Article | Open | Published:

Effects of NGM282, an FGF19 variant, on colonic transit and bowel function in functional constipation: a randomized phase 2 trial

The American Journal of Gastroenterologyvolume 113pages725734 (2018) | Download Citation

Subjects

Abstract

Objective

NGM282 is an analog of fibroblast growth factor 19 (FGF19), a potent inhibitor of bile acid (BA) synthesis in animals and humans. In phase 2 trials in type 2 diabetes and primary biliary cholangitis, NGM282 was associated with dose-related abdominal cramping and diarrhea. We aimed to examine effects of NGM282 on colonic transit, stool frequency and consistency, hepatic BA synthesis (fasting serum C4), fecal fat, and BA in functional constipation (FC).

Methods

Two-dose NGM282 (1 and 6 mg, subcutaneously daily), parallel-group, randomized, placebo-controlled, 14-day study in patients with FC (Rome III criteria) and baseline colonic transit 24 h geometric center (GC) <3.0. We explored treatment interaction with SNPs in genes KLB, FGFR4, and TGR5 (GPBAR1). Statistical analysis: overall ANCOVA at α = 0.025 (baseline as covariate where available), with three pairwise comparisons among the three groups (α = 0.008).

Results

Overall, NGM282 altered bowel function (number of bowel movements, looser stool form, and increased ease of passage) and significantly accelerated gastric and colonic transit. Dose-related effects were seen with GC 24 h, but not with gastric emptying (GE) and GC 48 h. There were no differences in fecal fat or weight, but there was reduced fecal total BA excretion with NGM282. The most common adverse events were increased appetite (n = 0 with placebo, 2 with 1 mg, 9 with 6 mg), injection site reaction (n = 2 placebo, 4 with 1 mg, 8 with 6 mg), and diarrhea (n = 1 with 1 mg and 4 with 6 mg NGM282). There was treatment interaction with KLB SNP, with greater increase in colonic transit in participants with the minor A allele (p = 0.056).

Conclusion

NGM282 significantly impacts GE and colonic transit, consistent with the observed clinical symptoms. The specific mechanism of prokinetic activity requires further research.

Introduction

The protein KLB and receptor functional growth factor receptor 4 (FGFR4) are the targets on the hepatocyte of the hormone FGF19, which is produced by the ileal enterocytes in response to absorption of bile acids. These are pivotal in the enterohepatic circulation of bile acids. NGM282 is a recombinant protein with 95.4% of its 190 amino acids identical to fibroblast growth factor 19 (FGF19) and with similar binding affinity to FGFR4 and KLB co-receptor on human hepatocytes. The molecule NGM282 mimics the actions of FGF19 on bile acid synthesis by decreasing the gene expression in human hepatocytes of CYP7A1, which is responsible for catalyzing the rate-limiting step of cholesterol conversion to bile acids. This has been demonstrated by a rapid and dose-dependent reduction in levels of serum 7-alpha-hydroxy-4-cholesten-3-one (C4), a key intermediate metabolite of CYP7A1-mediated bile acid synthesis in mice and humans [1,2,3]. However, phase II trials in type 2 diabetes, primary biliary cholangitis, and non-alcoholic steatohepatitis have also shown an unexpected increase in abdominal cramping, diarrhea, or loose stools that were dose-dependent and present in all study populations [4, 5].

These gastrointestinal symptoms are inconsistent with the current understanding of the activity of human FGF19. Up to 30% of patients with diarrhea-predominant irritable bowel syndrome (IBS-D) have evidence of reduced serum FGF19 [6], and this is associated with either increased bile acid synthesis or increased fecal excretion of bile acids [7]. By inhibiting hepatic synthesis of bile acids, NGM282 should reduce fecal excretion of bile acids and, thus, reduce bowel function and other lower gastrointestinal symptoms such as diarrhea. Therefore, the occurrence of increased diarrhea or loose stools does not appear to be consistent with FGF19-mediated inhibition of CYP7A1. Recently, we have shown that there is a positive correlation between serum C4 and fecal bile acids and colonic transit, even among patients with IBS-D or functional diarrhea without overt bile acid malabsorption [8]. Prior literature documented association of the variants of three genes involved in the effects of FGF19: KLB and FGFR4 in effects of bile acids, and TGR5 in colonic transit in humans [9,10,11].

Our aim, therefore, was to examine the effects of NGM282 on gastrointestinal functions in order to understand the mechanism of increased stool frequency and looser consistency in patients with functional constipation. Specifically, we evaluated effects on gastric and colonic transit, stool frequency and consistency, hepatic bile acid synthesis (C4), fecal fat, bile acids, and clinical symptoms in a placebo-controlled, two-dose NGM282 trial. We also conducted an exploratory analysis to assess possible interactions of NGM282 treatment interactions with single-nucleotide polymorphisms of KLB, FGFR4, and TGR5.

Materials and methods

Trial design

This was a single-center, 14-day treatment, randomized, double-blind, placebo-controlled, parallel-group, outpatient study of two doses of NGM282 in patients with functional constipation. The study took place in the Clinical Research Unit at Mayo Clinic, Rochester, MN. The study was reviewed and approved by Mayo Clinical Institutional Review Board and was registered in ClinicalTrials.gov, NCT#02649062.

