Exhibit 99.3

Supplemental Risk Factors

Risks Related to Our Business

Favorable results seen to date in clinical trials of A4250, including our open label Phase 2 trial of A4250 in patients with cholestatic liver disease, may not be predictive of favorable results in our planned Phase 3 clinical trial of A4250 in patients with PFIC, which is expected to be placebo controlled and involve different doses, treatment duration, number of patients and outcome measures and may have other differences in design or execution.

A number of companies in the pharmaceutical and biotechnology industries have suffered significant setbacks in later-stage clinical trials, even after promising results in earlier trials or in preclinical studies. Similarly, companies have experienced disappointing outcomes in later phases of a multiphase clinical trial, even after promising results in an early phase of the trial. A4250 has been evaluated in an investigator-initiated Phase 2 clinical trial for the treatment of PBC and in an open label Phase 2 trial in children with chronic cholestasis. Based on data from the PBC trial that we received from the investigator, eight patients with elevated pruritus received A4250 and all of them reported a reduction in pruritus. In addition, preliminary data from our Phase 2 trial in children with chronic cholestasis showed a reduction in serum bile acids in a substantial majority of patients and improvement in pruritus that was significantly correlated with the reduction in serum bile acids. If the favorable findings on pruritus and serum bile acids seen in these two Phase 2 trials are not replicated in the final data from our chronic cholestasis trial, in our planned trial of A4250 in patients with PFIC or in any other future trial of A4250 in patients with PFIC or other pediatric cholestatic liver disease or disorder, we may not obtain marketing approval for A4250 to treat any indication, in which case our business would be materially and adversely affected.

Our planned Phase 3 trial of A4250 in patients with PFIC is expected to involve a greater number of patients, different outcome measures, doses and treatment duration and may have other differences in trial design, in addition to the difference in patient population, compared with either the PBC trial or the pediatric chronic cholestasis trial. For example, our planned Phase 3 trial of A4250 in patients with PFIC will be a randomized, double blind, placebo controlled, multicenter, clinical trial designed to enroll 60 patients, with a treatment duration of six months. The primary endpoint for FDA evaluation, and a key secondary endpoint for EMA evaluation, will be an assessment of change in pruritus. The primary endpoint for EMA evaluation, and a key secondary endpoint for FDA evaluation, will be serum bile acid responder rate, where a responder is a patient who achieves a predetermined reduction in serum bile acid levels. Although we have assessed the effects of A4250 on serum bile acids in prior clinical trials of A4250, we have not previously utilized this serum bile acid responder rate endpoint. Moreover, we do not have any prior data regarding the effect of a placebo in patients with PFIC on serum bile acid levels, pruritus or any other outcome variable to guide the planning for our planned Phase 3 PFIC trial, which increases the risk that the trial will not be powered adequately to show a statistically significant separation between A4250 and placebo.

Furthermore, the specific patient-reported or caregiver-reported measures used to assess change in pruritus can vary from trial to trial. For example, the pruritus scales that were used in the PBC trial are not the same as the scales used in our trial in children with chronic cholestasis, except that the visual analogue scale of itching, known as VAS-itch, is common to both trials. To assess change in pruritus in our planned Phase 3 trial of A4250 in patients with PFIC, we plan to use patient-reported and caregiver-reported outcome instruments that we are currently developing, which employ or rely on different questions or assessments, require a different outcome to establish a positive response or are otherwise different from the outcome instruments used in our Phase 2 trial in children with chronic cholestasis. The differences in these instruments may reduce the likelihood that data from the Phase 2 trials of A4250 in children with chronic cholestasis or in PBC patients will be predictive of favorable results in the planned Phase 3 PFIC trial.

In addition, the expected doses of A4250 for our planned Phase 3 trial for some PFIC patients may not be precisely the same as the weight-based doses that we evaluated in our Phase 2 trial of A4250 in children with chronic cholestasis. Although we expect that patients in the planned trial of approximately the same weight will receive the same total dose of A4250, any difference in relative weights between particular patients will result in a difference in weight-based doses between those patients. This may reduce the likelihood that weight-based data from the Phase 2 trial will be predictive of favorable results in our planned PFIC trial.

1

We plan to use, for purposes of FDA evaluation, change in pruritus as the primary endpoint in our planned Phase 3 trial in PFIC patients. Because the assessment of pruritus relies on subjective patient or caregiver feedback, it is challenging to evaluate and measure consistently and, for any patient, a self-reported outcome may vary from a caregiver-reported outcome.

