variant of the gene PNPLA3 (palatin-like phospholipase domain-containing 3) and liver fat content16 and liver enzyme levels17, respectively. Allele frequency is ~49% for Hispanics, 23% for European Americans and 17% for African Americans. PNPLA3 encodes a lipase, a fat-digesting enzyme, but how such an enzyme (or its absence) causes steatosis is not obvious. (PNPLA3 knockout mice do not accumulate fat in the liver.) The enzyme builds up in lipid droplets, which store triglycerides for use at times of nutrient scarcity, suggesting dysfunction there. Mutant PNPLA3 has also been implicated in hepatitis C progression and in alcoholic liver disease. “It may have a role in some conserved injury pathway in the liver,” says Rotman. “That by itself makes it an attractive target.” At the November liver meeting, AstraZeneca and Ionis Pharmaceuticals reported an antisense oligonucleotide targeting mutant PNPLA318. The compound reduced steatosis, inflammation and fibrosis in PNPLA3 mutant mice. But PNPLA3’s function remains obscure, and it’s not yet clear where its expression most affects the disease, whether in hepatocytes, stellate cells or fat cells, says Rotman. And the variant confers only a 3–4 times additional risk of liver steatosis—an important gene, clearly, but not yet a validated target. A second gene, TM6SF2, has also been implicated in fatty liver disease19.?

A more consequential genetic association emerged last year. Two groups independently discovered a common genetic variant that strongly protects against progression of steatosis to NASH20,21. That directly implicates the gene product, the enzyme HSD17B13, in NASH pathogenesis. “It’s a strong case, I would call it a confirmed hit,” says Rotman, an investigator in one of the groups, led by the NIH. The other was a collaboration between Regeneron and Geisinger Health System in Pennsylvania, which provided DNA and electronic health records from almost 50,000 patients for exome sequencing and analysis by Regeneron. Exome sequencing revealed an association between a HSD17B13 variant and lower liver enzymes, and Regeneron went on to show that the homozygous variant provides 30% protection against nonalcoholic liver disease, 49% protection against cirrhosis, 53% protection against alcoholic liver disease and 73% protection against alcoholic cirrhosis. (Heterozygotes are protected to a lesser degree.) About a quarter of the population carries at least one copy of the protective variant. A recent study also showed protection against hepatitis C fibrosis22. Unlike PNPLA3 and TM6SF2, which mainly affect liver fat accumulation, “the effects of this gene seem to be really on the inflammation-fibrosis side,” says Aris Baras, head of the Regeneron Genetics Center. “Which is really exciting, if we can create therapeutics.”

Regeneron is now collaborating with Alnylam to advance a small interfering RNA (siRNA) therapeutic targeting HSD17B13 in NASH. Regeneron is also doing its own drug discovery in-house. “We’re pursuing all modalities that make sense,” says Baras. Meanwhile, several other companies are rumored to be working on drugs targeting HSD17B13.