Computer-based randomization generated by the office of the biostatistician was performed in fixed block sizes with equal allocation to the two doses of NGM282 and placebo according to the schedule provided by Mayo Clinic Department of Health Sciences Research. The research pharmacy of the Clinical Research Unit maintained the randomization schedule. The participants, investigators and statistician were blinded to group assignments. Sequentially numbered containers were maintained in the Mayo Clinic Research Pharmacy.


Participants

Eligibility criteria

The main eligibility criteria for participants were females who were not pregnant or nursing, 18–65 years of age, diagnosed with functional constipation by Rome III criteria and baseline colonic transit geometric center of <3.0 at 24 h, no evidence of dyssynergic defecation on rectal exam, no evidence of gastrointestinal diseases, and not on medications that would affect the gastrointestinal system. Thus, patients were included based on combination of symptoms as well as objective evidence that the participants did not have evidence of accelerated colonic transit. Details on inclusion and exclusion criteria can be reviewed in the Supplemental Materials. Patients were required to be off laxatives (as well as other medications detailed in Supplemental Materials) from 48 h prior to and throughout the study period.

Participants completed 10 visits during the study (Fig. 1) including a screening medical and physical examination, fasting screening laboratory tests including a comprehensive metabolic profile, a 12-lead electrocardiogram, and completion of a bowel disease questionnaire. Eligible participants underwent baseline colonic transit at 24 h, baseline 48-h stool fat, and bile acid measurement after eating a high-fat (100 g) diet for 4 days with stool collection in the final 2 days of the high-fat diet. Participants received a study drug kit and diary for recording bowel functions and date/time of study drug administration. Patients then received treatment (double-blind) for 14 days, with transit measurements conducted on days 8–10, ingestion of 100 g fat diet on days 10–14, and stool collections on days 13 and 14 (Fig. 1). The patients were contacted 7–10 days after finishing treatments in order to check for safety or adverse events.

Fig. 1
Fig. 1

Experimental protocol

Participants were enrolled by the study coordinators. The Research Pharmacy assigned the interventions for participants, based on the randomization provided by the study statistician, and using sequentially numbered containers. All study personnel and participants were blinded to study interventions. The study was approved by Mayo Clinic’s Institutional Review Board, and all participants gave written informed consent.


Study interventions

The study doses of NGM282, 1 and 6 mg, were selected based on observed gastrointestinal effects in phase 2 trials in participants with primary biliary cholangitis and type 2 diabetes mellitus. NGM282 was provided as a sterile solution of identical volume (0.3 ml) in pre-filled syringes delivering doses of 1 and 6 mg and a matched placebo. Both NGM282 and placebo were self-administered as a daily subcutaneous injection by the study participants for 14 days. The doses selected were based on prior reports on effects in liver diseases [4, 5].


Parameters measured

Clinical symptoms, adverse events, and safety labs

All participants underwent history, physical examination, vital signs, electrocardiogram (EKG), and laboratory safety assessments at specified times, including serum chemistries, hematology, and urinalysis. Pregnancy testing was performed at the start of the study and prior to any radioisotope ingestion.

Safety and tolerability were assessed by clinical evaluation of the following parameters: treatment emergent adverse events (TEAEs), clinical laboratory assessments, physical examination, EKGs, and vital signs. TEAEs were defined as AEs that commenced on or after the time of start of first study drug administration. AEs were categorized by intensity, causality, and relationship to study drug.

Stool frequency and consistency

Participants completed a daily stool diary (Supplemental Materials) at baseline and during the period of administration of the study drug to record and assess the timing of each bowel movement, consistency based on Bristol stool form scale [12], ease of passage [13, 14], and sense of completeness of evacuation [13, 14]. The stool diary was dispensed at visit 1 (screening) and at visit 5 (randomization). Ease of stool passage was also assessed based on a 7-point scale (1: manual disimpaction; 2: enema needed; 3: straining needed; 4: normal; 5: urgent w/o pain; 6: urgent w/pain; 7: incontinent).

Gastric and colonic transit

Gastric and colonic transit measurements were performed after an overnight fast using a well-validated scintigraphy method. Participants ingested a 99mTc egg meal (320 kcal, 30% fat; two scrambled eggs with one slice of whole wheat bread and one glass of skim milk) and 111In-labeled charcoal in a methacrylate-coated capsule. Abdominal images were obtained at standard times including scans every 15 min for the first 4 h for estimation of gastric emptying, as well as at 6, 8, 24, 32 and 48 h for detailed estimation of colonic transit summarized as geometric center at 24 and 48 h, and ascending colon emptying T1/2 (by linear extrapolation of counts in the ascending colon).

Hepatic bile acid synthesis

Fasting morning serum C4 was measured by liquid chromatography-tandem mass spectrometry [15] at baseline and during administration of study drug for estimation of hepatic bile acid synthesis.

Fecal bile acids and fat

These measurements were performed after completion of the baseline and on-treatment transit measurements. Fecal bile acids and fat were measured over 48 h. After consuming a 100-g fat diet for 48 h, participants collected all stool samples during the next 48 h while ingesting the same diet. The samples were kept cold or frozen and aliquots of the 48-h stool collection were analyzed using the van de Kamer method for fat and high performance liquid chromatography (HPLC)/tandem mass spectrometry for bile acids [16].