The primary endpoint in our planned Phase 3 trial of A4250 in patients with PFIC, for FDA evaluation, and a key secondary endpoint for EMA evaluation, will be an assessment of change in pruritus. Because the assessment of pruritus relies on subjective patient or caregiver feedback, it is challenging to evaluate and measure consistently and, for any patient, a self-reported outcome may vary from a caregiver-reported outcome. The measure of pruritus can be influenced by factors outside of our control and can vary widely from measurement point to measurement point for a particular patient, from patient to patient and from site to site within a clinical trial. Moreover, patients given an inactive comparator, or placebo, in a clinical trial may perceive a change in pruritus that is greater than we anticipated when designing the trial or that is comparable to the change experienced by patients given A4250, which could obscure the effect of A4250 in the trial and reduce the likelihood that the trial will be successful. In addition, the FDA or EMA may determine that any particular outcome instrument that we use in our planned Phase 3 trial to assess pruritus is not adequately reliable or valid for use with PFIC patients, which would delay and potentially prevent our receipt of marketing approval for A4250.

The clinical trial designs, durations, endpoints and outcomes that will ultimately be required to obtain marketing approval of A4250 to treat PFIC patients are uncertain and, in any case, may vary among the FDA, EMA and other regulatory authorities outside of the United States and European Union. Based on feedback that we have received from the FDA and the EMA, we expect both regulatory authorities to place a greater emphasis on the totality of the data from our planned Phase 3 clinical trial, including secondary endpoints, than may generally be expected. As a result, there is risk that, even if the primary endpoint of our planned Phase 3 clinical trial of A4250 for FDA evaluation purposes or for EMA evaluation purposes is met with statistical significance, the applicable regulatory authority may not find the overall results of our planned Phase 3 trial to be sufficient to support marketing approval of A4250 to treat PFIC, a symptom of PFIC such as pruritus or any other indication, and we may never receive marketing approval. Similar risks also apply for A3384, which we are developing as a treatment for BAM.

No product is currently approved for the treatment of either of PFIC or BAM in the United States, the European Union or, to our knowledge, any other jurisdiction, and there is limited clinical experience in PFIC and in BAM. Accordingly, there is not a well-established development path that, with positive outcomes in clinical trials, would be reasonably assured of receiving marketing approval for these indications.

Our planned Phase 3 PFIC program includes a single randomized, double blind, placebo controlled, multicenter clinical trial and an open label long-term extension study. The double blind trial is designed to enroll 60 patients with PFIC (subtypes 1 or 2), ages six months to 18 years, at sites in the United States, Canada, western Europe, the Middle East and Australia. Patients will be assigned to receive either 40 µg/kg/day or 120 µg/kg/day of A4250, or placebo, for six months. Patients taking a stable dose of medication to manage pruritus when entering the trial will be permitted to continue such background medication during the trial, subject to specified exceptions. The trial will have a primary endpoint for U.S. purposes, a different primary endpoint for E.U. purposes, and several secondary endpoints, including progression to surgery, change in growth markers and liver biochemistry variables, and others. The primary endpoint for FDA evaluation, and a key secondary endpoint for EMA evaluation, will be an assessment of change in pruritus using a proprietary tool that we developed. The trial’s primary endpoint for EMA evaluation, and a key secondary endpoint for FDA evaluation, will be serum bile acid responder rate, where a responder is a patient who achieves a predetermined reduction in serum bile acid levels. Patients in the trial will have the opportunity to participate in the open label extension study to assess long-term safety and durability of response.

Based on feedback that we have received from the FDA and the EMA, we expect both regulatory authorities to place a greater emphasis on the totality of the data from our planned Phase 3 clinical trial, including secondary endpoints, than may generally be expected. As a result, there is risk that, even if the primary endpoint of our planned Phase 3 clinical trial of A4250 for FDA evaluation purposes or for EMA evaluation purposes is met with statistical significance, the applicable regulatory authority may not find the overall results of our planned Phase 3 trial to be sufficient to support marketing approval of A4250 to treat PFIC, a symptom of PFIC such as pruritus or any other indication, and we may never receive marketing approval.

2

Furthermore, the FDA has informed us that showing a clinically meaningful effect only on pruritus could support approval for the treatment of pruritus associated with PFIC and has indicated to us that meaningful improvements on one or more additional clinical benefit endpoints and/or corroborative real-world clinical evidence would be required to support approval for the treatment of PFIC itself. Certain secondary endpoints that we expect to use in our planned Phase 3 trial may be considered clinical benefit endpoints, including progression to surgery and change in growth markers. The likelihood that a treatment duration of six months will be long enough for a placebo controlled trial to demonstrate an effect of A4250 on any particular clinical benefit endpoint is uncertain, and, if not, FDA approval of A4250 could be limited to the treatment of pruritus in PFIC patients, if A4250 is approved at all.

