|
Items in cart [0] |
Questions? Call 888-624-8373 |
| Copyright © 2009. National Academy of Sciences. All rights reserved. Terms of Use and Privacy Statement |
Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 42
5
Biomarkers of Acute Idiosyncratic
Hepatocellular Injury in Clinical Trials1
Hepatotoxicity is the adverse event that most frequently leads to regula-
tory action on drugs, including failure to approve, postmarketing warnings
added to the label, and withdrawal from the market (Temple, 2001). Among
research priorities in adverse drug events, hepatotoxicity was ranked first
in a 2006 survey of pharmaceutical companies (Holden, 2008). Because
most events are inaccurately classified (Aithal et al., 1999), the population
incidence of drug-induced liver injury is unknown. Yet drugs are the most
frequent cause of acute liver failure among those under consideration for
liver transplantation in the United States (Lee, 2003).
Animal studies in rodents, dogs, and monkeys detect approximately
half of compounds exhibiting hepatotoxicity in humans (Olson et al.,
2000). In vitro human hepatocyte testing similarly detects 50–60 percent
of drugs that can cause severe liver injury in humans, including some not
detected in animal testing (Xu et al., 2008). However, no currently available
preclinical tests detect the potential for serious human hepatotoxicity with
both high sensitivity and high specificity.
A recent example reveals some of the issues involved in drug safety test-
ing. A major pharmaceutical company submitted a new drug application
for treatment of a chronic disease. The FDA agreed with the sponsor’s effi-
1 This chapter is derived from a white paper prepared by Paul Watkins, Director, Hamner
Center for Drug Safety Sciences, University of North Carolina, Chapel Hill; John Bloom, Dis-
tinguished Medical Fellow, Diagnostic and Experimental Medicine, Eli Lilly and Company;
and Christine Hunt, Vice President, Clinical Safety Systems, GlaxoSmithKline, with additional
input from workshop discussions.
OCR for page 43
BIOMARKERS OF ACUTE IDIOSYNCRATIC HEPATOCELLULAR INJURY
cacy data. However, it was noted that among approximately 4,000 treated
patients in clinical trials, two developed elevations in both serum alanine
aminotransferase (ALT) and bilirubin. As a prerequisite for approval, the
company was told to conduct a new safety study of 10,000 patients treated
with the drug for 1 year, and to include an additional 10,000 subjects
receiving comparator treatment for 1 year to exclude an unacceptable
level of risk for clinically serious acute idiosyncratic hepatocellular injury
(AIHI).2 The delay and additional investment required to bring such drugs
to market can be detrimental not only to their manufacturers, but also to
patients with unmet medical needs.
This chapter begins with an overview of AIHI. It then describes the
current state of biomarkers for AIHI and reviews potential new biomarkers
now emerging from various lines of investigation. The chapter ends with
highlights of the breakout session on liver safety biomarkers.
ACuTE IDIOSyNCRATIC HEPATOCELLuLAR INJuRy (AIHI)
The clinical and histologic presentation of drug-induced liver injury can
take many forms, mimicking most types of liver disease. AIHI is of greatest
concern in drug development because of its potential rapidity of develop-
ment and high morbidity and mortality (Andrade et al., 2005; Bjornsson
and Olsson, 2005). Table 5-1 lists marketed drugs that have been subject
to regulatory actions since 1995 because of liver safety concerns. All of
the drugs listed can cause AIHI, with the exception of terbenafine (mixed
hepatocellular/cholestatic injury), valproate (microvesicular steatosis), and
acetaminophen (hepatocellular injury, but without the characteristics of
AIHI discussed below). The discussion at the workshop focused exclusively
on AIHI and not on other forms of drug-induced liver injury.
Figure 5-1 shows a typical presentation of AIHI. The patient exhibited
normal liver chemistries at baseline and for several weeks while receiving
treatment, but then developed serious liver injury with loss of overall liver
function, manifested as a rise in serum bilirubin and ultimately death.
During AIHI, if treatment is not withdrawn promptly, and in some
cases even with prompt discontinuation, the progressive loss of hepatocytes
leads to liver dysfunction and ultimately death (absent liver transplant). The
event is frequently termed “idiosyncratic” because the majority of treated
patients are able to take the drug safely at the recommended dose range; the
affected individuals are different from the majority in ways that make them
susceptible to injury or less able to recover from injury. With most of the
drugs listed in Table 5-1, fatal AIHI typically occurs in 1 in every 10,000
2 This chapter uses the term “AIHI” to refer specifically to acute and idiosyncratic hepatocel-
lular injury that can progress to liver failure.