Genotyping

Genomic DNA was isolated from peripheral blood leukocytes as described in previous studies (Qiagen Kit, Qiagen Corp., Germantown, MD, USA). Our lab has previously published assays to detect the variations in the genes of interest (KLB rs17618244, FGFR4 rs351855, and TGR5 rs11554825) by TaqMan® SNP Genotyping Assays (Applied Biosystems; Foster City, CA, USA) according to the manufacturer’s instructions, using 10–20 ng DNA [9, 10]. Following PCR amplification, end reactions were read on the ABI 7500 Fast Real-Time PCR System using Sequence Detection Software version 1.3.1 (Applied Biosystems, Thermo Fisher Scientific, Waltham, MA, USA).


Statistical considerations

End points

The primary end point was colonic transit as measured by geometric center at 48 h. Secondary end points were: colonic geometric center at 24 h, the half-time of ascending colon emptying, daily bowel function including stool frequency and consistency as measured by the Bristol Stool Form Scale (BSFS) and ease of passage from baseline to end of treatment as measured by a study diary (Supplemental Materials), total 48-h fecal bile acid secretion and percentage of main fecal bile acids (chenodeoxycholic acid, cholic acid, deoxycholic acid, lithocholic acid, and ursodeoxycholic acid), fasting serum C4, and total and percentage of serum bile acids.

An exploratory pharmacogenetics association study was performed to assess the interaction of SNPs in genes of KLB, FGFR4, and TGR5 (GBPAR1) with the effects of NGM282 on colonic transit.


Study power and analysis

Table 1 shows the effects sizes detectable with 80% power based on a two-sample t test using a two-sided alpha level of 0.025, with 10 participants per group. These data were obtained for the primary end points in the administrative interim analysis conducted after 15 participants had completed the study; this was performed to assess more accurately the coefficient of variation (COV) in transit responses to the medication in the trial. This showed that the COV was smaller than predicted and, therefore, the total sample size could be reduced while still achieving the anticipated treatment effect size.

Table 1 Statistical power expressed as effect size detectable with 80% power (α = 0.025) based on interim analysis and 10 participants per treatment group

Based on data obtained for the primary end points in the interim analysis, the effect sizes detectable with 80% power based on a two-sample t test using a two-sided alpha level of 0.025 are shown in Table 1. The effect size is the difference in group means relative to the overall mean (assuming 10 per group). An analysis of covariance (ANCOVA) should provide 80% power to detect similar differences using a pooled estimate of variation across all groups and potentially even smaller effect sizes by adjusting for important covariates.

Statistical analysis included overall ANCOVA at α = 0.025 (using baseline as covariate where available), with three pairwise comparisons among the three groups at α = 0.008 to correct for multiple comparisons. All randomized participants were included in the intent-to-treat analysis, and analysis was according to the original assigned groups. All baseline characteristics and demographic data were summarized using the intent-to-treat analysis. The intent-to-treat analysis was based on randomized treatment, if this differed from actual treatment received.

For each missing data point, we imputed the average value for all participants in the study and reduced the degrees of freedom by one for each data value imputed for that end point.

Analysis of genotypes did not include imputed data and used a dominant genetic model. Also included in the analysis of variance models was a genotype by treatment interaction.

All authors had access to the study data and reviewed and approved the final manuscript.

All of the analyses (including power calculation, interim analysis conducted for ensuring adequate power, and final study analyses) were conducted by the academic authors (chiefly authors A.R.Z. and M.C.), completely independently of the study sponsor.


Role of the funding source

This was a single-center study at Mayo Clinic in Rochester, MN. The trial was sponsored by NGM Biopharmaceuticals, South San Francisco, CA. NGM282 was supplied by NGM Biopharmaceuticals. One investigative team at one site in the United States was responsible for recruitment, enrollment, and follow-up of subjects.

Results

Participant disposition and baseline data

Forty-two female patients were assessed for eligibility in the study from 22 December 2015 through 5 December 2016 (Fig. 2). Four patients declined to participate, four did not meet study criteria, and three had colonic transit GC >3.0 at 24 h; thus, 31 participants were randomized to the study. Ten participants received placebo, 10 received 1 mg NGM282, and 11 received 6 mg NGM282. Four participants in the NGM282 6 mg group discontinued treatment, three because of diarrhea and one because of injection site reaction. However, all completed all the transit measurements, and all but one completed post-treatment fecal fat and bile acid measurements. The four participants who were randomized to the 6 mg NGM282 dose and withdrew secondary to diarrhea or injection site reaction were not taking study drug during the final 48-h stool collection; however, all participants completed all transit measurements. One participant in the placebo group did not complete post-treatment fecal fat and bile acid measurements.

Fig. 2
Fig. 2

CONSORT flow chart

Participants in the three groups were similar in all baseline characteristics (Table 2). Baseline colonic transit geometric center was 1.91 ± 1.0, which is similar to the 2.0 ± 1.0 (SEM) that has been reported in the past in patients with functional constipation [17].

Table 2 Demographics and baseline characteristics of the three groups of participants

Effects of NGM282 on gastric emptying and colonic transit

NGM282 significantly accelerated gastric emptying and overall colonic transit at 24 and 48 h compared to placebo (Table 3). Both NGM282 dose groups had acceleration of gastric emptying compared to placebo, but there was no difference between the two NGM282 groups.