To support marketing approval of a drug, the FDA requires a demonstration of efficacy based on an endpoint reflecting clinical benefit. However, under Subpart H regulations, the FDA can grant accelerated approval based on a surrogate endpoint that is reasonably likely to predict clinical benefit. If we elect to seek accelerated approval under Subpart H, the FDA (or EMA under its regulations) may determine that the surrogate endpoint, or that the outcome shown in the planned trial on the surrogate endpoint, does not establish a reasonable likelihood of predicting clinical benefit or otherwise is not sufficient to support approval, even if the surrogate endpoint is met with statistical significance. For example, if we pursue an accelerated approval of A4250 under Subpart H and the results of the change in pruritus endpoint are not of a magnitude that will be clearly meaningful, it is possible that the FDA will not consider the result to be sufficient to support approval under the Subpart H pathway. If this occurs, our business would be materially harmed.

In addition, if we pursue an accelerated approval under Subpart H for A4250, we will be required to conduct a post-approval clinical outcomes trial to confirm its clinical benefit in PFIC. There can be no assurance that any post-approval trial that we conduct will confirm that the surrogate endpoint used for accelerated approval will eventually show an adequate correlation with clinical outcomes. If a clinical outcomes confirmatory trial that we conduct fails to show such adequate correlation, we may not be able to maintain any previously granted marketing approval for A4250 in PFIC that we may obtain.

We have selected serum bile acid responder rate as the primary endpoint for purposes of EMA evaluation and as a key secondary endpoint for FDA evaluation. To support the clinical utility of reduction in serum bile acids in the treatment of patients with PFIC, we are supporting an independent study pooling and analyzing long-term PFIC patient data from a number of leading PFIC academic centers, which we refer to as the PFIC Study Group. Should the data that are being accumulated by the PFIC Study Group come to a different conclusion about the relationship of serum bile acids to beneficial clinical outcomes, the EMA or the FDA may determine the results of the Phase 3 PFIC study not to be adequate to support approval.

Also, it is possible that any marketing authorization we may receive in the future from the EMA for A4250 for the treatment of PFIC could be conditional on post-authorization studies and not be considered a full authorization. Our ability to obtain and maintain conditional marketing authorization in the European Union will be limited to specific circumstances and subject to several conditions and obligations, if obtained at all, including the completion of a clinical outcomes trial to confirm the clinical benefit of A4250 in PFIC. Conditional marketing authorizations based on incomplete clinical data may be granted for a limited number of listed medicinal products for human use, including products designated as orphan medicinal products under European Union law, if (1) the risk-benefit balance of the product is positive, (2) it is likely that the applicant will be in a position to provide the required comprehensive clinical trial data, (3) unmet medical needs will be fulfilled and (4) the benefit to public health of the immediate availability on the market of the medicinal product outweighs the risk inherent in the fact that additional data are still required. Specific obligations, including with respect to the completion of ongoing or new studies, and with respect to the collection of pharmacovigilance data, may be specified in the conditional marketing authorization. Conditional marketing authorizations are valid for one year, and may be renewed annually, if the risk-benefit balance remains positive, and after an assessment of the need for additional or modified conditions.

The FDA and EMA generally require two pivotal clinical trials to support marketing approval of a drug. It is our objective to conduct a single Phase 3 clinical trial in patients with PFIC that, together with data from a long-term, open label extension study, forms the primary support for applications for marketing approval of A4250 in

3

both the United States and European Union for treatment of patients with PFIC. If the FDA or EMA requires us to conduct additional clinical trials beyond the ones that we currently contemplate in order to support marketing approval of A4250 to treat patients with PFIC in the United States or European Union, it would result in a more expensive and potentially longer development program for A4250 than we currently contemplate, which could delay our ability to generate product revenues with A4250, interfere with our ability to enter into any potential licensing or collaboration arrangements with respect to this program, cause our value to decline, and limit our ability to obtain additional financing that may be necessary to complete the planned pivotal program. Even though it is our objective to conduct a single Phase 3 trial of A4250 as the basis, together with safety data from an extension study to evaluate long-term outcomes, an application for marketing approval for A4250 in PFIC, either the FDA or the EMA may require that we meet the primary endpoint or endpoints in the trial at a higher level of statistical significance than would otherwise be required for a trial to be successful, which would reduce the likelihood of a positive trial.

Likewise, if we conduct any future clinical trial designed to support marketing approval of A3384 as a treatment for BAM, the FDA, EMA or any regulatory authority outside of the United States or the European Union may determine that the designs or endpoints of the trial, or that the outcomes shown on any particular endpoints in the trial, are not sufficient to establish a clinically meaningful benefit for A3384 in the treatment of BAM or otherwise to support marketing approval, even if the primary endpoint or endpoints of the trial is met with statistical significance.

The design of our planned Phase 3 clinical trial of A4250 in patients with PFIC may not conform precisely in all respects to the recommendations or preferences expressed by either the FDA or EMA.