OCR for page 44
DEVELOPMENT OF BIOMARKERS FOR DRUG SAFETY
TAbLE 5-1 Regulatory Actions on Approved Drugs Due to
Hepatotoxicity, 1995–2008
Withdrawals Second Line Warnings
• bromfenac • felbamate • acetaminophen
• troglitazone • tolcapone • atomoxetine
• pemoline • trovafloxacin • leflunomide
• bosentan
• nefazodone
• infliximab
• nevirapine
• saquinavir
• pyrazinamide/rifampin
• interferon 1β, 1β
• terbinafine
• telithromycin
• valproic acid (kava, lipokinex)
• zifirlukast
SOURCE: Guo et al., 2008.
100.0
start stop died
Liver Test Values, xULRR
10.0
ALTx
TBLx
ASTx
ALPx
1.0
probably
drug-induced
hosp
0.1
–30
0
30
60
90
120
150
180
Study Day
FIGuRE 5-1 Acute idiosyncratic hepatocellular injury. An 80-year-old man who
experienced acute idiosyncratic hepatocellular injury exhibited marked increases
in alanine aminotransferase (ALT) and aspartate aminotransferase (AST) starting
5-1
about 45 days after treatment began. Total bilirubin (TLB) rose dramatically before
death, while alkaline phosphatase (ALP) increased less markedly. Measures are
compared with the upper limit of the reference range (ULRR).
SOURCE: Watkins slide presentation.
OCR for page 45
BIOMARKERS OF ACUTE IDIOSYNCRATIC HEPATOCELLULAR INJURY
to 100,000 treated patients. It is rare for fatal AIHI to occur in preapproval
clinical trials, in part because most such trials involve insufficient numbers
of patients treated for long enough to have a high likelihood of identifying
such rare individuals.
CuRRENT STATE OF bIOMARkERS FOR AIHI
The primary biomarkers discussed at the workshop were those that
detect a drug’s potential to cause AIHI in a preapproval clinical trial.
Biomarkers to identify individual susceptibility to a drug with established
AIHI potential were discussed to the extent that they were relevant in this
context. Other types of biomarkers, such as those that may aid in causality
assessment, were not discussed.
Serum ALT activity is the biomarker used most frequently to detect
hepatocellular injury in clinical trials and is more liver-specific than aspartate
aminotransferase (AST) (Green and Flamm, 2002). Serum ALT can increase,
even markedly (for example, to levels exceeding 20 times the upper limits
of normal [ULN]), as a result of events other than hepatocyte necrosis,
including hepatocyte autophagy in anorexia nervosa (Rautou et al., 2008) or
hepatic glycogen accumulation in uncontrolled type 1 diabetes (Olsson et al.,
1989; Sayuk et al., 2007). Lesser ALT elevations are observed with hepatic
steatosis (Browning et al., 2004). Activation of ALT gene transcription can
occur in response to Peroxisome proliferators activated receptors (PPARs)
agonists in cell culture (Thulin et al., 2008). In addition, it is theoretically
possible that drugs could interfere with ALT degradation or blood clearance.
However, no published data support transcriptional or clearance-related
ALT increases due to these events occurring in humans.
Drugs recognized to cause AIHI have demonstrated an increased inci-
dence of ALT elevations of more than three times ULN relative to placebo
or controls in preapproval clinical trials (Temple, 2001). However, ALT
elevations have a limited specificity to predict AIHI. Even frequent and
fairly high ALT elevations do not reliably predict the potential to cause
AIHI in clinical trials, as evidenced by tacrine. In clinical trials, about
25 percent of Alzheimer’s disease patients receiving tacrine developed ALT
elevations of greater than three times ULN, and 2 percent exhibited ALT
elevations of greater than 20 times ULN (Watkins et al., 1994). However,
tacrine exhibits a very low risk of causing AIHI. Similarly, although statins
have demonstrated up to a 3 percent incidence of ALT of greater than three
times ULN in clinical trials relative to placebo or controls, statin use has
not been associated with an increased risk of acute liver failure (Kaplowitz,
2005). Heparins also can cause ALT elevations but pose a very low or zero
risk of causing AIHI. Drugs such as tacrine, heparins, and statins gener-
ally exhibit transient, self-limited liver injury that resolves with continued
OCR for page 46
DEVELOPMENT OF BIOMARKERS FOR DRUG SAFETY
treatment in a process termed “adaptation.” Adaptation is observed not
only with drugs that pose a low risk of causing AIHI but also in many or
most patients who experience ALT elevations while receiving drugs that can
cause AIHI, such as troglitazone (Watkins, 1998) and isoniazid (Mitchell
et al., 1975).