Table 3 Effects of NGM282 on gastric emptying, colonic transit, bowel function, stool fat, bile acids, and bile synthesis at the end of treatment

The effect of NGM282 on colonic transit at 24 h was dose-dependent; there was borderline acceleration with NGM282 1 mg and significant acceleration with NGM282 6 mg compared to the NGM282 1 mg and placebo groups. At 48 h, there was significant acceleration with NGM282, 1 and 6 mg, compared to placebo, but there were no differences between the two NGM282 groups (Fig. 3). The colonic transit at 24 h on treatment with NGM282 1 and 6 mg was accelerated relative to baseline as shown in Tables 1 and 2. The overall effect of NGM282 (for both the 1 and the 6 mg doses) on ascending colon emptying compared to placebo was borderline (Table 3).

Fig. 3
Fig. 3

Effect of NGM282 on colonic transit (CT) at 24 and 48 h estimated as geometric center (overall p < 0.001 at 24 and 48 h). Note that the dose-related effect is significant at 24 h (p = 0.007) for NGM282 6 mg vs. 1 mg, and each dose is significant compared to placebo at 48 h

Effects of NGM282 on bowel function and stool fat

Table 3 shows treatment effects on bowel function and fecal fat, adjusting for baseline values using estimated least square means. Participants who received NGM282 6 mg had significantly more bowel movements per week compared to participants who received NGM282 1 mg or placebo, with average bowel movements of 15, 8, and 5 per week, respectively, in the three groups (Fig. 4). The difference in the number of bowel movements per week between NGM282 1 mg and placebo was not statistically significant.

Fig. 4
Fig. 4

Effects of NGM282 on stool frequency, form (consistency using Bristol stool form scale), and ease of passage

Stool form was significantly softer and easier to pass for participants who received NGM282 6 mg compared to placebo, with no difference between the two doses of NGM282.

Stool weight and fecal fat excretion were normal and similar in the three groups (Table 3).

Effects of NGM282 on fecal bile acids and serum C4

Bile acid excretion decreased and the proportion of primary bile acids increased in participants who received NGM282 compared to placebo (Table 3). The decrease in bile acid excretion was significant for NGM282 1 mg and borderline for NGM282 6 mg compared to placebo; there were no significant differences in fecal bile acid excretion between the NGM282 1 and 6 mg dose groups.

Bile acid synthesis, as measured by fasting serum C4, similarly decreased in participants who received NGM282 compared to placebo, although this overall difference was not statistically significant (overall p = 0.056, Table 3). None of the baseline samples or samples obtained on placebo treatment showed serum C4 levels below the limit of detection of the assay (that is <0.9 ng/mL). However, 1 participant who received NGM282 1 mg treatment and 8 patients who received NGM282 had serum C4 levels below 0.9 ng/mL. Table 3 shows that numerically. The suppression of fasting serum C4 was greater with the 6 mg dose than the 1 mg dose, with the mean changes from baseline of −2.9 ng/mL for the NGM282 1 mg group and −8.3 ng/mL for the NGM282 6 mg group.

Exploratory analysis of GPBAR1, FGFR4, KLB genotypes, and effects on colonic transit by treatment group

While both genotype groups showed acceleration of colonic transit with NGM282, we detected a greater acceleration of colonic transit at 24 h in participants with the KLB rs17618244 A (minor) allele (KLB Gln728) compared to KLB major (G) allele (KLB Arg728) in response to NGM282 6 mg, suggesting an interaction between this genotype and the response to colonic transit (Fig. 5).

Fig. 5
Fig. 5

Klothoß genotype by treatment interaction

There were no significant interactions detected between treatment and FGFR4 rs351855 or TGR5 rs11554825 genotypes. There were no significant differences in gastric emptying and bile acids in the different genotype polymorphisms in genes KLB, FGFR4, and TGR5 (GBPAR1).

Tolerability and safety

The majority of adverse events were mild and all resolved upon discontinuation of study medication. There were four early withdrawals in the NGM282 treatment arms (three due to diarrhea and one due to injection site reactions) vs. none in the placebo arm. However, all participants completed planned colonic transit on study medication. Increased bowel movement frequency was reported as an adverse event (in contrast to the information garnered from the assessment of patient response outcomes in the daily diary) in one participant receiving NGM282 1 mg and in four participants receiving NGM282 6 mg, compared to none on placebo (p = 0.0964). There was a higher frequency of increased flatulence reported in participants on NGM282 1 mg (n = 1) and NGM282 6 mg (n = 4) when compared to none in the placebo arm (p = 0.0964). Similarly, four participants who received NGM282 6 mg reported bloating compared to none with NGM282 1 mg and one who received placebo. Headaches were seen at both doses (n = 6 in each arm compared to one participant on placebo (p = 0.0531)). There were more injection site reactions observed in NGM282 treated participants (n = 8 with 6 mg and n = 4 with 1 mg) compared to two in the placebo arm (p = 0.0504).

Hyperphagia was observed in two of the NGM282 1 mg and nine of the NGM282 6 mg treated subjects, respectively, while none on placebo (p < 0.0001).