Although the feedback on our proposed Phase 3 clinical trial in patients with PFIC that we have received from the FDA and EMA is generally consistent, it is not identical and the design of our planned Phase 3 clinical trial of A4250 in patients with PFIC may not conform precisely in all respects to the recommendations or preferences expressed by either regulatory authority. As a result, there is increased risk that, even if we view the results of our planned trial as favorable, the FDA or EMA may not find the overall results to be sufficient to support marketing approval of A4250 to treat PFIC, a symptom of PFIC such as pruritus or any other indication.

If the commercial opportunity in PFIC is smaller than we anticipate, or if A4250 receives approval to treat only a specific subpopulation of patients with PFIC or only a specific symptom of PFIC such as pruritus, our future revenue from A4250 may be adversely affected and our business may suffer.

If the size of the commercial opportunities in any of our target indications is smaller than we anticipate, we may not be able to achieve profitability and growth. We are developing A4250 initially as a treatment for patients with PFIC and potentially also as a treatment for patients with other pediatric cholestatic liver diseases and disorders. PFIC and these other diseases and disorders are each rare, with a limited patient population. Moreover, we expect that the addressable PFIC patient population for A4250 is only a subset of the overall patient population, specifically patients who have not yet received partial external biliary diversion, or PEBD, surgery or liver transplant surgery or patients who have had PEBD reversal surgery. We are not aware of any available patient registries for PFIC, and we rely on various estimates and assumptions to estimate the addressable PFIC population. In addition, there are different subtypes of PFIC and the beneficial effects of A4250 may vary among patients with different subtypes or among children of different ages. We expect that the inclusion criteria for our planned Phase 3 clinical trial will include patients with PFIC subtypes 1 and 2, but not PFIC subtype 3. A4250 may ultimately receive regulatory approval, if at all, as a treatment for some but not all of the subtypes, or for children of some ages but not others. Moreover, we plan to use change in pruritus as the primary endpoint of our planned Phase 3 trial for purposes of FDA evaluation. The FDA has informed us that showing a clinically meaningful effect only on pruritus could support approval for the treatment of pruritus associated with PFIC and has indicated to us that meaningful improvements on one or more additional clinical benefit endpoints and/or corroborative real-world clinical evidence would be required to support approval for the treatment of PFIC itself. If the commercial opportunity in PFIC is smaller than we anticipate, whether because our estimates of the addressable patient population prove to be incorrect, because A4250 receives marketing approval, if at all, as a treatment for some but not all PFIC subtypes, for children of some ages but not others, for pruritus associated with PFIC but not PFIC itself or for any other reason, our future revenue from A4250 may be adversely affected and our business may suffer.

It is critical to our ability to grow and become profitable that we successfully identify patients with PFIC and any other rare cholestatic liver diseases and disorders that we may target in the future. Our projections of the number

4

of people who have PFIC or our other potential target cholestatic liver diseases and disorders, as well as the subset who have the potential to benefit from treatment with A4250, are based on a variety of sources, including third-party estimates and analyses in the scientific literature, and may prove to be incorrect. Further, new information may emerge that changes our estimate of the prevalence of these diseases or the number of patient candidates for A4250. The effort to identify patients with PFIC or our other potential target indications is at an early stage, and we cannot accurately predict the number of patients for whom treatment might be possible. Additionally, the addressable patient population for A4250 may be limited or may not be amenable to treatment with A4250, and new patients may become increasingly difficult to identify or gain access to, which would adversely affect our results of operations and our business. Further, even if we obtain significant market share for A4250, we may never achieve profitability because the potential target patient population for A4250 is small.

The terms of our loan facility agreement and the associated security agreements may restrict our ability to engage in certain transactions and adversely affect our operating flexibility.

In December 2014, we entered into a loan facility agreement with Kreos Capital IV (UK) Limited, or Kreos, which was revised in February 2016 and further amended and restated pursuant to a supplemental deed entered into in May 2016. We have also entered into certain guaranties and security agreements with Kreos. Pursuant to the terms of the loan facility agreement and the associated security agreements, subject to certain exceptions, we cannot engage in specified transactions unless certain conditions are met or we receive the prior approval of Kreos. These transactions include:

If Kreos does not provide its consent to such actions, we could be prohibited from engaging in transactions that could be beneficial to our business and our stockholders unless we repay the loan, which may not be desirable or possible. The obligations under the loan facility agreement are secured by substantially all of our assets. If we default under the loan facility agreement or the associated security agreements, including for an inability to repay amounts as they become due, and we are unable to obtain a waiver for such a default, Kreos would have a right to accelerate our obligation to repay the entire loan, obtain control of our cash accounts and foreclose on the secured assets in order to satisfy our obligations under the loan facility agreement. Any such action on the part of Kreos against us could have a materially adverse impact on our business, financial condition and results of operations.

5