Common pathways likely underlie initial injury, regardless of whether
the injury progresses or resolves with continued treatment. These path-
ways include those that determine the intracellular “dose” of hepatotoxic
metabolites or bile acids (for example, cellular transporters, Phase 1 and 2
drug metabolism, and concomitant medications) and the production of
hepatocyte injury (for example, oxidative stress or mitochondrial impair-
ment), as well as regenerative or hepatoprotective abilities (including
hepatic glutathione redox status, Nrf2 activation of cell defense systems,
and liver regeneration). All of these pathways may be influenced by genetic
(Larrey, 2002; Daly et al., 2007; Kindmark et al., 2007; Wilke et al., 2007),
epigenetic (Murata et al., 2007), demographic (Kaplowitz, 2005; Uetrecht,
2007), infectious/inflammatory (Roth et al., 2003), and environmental fac-
tors (Larrey, 2002; Kaplowitz, 2005).
A popular theory is that progressive liver injury occurs in those indi-
viduals who fail to adapt to the initial insult. However, data to support this
theory are sparse. It has been claimed that with drugs capable of causing
AIHI, such as isoniazid, troglitazone, and ximelagatran, the AIHI events typi-
cally occur with a latency similar to that of the more frequent ALT elevations
observed in clinical trials (personal communication, D. Larrey, J. Uetrecht,
P. Watkins), a view consistent with a mechanistic link between isolated ALT
elevations and AIHI events. The temporal relationship between ALT eleva-
tions observed in clinical trials and postmarketing AIHI events has not been
systematically examined.
A logical conclusion would be that drugs that cause ALT elevations can
be placed along a spectrum. On one end are drugs causing ALT elevations
that may reflect liver injury but never cause AIHI. On the other end are
drugs that cause liver injury that progresses to AIHI with relatively high
frequency (perhaps 1 in 10 subjects with ALT elevations who are continued
on treatment). While ALT is a sensitive biomarker for liver injury, it alone
cannot differentiate between drugs at opposite ends of this spectrum.
Drug-Induced Liver Injury with Jaundice:
The Current Gold Standard biomarker for AIHI Potential
Hy Zimmerman (1968) first noted that a patient who presented with
drug-induced hepatocellular injury with jaundice had at least a 10 percent
chance of dying from liver failure (before liver transplantation was avail-
able). In hepatocellular injury, a rise in total and direct bilirubin reflects a
OCR for page 47
BIOMARKERS OF ACUTE IDIOSYNCRATIC HEPATOCELLULAR INJURY
substantial risk because it indicates a major loss of functioning hepatocytes
(when other causes for increased bilirubin are excluded). The approxi-
mately 10 percent mortality or transplant rate for drug-induced hepato-
cellular jaundice—known as “Hy’s Law”—has been confirmed in recent
reports from Sweden and Spain (Andrade et al., 2005; Bjornsson and
Olsson, 2005).
FDA draft guidance for evaluating liver safety in a clinical trial defines
“Hy’s Law cases” as subjects in a clinical trial who experience ALT eleva-
tions of more than three times ULN and total bilirubin of more than two
times ULN and satisfy the following three criteria: (1) the liver injury
should be hepatocellular in nature, and there should not be a prominent
cholestatic component (e.g., serum alkaline phosphatase of more than two
times ULN); (2) there should be no more likely alternative cause than drug-
induced liver injury, such as acute viral hepatitis A or B or C, preexisting or
other acute liver disease, or another drug capable of causing the observed
injury; and (3) there should be evidence that the drug causes more frequent
but less severe hepatocellular injury as shown by more frequent ALT eleva-
tions of greater than three times ULN in the treated group relative to the
control group (FDA, 2007). The FDA has placed great confidence in the
specificity of Hy’s Law cases as a biomarker for identifying drugs capable
of inducing serious AIHI, reporting, “We are not aware of false positive
Hy’s Law findings” (FDA, 2007).
Hy’s Law cases are, however, a specific but imperfect biomarker for
drugs capable of causing AIHI. In an FDA review of 26 new drug appli-
cations (13 of the drugs were known to be “hepatotoxic”), Hy’s Law
events were seldom observed in the clinical trials, even with compounds
later demonstrated to be capable of causing severe AIHI (Pauls, 2004).