Discussion

Our study has shown that NGM282 increased the number of bowel movements and resulted in looser stool consistency and greater ease of passage of stool, and this was associated with acceleration of colonic transit. As an FGF19 analog, NGM282 reduced fecal bile acid excretion; however, there was no evidence of steatorrhea, suggesting that steatorrhea secondary to bile acid deficiency could not be the cause of the increased bowel functions and acceleration of colonic transit or gastric emptying. The acceleration of gastric emptying T1/2 and colonic transit at 48 h was significant with both the 1 mg and 6 mg doses of NGM282. In contrast, the effect of the 1 mg dose on colonic transit at 24 h was borderline (p = 0.07). This greater colonic transit effect with the 6 mg dose is consistent with the propensity for NGM282 to induce diarrhea and loose stools with increasing dosage [4, 5]. The medication may provide novel approaches for accelerating gastric emptying and colonic transit as an alternative to the prevailing therapeutic approach with 5-HT4 receptor agonists, some of which may have effects on cardiac rhythm.

Previous trials on bile acids [13], colesevelam [14], and ileal bile acid transporter inhibitor (IBAT) [18] have shown a modest decrease in gastric emptying, with an inverse relationship between gastric emptying and colonic transit, that is, acceleration of colonic transit and delay of gastric emptying. This action is presumed to result from bile acid binding of the G protein-coupled receptor, TGR5 [19], on enteroendocrine cells with resultant secretion of glucagon-like peptide 1 (GLP-1). However, in our current study of NGM282, gastric and colonic transit were both accelerated, suggesting this was not an effect mediated by bile acids, but rather the action of NGM282 on nerve cells with neuromuscular stimulation in the gastrointestinal tract. These effects of NGM282 appear to be related to increased colonic motility rather than increased small bowel or colonic secretion. This hypothesis is supported by the fact that stool weight was similar in the placebo and NGM282 groups, and the relative increase in stool consistency was low compared to the increase in the numbers of stools in participants who received NGM282. This change in colonic transit occurred despite a decrease in bile acid excretion, suggesting a different mechanism of action.

Potential mechanisms

There are several potential mechanisms that could result in the stimulation of gastric and colonic transit with NGM282.

  1. a.

    Repression of ASBT (IBAT): it is conceivable, based on prior studies of FGF19, that there was a repression of the apical sodium-dependent bile acid transporter [20]. This would result in higher concentrations of bile acids in the colon and, hence, increased colonic secretion or motility. However, we noted decreased fecal bile acids, suggesting that this mechanism is unlikely to explain the acceleration of colonic transit.

  2. b.

    Upregulation of BA receptors associated with stimulation of colonic transit: in our current studies, we did not measure changes in BA transporter expression or function and, therefore, we cannot exclude that as a potential mechanism for the acceleration of gastrointestinal transit. It is conceivable that the decrease in bile acid excretion may have caused a compensatory increase in the number or function of BA transporters or the TGR5 receptor which mediates effects of bile acids on colonic motility [21] and may explain at least some of the changes in gastrointestinal transit observed in our study.

  3. c.

    Stimulation of excitatory neural control: an alternative explanation would be stimulation of excitatory neural control of gastric and colonic motility. This is supported by the following observations in the literature:

(i) The FGF family of signaling molecules are regulatory factors in autocrine/endocrine signaling during ENS development [22]. Given recent evidence that enteric neurons in the adult state are maintained by neurogenesis [23], the FGF family of ligands may have an effect on enteric neural function beyond the embryonic stages.

(ii) Effects on central neural mechanisms: studies in mice have shown that Klothoβ deficient mice age prematurely, and this may be linked to decrease in cholinergic signaling through the M1 muscarinic cholinergic receptor [24]. FGF15 (ortholog of human FGF19) has been shown to regulate neural tube formation in mice [25, 26]. Thus, FGF family of proteins, including FGF19 analog, acting more effectively through the increased stability of Klotho/β-Klotho transmembrane proteins could enhance neural functions. Thus, Klothoβ controls neural tube development, and Klotho mutant mice have lower cholinergic expression.

There are functional interactions between muscarinic Ach receptor and FGF receptors in hippocampal neurons, which impact cholinergic neuronal plasticity [27].

It is unclear whether central or peripheral effects on neural control predominate. Given the demonstration that, in the presence of excess FGF19, there is increased permeation across the blood brain barrier, it is conceivable that the effect of FGF19 on stimulation of gastric and colonic transit may be centrally mediated [28]. In fact, there are FGF19 receptors in the hypothalamus, and intracerebroventricular injection of FGF15/19 have been shown to act on FGF receptors 1 and 4 in rat hypothalamus, to alter appetite [29], and this could conceivably explain, at least in part, the hyperphagia reported by some participants in the current study.

Overall, our studies suggest that the action of NGM282 on colonic transit is not related to the effects of the FGF19 analog through the FGFR4 receptor in the liver, but at other sites, potentially cholinergic nerve cells that modulate colonic transit. The potential excitation of neural mechanisms is supported by the exploratory, hypothesis-generating pharmacogenetics observations in our study, which demonstrated a drug-by-genotype interaction between KLB variant and the effect of NGM282 on colonic transit.

Previous work in IBS-D has demonstrated that the protein obtained from the major allele is less stable than that produced by the minor allele, with less feedback inhibition of bile acid synthesis and faster colonic transit in a proportion of IBS-D patients with the major phenotype [9].