This is not surprising since the rarity of susceptible individuals and the
delayed appearance of the event generally would require very large and
prolonged clinical trials for detection. Also, clinical protocols usually
mandate frequent monitoring of and strict stopping rules based on serum
ALT, especially once a liver safety issue has been established for a drug in
development. Stopping treatment at a low level of liver injury may allow
a patient susceptible to AIHI to recover without demonstrating a rise in
serum bilirubin. The only way to determine whether a patient with ALT
elevation will adapt or progress is to continue treatment and observe the
patient closely, with frequent monitoring of liver chemistries. The draft
FDA guidance on liver safety (FDA, 2007) suggests that continued treat-
ment may be considered in subjects with asymptomatic ALT elevations
exceeding three times ULN. This practice may place these study subjects
at greater health risk than subjects without ALT elevations, which raises
ethical concerns.
OCR for page 48
DEVELOPMENT OF BIOMARKERS FOR DRUG SAFETY
Do better biomarkers Than Hy’s Law Cases Exist?
The ideal biomarker would not require placing patients at significant
risk in the course of distinguishing between drugs capable of causing AIHI
and drugs, such as heparin and statins, that do not appear to cause AIHI.
The ideal biomarker also would be able to make this distinction in a rela-
tively small clinical trial of short duration. The plausibility of such a bio-
marker rests on the mechanistic differences between drugs that have the
potential to cause AIHI and those that cause only reversible ALT elevations.
At least two possibilities exist:
• The mechanisms that distinguish AIHI are many, complex, and
agent-specific, to the extent that identifying a manageable number
of predictive markers applicable to most drugs capable of causing
severe AIHI is impractical or impossible.
• Common mechanisms for AIHI exist and can be translated to a
manageable number of validated biomarkers that could be applied
to better understand the hepatic safety of candidate drugs in both
the preregistration and postmarketing settings.
The first of these possibilities suggests the need for drug-specific bio-
markers for those agents whose risk/benefit balance warrants continued
marketing (as when agent-specific markers for patients at risk are available)
when the incidence of liver safety events is high enough to characterize pre-
marketing development. The existence of a small group of markers that, in
the aggregate, have predictive value for AIHI in many or most cases relies
on the second possibility being correct. What follows are three lines of
thought regarding the pathogenesis of AIHI that can be used to examine
possible biomarkers.
Cumulative Injury Theory
Cumulative injury theory maintains that drugs capable of causing AIHI
induce progressive impairment of critical functions of hepatocytes that may
start soon after the initiation of treatment but is not detected by elevation
in serum ALT. An example is progressive mitochondrial injury (for example,
as demonstrated for fialuridine and in cell culture for other drugs, such as
nefazadone and troglitazone) where adenosine triphosphate (ATP) genera-
tion is progressively compromised over a period of weeks or months during
treatment (McKenzie et al., 1995; Dykens et al., 2008; Xu et al., 2008).
When mitochondrial function deteriorates to a critical level, hepatocellular
necrosis may ensue, releasing or recruiting injury-propagating factors and/
OCR for page 49
BIOMARKERS OF ACUTE IDIOSYNCRATIC HEPATOCELLULAR INJURY
or increasing metabolic demands on neighboring hepatocytes, which may
progress to liver failure.
Some recent data suggest that a small number of critical pathways
may compromise hepatocyte function to produce AIHI. In a recent study,
fluorescent imaging of human hepatocytes was used to examine the effects
of 300 hepatotoxins and nonhepatotoxins on mitochondrial damage, oxi-
dative stress, and intracellular glutathione. These in vitro studies predicted
approximately 60 percent of drugs capable of causing AIHI (many of which
had not been detected in preclinical testing) with a high specificity (a false
positive rate of 0–5 percent) (Xu et al., 2008). Because mitochondrial
damage, oxidative stress, or depletion of intracellular glutathione may
be downstream of molecule-specific events, such as reactive metabolite
accumulation, a drug capable of causing severe AIHI could induce char-
acteristic changes in the serum proteome or metabolome or in the urinary
metabolome that would not be present in patients treated with drugs
incapable of causing AIHI.
Immune Response Theory
Another mechanism proposed to account for the temporal delay in
the onset and progression of liver injury is the production of reactive
metabolites resulting in immune activation (Uetrecht, 2007). Within the
hepatocyte, a drug is bioactivated to a reactive metabolite that binds to and
modifies hepatocellular proteins. When this modified protein or hapten is
presented by antigen-presenting cells to T cells, they transform to cytotoxic
T cells and antibody-producing B cells (Kaplowitz, 2005; Park et al., 2005).
Such drug-induced immune reactions typically occur within the first month
of treatment and more rapidly with rechallenge (Kaplowitz, 2005), as seen
with halothane (Mushin et al., 1971), and may be accompanied by clinical
signs of hypersensitivity, such as fever, rash, and eosinophilia. The role of a
specific hepatotoxin/metabolite in this immune response can be assessed in
some cases by the lymphocyte stimulation test (Kaplowitz, 2005; Sanderson
et al., 2006). These immune responses may be enhanced by acute inflam-
mation or circulating lipopolysaccharide in rodents (Roth et al., 2003), and
may explain why immunoallergic hepatotoxicity is more common in AIDS
patients (Kaplowitz, 2005). It is quite probable that many AIHI cases result
from episodic environmental/infectious/inflammatory changes that occur
during drug therapy and that affect susceptibility or directly trigger a toxic
interaction with a drug.
A variety of data suggest that immune mechanisms may underlie AIHI
even when there are no clinical signs of hypersensitivity; an example is a
report of human leukocyte antigen (HLA) associations with zimelagatran
OCR for page 50
0 DEVELOPMENT OF BIOMARKERS FOR DRUG SAFETY
hepatotoxicity (Kindmark et al., 2007). It is possible that biomarkers of
immune activation could be useful in distinguishing benign ALT elevations
from those that can portend AIHI. In support of this concept, ALT eleva-
tions accompanied by hepatitis symptoms (fatigue, nausea, right upper
quadrant pain) appear to be more predictive of AIHI potential than are
asymptomatic ALT elevations (Nolan et al., 1999). These symptoms may
be mediated by cytokines or other endogenous proteins, which may be
detectable long before symptoms appear.
Failure of Adaptation
If the critical issue in AIHI is failure to adapt to the initial injury, there
may be biomarkers that could identify patients likely to adapt—and, con-
versely, those likely to progress to severe liver injury—at a very early stage
in the injury process.
POTENTIAL NEW bIOMARkERS FOR AIHI
Sources of Candidate biomarkers
Candidate biomarkers for AIHI are emerging from many lines of inves-
tigation, including extensive transcriptomic profiling of rats treated with
a variety of hepatotoxic drugs. In the Liver Toxicity Biomarker Study
(LTBS), pangenomic approaches are being used in rats to identify bio-
markers capable of distinguishing pairs of drugs that are structurally and
pharmacologically similar but differ in that one is capable of causing AIHI
and the other is not. The Predictive Safety Testing Consortium has been
identifying potential liver safety biomarkers but has not yet focused on
detecting those for AIHI.
Another path to identifying potential biomarkers for AIHI is the ongo-
ing effort to study patients who have actually experienced the condition.
The Severe Adverse Event Consortium (SAEC) (Holden, 2008) has begun
whole-genome single nucleotide polymorphism (SNP) analysis on germ-line
DNA from patients who have experienced varying degrees of drug-induced
liver injury, including AIHI. The expanding U.S.-based Drug Induced Liver
Injury Network (DILIN) (Hoofnagle, 2004) will begin genetic analysis on
a similar cohort and has the advantage of maintaining identity links to
the participants so that additional phenotyping studies can be performed.
Because subjects are enrolled in these registries only after a diagnosis of
drug-induced liver injury has been made, it is generally not possible to
obtain blood or urine early in the course of or prior to the injury.
One research priority will be to generate hypotheses that can be tested
in gene banks and in the DILIN subjects themselves. International drug-
OCR for page 51
BIOMARKERS OF ACUTE IDIOSYNCRATIC HEPATOCELLULAR INJURY
induced liver injury registries in the United States, the United Kingdom,
Japan, Spain, Sweden, and Denmark now contain thousands of expert-
adjudicated cases, which can be combined and analyzed for risk factors
predicting progression to AIHI. Mining of large postmarketing adverse
event databases also may suggest drug–environment susceptibility factors
that could lead to testable hypotheses or be used to provide support-
ive data for genetic associations observed in these networks. In addition,
analysis of blood/urine samples obtained in clinical trials of drugs known
to cause AIHI may be useful in identifying biomarkers, especially when
compared with blood/urine samples obtained in clinical trials of drugs
that cause ALT elevations but do not have the potential to cause AIHI. A
large prospective trial in isoniazid-treated patients has been proposed for
this purpose (Watkins et al., 2008). Studying differences in susceptibility
to hepatotoxicity across panels of inbred strains of mice and performing
quantitative trait loci mapping may be a promising approach to gener-
ating hypotheses that would be testable in relatively small numbers of
human subjects. Recent models have emerged in which drugs that cause
AIHI in humans also cause liver injury in animals. It may be productive to
explore biomarkers in these animal models, especially for biomarkers that
are related to injury progression and adaptation and that predict serious
downstream injury. In choosing drugs for study, consideration should be
given to AIHI-associated drugs having a negative comparator in the same
pharmacologic/structural class that is devoid of an equivalent degree of
AIHI liability (e.g., trovafloxacin-levofloxacin).