Klotho/β-Klotho transmembrane proteins are required for efficient signaling via FGF receptors [30, 31]. The G allele in KLB confers lower KLB protein stability, whereas the A allele is associated with high KLB protein stability [9]. In our study, NGM282 decreased bile acid excretion by decreasing bile acid synthesis, which should result in delayed colonic transit as the FGF19 analog would activate the KLB-FGFR4 receptor complex on hepatocytes to reduce bile acid synthesis in participants with the minor A allele who have stable Klothoβ protein. However, this is the opposite of what we observed, suggesting an entirely different mechanism for the action of NGM282 on colonic transit. We propose that NGM282 in the systemic circulation acts on Klothoβ in nerve cells, that the A allele is associated with normal synthesis or survival of the KLB protein in cholinergic neurons, and that this results in increase in colonic transit.

Further studies are needed to determine the specific cells on which NGM282 acts to increase colonic transit.

Limitations

Our study limitations include the relatively small sample size, inclusion of only females, short duration of treatment when appraising the clinical efficacy of NGM282 as a treatment for functional constipation, and failure to record effects on quality of life, which would be more pertinent in a trial focused on patient response outcomes rather than pharmacodynamics end points. However, the data are generalizable to patients with functional constipation in view of the randomized, controlled trial, and the positive results in the assessment of the effect on colonic transit in a well-powered study prove that the pharmacodynamics effects are robust, Indeed, such results have been accurate predictors of efficacy in phase 2B and 3 clinical trials with other agents, including secretagogues such as lubiprostone and linaclotide, and prokinetic agents such as prucalopride and tegaserod [32]. As summarized elsewhere, scintigraphic colonic transit measurement is a valid biomarker of efficacy in lower functional gastrointestinal disorders [33]. There were a large number of side effects seen more frequently in patients who took NGM282 compared to placebo; however, only a few of these were significant enough to cause discontinuation of the study drug. Further studies in larger patient cohorts are needed to confirm our general results on efficacy and safety of NGM282, and the genetic analytic findings and to determine the specific mechanisms through which NGM282 increases colonic transit.

Conclusion

NGM282 increases gastrointestinal transit through a unique, but unidentified mechanism, likely outside the liver. Further research is needed to fully elucidate the exact mechanism in order to identify optimal mitigation strategies for patients treated with NGM282. Furthermore, these prokinetic activities may have a therapeutic potential in patients with functional constipation or delayed gastric emptying and warrant further study.

Study Highlights

WHAT IS CURRENT KNOWLEDGE

  • Increased bile acid synthesis in response to decreased endogenous FGF19 feedback to hepatocytes is associated with diarrhea.

  • NGM282, an FGF19 analog, decreases bile acid synthesis as demonstrated by a decrease in levels of fasting serum C4, a surrogate of hepatocyte bile synthesis rate.

  • Contrary to expectation in association with decreased bile acid synthesis, patients with cholestatic liver diseases treated with NGM282 reported more frequent and looser bowel movements when compared to placebo.

  • Klothoβ (KLB) controls neural tube development, and Klotho mutant mice have lower cholinergic expression.

WHAT IS NEW HERE

  • NGM282 accelerated gastric and colonic transit without increase in fecal fat and independent of its effect on decreased bile acid synthesis, suggesting a mechanism unrelated to bile acid kinetics or induction of steatorrhea.

  • The acceleration of colonic transit was increased in patients with KLB rs17618244 minor allele, suggesting an effect on peripheral nerves.

  • FGF19 analogs may be developed for use as prokinetics in patients with gastrointestinal dysmotility such as gastroparesis and colonic inertia.

  • The use of FGF19 analogs in cholestatic liver diseases may require the use of antidiarrheals for management of gastrointestinal side effects.

References

  1. 1.

    Luo J, Ko B, Elliott M, et al. A nontumorigenic variant of FGF19 treats cholestatic liver disease. Sci Transl Med. 2014;6:247ra100.

  2. 2.

    Zhou M, Wang X, Phung V, et al. Separating tumorigenicity from bile acid regulatory activity for endocrine hormone FGF19. Cancer Res. 2014;74:3306–16.

  3. 3.

    Zhou M, Learned RM, Rossi SJ, et al. Engineered fibroblast growth factor 19 reduces liver injury and resolves sclerosing cholangitis in Mdr2-deficient mice. Hepatology. 2016;63:914–29.

  4. 4.

    Harrison SA, Abdelmalek MF, Trotter JF, et al. NGM282, a novel variant of FGF19, significantlyreduces hepatic steatosis and key biomarkers of NASH: results of a phase 2, multicenter, randomized, double-blinded, placebo-controlled trial in biopsy-confirmed NASH patients. J Hepatol. 2017;66:S92–3.

  5. 5.

    Mayo MJ, Wigg AJ, Thompson AJV, et al. Impact of NGM282 on the incidence and severity of pruritus in primary biliary cirrhosis patients and correlations with liver chemistries and serum bile acids. Hepatology. 2015;62:56–7.

  6. 6.

    Walters JR, Tasleem AM, Omer OS, et al. A new mechanism for bile acid diarrhea: defective feedback inhibition of bile acid biosynthesis. Clin Gastroenterol Hepatol. 2009;7:1189–94.