Finally, the FDA has sponsored a cooperative research and develop-
ment agreement to develop a computer-based model for understanding
and predicting drugs capable of causing AIHI.3 The goal of this effort is
to incorporate current mechanistic knowledge, as well as data and insights
gained from ongoing efforts such as the SAEC and DILIN analyses. This
evolving model could suggest novel biomarkers and provide a biological
rationale for biomarkers discovered by other means.
validation of Candidate biomarkers
Biomarkers that are predictive in small clinical trials of short duration
would be extremely useful. Potential biomarkers could be tested by admin-
istering examples of drugs both with and without AIHI liability to small
groups of closely monitored patients or healthy volunteers and analyzing
prospectively collected blood/urine samples. For example, the first pair of
drugs studied in the LTBS were tolcapone (whose use is restricted because of
3 More information about this agreement can be found at http://www.entelos.com/newsReleases.
php?ID=press101.
OCR for page 52
DEVELOPMENT OF BIOMARKERS FOR DRUG SAFETY
liver toxicity) and entacapone. Since both drugs are in clinical use, it should
be possible to test candidate biomarkers that emerge from the LTBS effort
in patients or possibly healthy volunteers treated with these drugs. Short-
term studies of this design with healthy volunteers may be ethical since the
onset of liver injury is typically delayed weeks or months with tolcapone
(Olanow and Watkins, 2007). However, it is possible that drugs capable of
causing AIHI will be distinguishable only once liver injury has begun as sig-
naled by ALT elevations, so that longer-term treatment would be required
to evoke the phenotype. In this case, blood/urine samples would have to be
obtained from patients with ALT elevations induced by drugs capable of
causing AIHI and then compared with blood/urine samples obtained from
patients with ALT elevations induced by drugs that do not cause AIHI. If
(for at least some drugs) human AIHI occurs via interaction with an acute
inflammatory stress, then plasma biomarkers based on this mode of action
can be examined. For example, prolonged elevation of plasma cytokines,
hemostatic biomarkers, and/or markers of neutrophil activation, when
used in conjunction with traditional biomarkers such as ALT, might prove
predictive. Generally, biomarkers with mechanistic/mode-of-action under-
pinnings are likely to be the most consistent predictors of AIHI.
True validation of biomarkers will ultimately require large numbers of
samples obtained from individuals with well-known phenotypes, includ-
ing both healthy and diseased populations, as well as populations treated
with many different drugs. One path would be to institute protocols for
standard data and blood/urine collection once ALT elevations have been
observed in a clinical trial. An example of a liver safety data management
system is eDish (Guo et al., 2008). This or a similar format could be directly
linked to the sample bank, and would allow immediate identification of
individuals of interest and immediate access to all pertinent clinical and
laboratory data for those patients for detailed evaluation. Because the true
potential of a drug to cause AIHI may not be evident preapproval, the
blood/urine samples and data bank would need to be maintained for some
time postmarketing. It would obviously be ideal if scientists had access to
samples and clinical data from the clinical trials of many of the drugs listed
in Table 5-1.
HIGHLIGHTS OF THE bREAkOuT DISCuSSION
In plenary session, John Bloom of Eli Lilly presented the main con-
clusions of participants in the breakout session on biomarkers for liver
toxicity. Discussants in the breakout session observed that candidate AIHI
biomarkers are best identified and validated in three relevant human popu-
lations: Hy’s Law cases; subjects in prospective, controlled clinical trials
with established and well-characterized AIHI agents, including isoniazid;
OCR for page 53
BIOMARKERS OF ACUTE IDIOSYNCRATIC HEPATOCELLULAR INJURY
and subjects in clinical trials who receive a drug known to cause ALT eleva-
tions but not yet known to cause AIHI. Some data suggest that subpopula-
tions of these groups may exhibit changes that share common mechanisms
with those associated with AIHI. Discussants identified the following six
priority research efforts.
Accessing and Characterizing Hy’s Law Cases
The first priority is to develop methods for overcoming key barriers
to accessing clinical information and biospecimens from Hy’s Law cases,
which arguably constitute the most relevant population to study. This pri-
ority need raises several important questions. How can these rare cases be
better accessed and characterized? How can well-annotated specimens be
obtained, including specimens from matched controls? Can ongoing initia-
tives such as those of SAEC, DILIN, and other partnerships be integrated
more effectively? How can electronic health records and large databases,
such as those from the U.S. Department of Veterans Affairs or private
insurers, be better leveraged? Can a warehouse for data be established?