  7. 7.

    Valentin N, Camilleri M, Altayar O, et al. Biomarkers for bile acid diarrhea in functional bowel disorder with diarrhea: a systematic review and meta-analysis. Gut. 2016;65:1951–9.

  8. 8.

    Peleman C, Camilleri M, Busciglio I, et al. Colonic transit and bile acid synthesis or excretion in patients with irritable bowel syndrome-diarrhea without bile acid malabsorption. Clin Gastroenterol Hepatol. 2017;15:720–7.

  9. 9.

    Wong BS, Camilleri M, Carlson PJ, et al. A Klothoβ variant mediates protein stability and associates with colon transit in irritable bowel syndrome with diarrhea. Gastroenterology. 2011;140:1934–42.

  10. 10.

    Camilleri M, Shin A, Busciglio I, et al. Genetic variation in GPBAR1 predisposes to quantitative changes in colonic transit and bile acid excretion. Am J Physiol Gastrointest Liver Physiol. 2014;307:G508–16.

  11. 11.

    Camilleri M, Carlson P, Acosta A, et al. Colonic mucosal gene expression and genotype in irritable bowel syndrome patients with normal or elevated fecal bile acid excretion. Am J Physiol Gastrointest Liver Physiol. 2015;309:G10–20.

  12. 12.

    Heaton KW, Radvan J, Cripps H, et al. Defecation frequency and timing, and stool form in the general population: a prospective study. Gut. 1992;33:818–24.

  13. 13.

    Rao AS, Wong BS, Camilleri M, et al. Chenodeoxycholate in females with irritable bowel syndrome-constipation: a pharmacodynamic and pharmacogenetic analysis. Gastroenterology. 2010;139:1549–58.

  14. 14.

    Odunsi-Shiyanbade ST, Camilleri M, McKinzie S, et al. Effects of chenodeoxycholate and a bile acid sequestrant, colesevelam, on intestinal transit and bowel function. Clin Gastroenterol Hepatol. 2010;8:159–65.

  15. 15.

    Camilleri M, Nadeau A, Tremaine WJ, et al. Measurement of serum 7α-hydroxy-4-cholesten-3-one (or 7αC4), a surrogate test for bile acid malabsorption in health, ileal disease and irritable bowel syndrome using liquid chromatography-tandem mass spectrometry. Neurogastroenterol Motil. 2009;21:734–e43.

  16. 16.

    Wong BS, Camilleri M, Carlson P. et al. Increased bile acid biosynthesis is associated with irritable bowel syndrome with diarrhea. Clin Gastroenterol Hepatol. 2012;10:1009–15, e3.

  17. 17.

    Manabe N, Wong BS, Camilleri M, et al. Lower functional gastrointestinal disorders: evidence of abnormal colonic transit in a 287 patient cohort. Neurogastroenterol Motil. 2010;22:293–e82.

  18. 18.

    Wong BS, Camilleri M, McKinzie S, et al. Effects of A3309, an ileal bile acid transporter inhibitor, on colonic transit and symptoms in females with functional constipation. Am J Gastroenterol. 2011;106:2154–64.

  19. 19.

    Hansen M, Scheltema MJ, Sonne DP, et al. Effect of chenodeoxycholic acid and the bile acid sequestrant colesevelam on glucagon-like peptide-1 secretion. Diabetes Obes Metab. 2016;18:571–80.

  20. 20.

    Ghosh A, Chen F, Banerjee S, et al. c-Fos mediates repression of the apical sodium-dependent bile acid transporter by fibroblast growth factor-19 in mice. Am J Physiol. 2014;306:G163–71.

  21. 21.

    Alemi F, Poole DP, Chiu J, Schoonjans K, Cattaruzza F, Grider JR, Bunnett NW, Corvera CU. The receptor TGR5 mediates the prokinetic actions of intestinal bile acids and is required for normal defecation in mice. Gastroenterology. 2013;144:145–54.

  22. 22.

    Memic F, Knoflach V, Morarach K, et al. Transcription and signaling regulators in developing neuronal subtypes of mouse and human enteric nervous system. Gastroenterology. 2017. https://doi.org/10.1053/j.gastro.2017.10.005.

  23. 23.

    Kulkarni S, Micci MA, Leser J, et al. Adult enteric nervous system in health is maintained by a dynamic balance between neuronal apoptosis and neurogenesis. Proc Natl Acad Sci USA. 2017;114:E3709–18.

  24. 24.

    Park SJ, Shin EJ, Min SS, et al. Inactivation of JAK2/STAT3 signaling axis and downregulation of M1 mAChR cause cognitive impairment in klotho mutant mice, a genetic model of aging. Neuropsychopharmacology. 2013;38:1426–37.

  25. 25.

    Yang SL, Yang M, Herrlinger S, et al. MiR-302/367 regulate neural progenitor proliferation, differentiation timing, and survival in neurulation. Dev Biol. 2015;408:140–50.

  26. 26.

    Fischer T, Faus-Kessler T, Welzl G, et al. Fgf15-mediated control of neurogenic and proneural gene expression regulates dorsal midbrain neurogenesis. Dev Biol. 2011;350:496–510.

  27. 27.