Could such an effort be integrated into the FDA’s Sentinel Initiative?
Developing and Implementing Protocols for Specimen and
Data Collection in Clinical Trials of Specific Marketed
Drugs known to Either Cause or Not Cause AIHI
A second priority is to develop and implement protocols for specimen
and data collection in prospective clinical trials of isoniazid and other drugs
known to cause AIHI or known to cause ALT elevations but not AIHI. A
number of key questions need to be addressed. What would a protocol look
like in terms of the subjects who are enrolled and controls for concomitant
treatments and diseases? Are there markers that can be used to enrich this
patient population? Can signals or markers for adaptation and severity
of liver injury be differentiated and stratified? Are there markers that can
predict, rather than simply demonstrate, the effect of the disorder? What
other agents should be considered for prospective trials? To what extent
will the identified markers be agent-specific and therefore not more broadly
applicable? How can this kind of study be sponsored or funded, and can it
be coupled with ongoing studies?
Investigating ALT Signals in Clinical Trials During Drug Development
A third need is to develop and implement protocols for standardized
data and biospecimen collection in clinical trials when an ALT signal is
identified. Important questions include the following: What should be the
OCR for page 54
DEVELOPMENT OF BIOMARKERS FOR DRUG SAFETY
trigger for collection, and which specimens should be collected? How can
standardization of specimens and data be achieved, including ascertainment
as well as phenotyping? What should be the role of regulators? Should
access to specimens and data be restricted? How can the risk for the spon-
sor be managed in such a situation?
Making use of Existing Databases
A fourth priority is to conduct a thorough examination of existing
FDA liver safety databases from Phase III clinical trials, and perhaps from
the Adverse Event Reporting System database, to test hypotheses for the
more frequent benign ALT elevations. An important hypothesis is that such
elevations in this population, or more likely a subset of these patients, are
mechanistically linked to AIHI, and that this link could be validated or at
least corroborated. How can these data be mined? How can privacy issues
be addressed? How can alignment among regulatory agencies and private
companies be achieved? What are the incentives for alignment? What
resources and oversight are needed? Can this research be decoupled from
regulatory decision making?
Prioritizing biomarker Discovery
A fifth research need is to prioritize biomarker discovery options using
the data and biospecimens from the three populations described above.
This work will require answers to a number of questions. How should the
right specimens be collected when candidate markers are not known? How
should options be kept open? Should candidate biomarker domains be pri-
oritized according to established and emerging hypotheses? For instance,
if a toxic metabolite or other form of injury leads to subsequent immune
responses, there are pathways from which one can derive biomarkers for
these responses. Should biomarker discovery searches be prioritized along
the lines of these hypotheses or enabling technology platforms? Should this
guide which specimens are collected?
Identifying and Prioritizing Nonclinical Research Options
Finally, there is a need to identify and prioritize nonclinical research
options for generating biomarker hypotheses for testing in clinical speci-
men banks. Can animal models enable method development? If relevant
models are identified, can they provide information on progression fac-
tors, reversibility, the kinetics of biomarker changes, and other questions
enabled by tightly controlled conditions? Can nonclinical studies be linked
to clinical studies to inform biomarker identification? Are there surrogate
OCR for page 55
BIOMARKERS OF ACUTE IDIOSYNCRATIC HEPATOCELLULAR INJURY
in vivo or in silico models that can suggest new candidate biomarkers for
human studies?
Next Steps
The questions identified by the discussants in each of the above six
areas attest to the complexity and challenges of implementing such a multi-
disciplinary and interinstitutional endeavor. Many additional questions
remain regarding coordination, oversight, and sponsorship of efforts in
these areas. Breakout session participants discussed potential roles for the
IOM in facilitating efforts of the research community in addressing these
questions. One suggested approach would be for the IOM to convene and
facilitate working groups in each of the six priority research areas.
REFERENCES
Aithal, G. P., M. J. Rawlins, and C. P. Day. 1999. Accuracy of hepatic adverse drug reaction
reporting in one English health region. British Medical Journal 319:1541.