    Di Liberto V, Borroto-Escuela DO, Frinchi M, et al. Existence of muscarinic acetylcholine receptor (mAChR) and fibroblast growth factor receptor (FGFR) heteroreceptor complexes and their enhancement of neurite outgrowth in neural hippocampal cultures. Biochim Biophys Acta. 2017;1861:235–45.

  28. 28.

    Hsuchou H, Pan W, Kastin AJ. Fibroblast growth factor 19 entry into brain. Fluids Barriers CNS. 2013;10:32.

  29. 29.

    Ryan KK, Kohli R, Gutierrez-Aguilar R, et al. Fibroblast growth factor-19 action in the brain reduces food intake and body weight and improves glucose tolerance in male rats. Endocrinology. 2013;154:9–15.

  30. 30.

    Itoh N, Ornitz DM. Functional evolutionary history of the mouse Fgf gene family. Dev Dyn. 2008;237:18–27.

  31. 31.

    Jones S. Mini-review: endocrine actions of fibroblast growth factor 19. Mol Pharm. 2008;5:42–8.

  32. 32.

    Camilleri M. Review article: biomarkers and personalized therapy in lower functional gastrointestinal disorders. Aliment Pharmacol Ther. 2015;42:818–28.

  33. 33.

    Zinsmeister AR, Burton D, Camilleri M. Pharmacodynamic and clinical endpoints for functional colonic disorders: statistical considerations. Dig Dis Sci. 2013;58:509–18.

Download references

Acknowledgements

We thank Mrs. Cindy Stanislav for excellent secretarial assistance and the nurses and staff of the Mayo Clinic Clinical Research Unit for nursing support and the care of participants.

Author information

Affiliations

  1. Clinical Enteric Neuroscience Translational and Epidemiological Research (C.E.N.T.E.R.), Division of Gastroenterology and Hepatology, Mayo Clinic, Rochester, MN, USA

    • Ibironke Oduyebo MD
    • , Michael Camilleri MD
    • , Alfred D. Nelson MBBS
    • , Disha Khemani MBBS
    • , Sara Linker Nord
    • , Irene Busciglio BS
    • , Duane Burton MHA
    • , Deborah Rhoten
    • , Michael Ryks
    •  & Paula Carlson BS
  2. Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, MN, USA

    • Leslie Donato PhD
    •  & Alan Lueke
  3. NGM Biopharmaceuticals, South San Francisco, CA, USA

    • Kathline Kim BA
    •  & Stephen J. Rossi PharmD
  4. Department of Health Sciences Research, Mayo Clinic, Rochester, MN, USA

    • Alan R. Zinsmeister PhD

Authors

  1. Search for Ibironke Oduyebo MD in:

  2. Search for Michael Camilleri MD in:

  3. Search for Alfred D. Nelson MBBS in:

  4. Search for Disha Khemani MBBS in:

  5. Search for Sara Linker Nord in:

  6. Search for Irene Busciglio BS in:

  7. Search for Duane Burton MHA in:

  8. Search for Deborah Rhoten in:

  9. Search for Michael Ryks in:

  10. Search for Paula Carlson BS in:

  11. Search for Leslie Donato PhD in:

  12. Search for Alan Lueke in:

  13. Search for Kathline Kim BA in:

  14. Search for Stephen J. Rossi PharmD in:

  15. Search for Alan R. Zinsmeister PhD in:

Guarantor of the article

M.C. accepts full responsibility for the conduct of the study. He has had access to the data and control of the decision to publish.

Specific author contributions

I.O.: fellow investigator, screening, and first-line care of participants, co-authorship; M.C.: PI, project concept and design, senior author of manuscript; A.D.N.: fellow investigator, patient screening, and care, co-authorship; D.K.: fellow investigator, patient screening, and care, co-authorship; S.L.N.: study coordinator, recruitment, and management of participant schedules; I.B.: study coordinator, recruitment, and management of participant schedules; D.B.: gastrointestinal and colonic function analysis and co-authorship; D.R.: technical measurements of gastrointestinal and colonic transit measurement; M.R.: technical measurements of gastrointestinal and colonic transit measurement; P.C.: genotype assays, co-authorship of manuscript; L.D.: laboratory measurement supervision, co-authorship of manuscript; A.L.: lab measurements of serum C4 and fecal bile acids, co-authorship of manuscript; K.K.: sponsor employee and stockholder, reviewed final data and edited the manuscript; S.J.R.: sponsor employee and stockholder, reviewed final data and edited the manuscript; A.R.Z.: study statistician, author of statistical analysis plan, interim and final analyses.

Financial support

This study was sponsored by NGM Biopharmaceuticals, South San Francisco, CA. The study drug, NGM282, was supplied by NGM Biopharmaceuticals. M.C. was supported by R01-DK92179 from National Institutes of Health. The study was conducted in the Mayo Clinic Clinical Research Unit, which is supported by grant UL1-TR002377 from NIH.

Potential competing interests

K.K. and S.J.R. are employees and stockholders of NGM Biopharmaceuticals. The remaining authors declare that they have no conflict of interest.

Corresponding author

Correspondence to Michael Camilleri MD.

Electronic supplementary material

About this article

Publication history

Received

Accepted

Published

DOI

https://doi.org/10.1038/s41395-018-0042-7