Andrade, R. J., M. I. Lucena, M. C. Fernández, G. Pelaez, K. Pachkoria, E. García-Ruiz,
B. García-Muñoz, R. González-Grande, A. Pizarro, J. A. Durán, M. Jiménez, L. Rodrigo,
M. Romero-Gomez, J. M. Navarro, R. Planas, J. Costa, A. Borras, A. Soler, J. Salmerón,
R. Martin-Vivaldi, and Spanish Group for the Study of Drug-Induced Liver Disease.
2005. Drug-induced liver injury: An analysis of 461 incidences submitted to the Spanish
registry over a 10-year period. Gastroenterology 129:512–521.
Bjornsson, E., and R. Olsson. 2005. Outcome and prognostic markers in severe drug-induced
liver disease. Hepatology 42(2):481–489.
Browning, J. D., L. S. Szczepaniak, and R. Dobbins. 2004. Prevalence of hepatic steatosis in an
urban population in the United States: Impact of ethnicity. Hepatology 40:1387–1395.
Daly, A. K., G. P. Aithal, J. B. S. Leathart, R. A. Swainsbury, T. S. Dang, and C. P. Day. 2007.
Genetic susceptibility to diclofenac-induced hepatotoxicity: Contribution of UGT2B7,
CYP2C8, and ABCC2 genotypes. Gastroenterology 132(1):272–281.
Dykens, J. A., J. D. Jamieson, L. D. Marroquin, S. Nadanaciva, J. J. Xu, M. C. Dunn, A. R.
Smith, and Y. Will. 2008. In vitro assessment of mitochondrial dysfunction and cytotoxicity
of nefazodone, trazodone and buspirone. Toxicological Sciences 103(2):335–345.
FDA (Food and Drug Administration). 2007. Guidance for industry (draft) drug-induced
liver injury: Premarketing clinical evaluation. www.fda.gov/cder/guidance/7507dft.htm
(accessed October 17, 2008).
Green, R. M., and S. Flamm. 2002. AGA technical review on the evaluation of liver chemistry
tests. Gastroenterology 123:1367–1384.
Guo, T., K. Gelperin, and J. Senior. 2008. A tool to help you decide: Detect potentially serious
liver injury. AASLD-FDA-NIH-PhRMA-Hepatotoxicity Special Interest Group meet-
ing. http://www.fda.gov/cder/livertox/presentations2008/D-GelperinGuo2.pdf (accessed
October 17, 2008).
Holden, A. 2008. Detecting and investigating drug-induced adverse events: The international
serious adverse event consortium’s experience. www.fda.gov/cder/livertox/presentations2008/
T-Holden2.pdf (accessed October 17, 2008).
Hoofnagle, J. H. 2004. Drug-Induced Liver Injury Network (DILIN). Hepatology 40(4):773.
Kaplowitz, N. 2005. Drug hepatotoxicity. Nature Reviews. Drug Discovery 4:489–499.
OCR for page 56
DEVELOPMENT OF BIOMARKERS FOR DRUG SAFETY
Kindmark, A., A. Jawaid, C. G. Harbron, B. J. Barratt, O. F. Bengtsson, T. B. Andersson, S.
Carlsson, K. E. Cederbrant, N. J. Gibson, M. Armstrong, M. E. Lagerström-Fermér, A.
Dellsén, E. M. Brown, M. Thornton, C. Dukes, S. C. Jenkins, M. A. Firth, G. O. Harrod,
T. H. Pinel, S. M. Billing-Clason, L. R. Cardon, and R. E. March. 2007. Genome-
wide pharmacogenetic investigation of a hepatic adverse event without clinical signs of
immunopathology suggests an underlying immune pathogenesis. The Pharmacogenomics
Journal 8(3)186–195.
Larrey, D. 2002. Epidemiology and individual susceptibility to adverse drug reactions affecting
the liver. Seminars in Liver Disease 22:145–155.
Lee, W. M. 2003. Drug-induced hepatotoxicity. New England Journal of Medicine
349(5):474–485.
McKenzie, R., M. W. Fried, R. Sallie, H. Conjeevaram, A. M. Di Bisceglie, Y. Park, B.
Savarese, D. Kleiner, M. Tsokos, and C. Luciano. 1995. Hepatic failure and lactic acidosis
due to fialuridine (FIAU), an investigational nucleoside analogue for chronic hepatitis B.
New England Journal of Medicine 333:1099–1105.
Mitchell, J. R., M. W. Long, U. P. Thorgeirsson, and D. J. Jollow. 1975. Acetylation rates
and monthly liver function tests during one year of isoniazid preventive therapy. Chest
68(2):181–190.
Murata, K., M. Hamada, K. Sugimoto, and T. Nakano. 2007. A novel mechanism for drug-
induced liver failure: inhibition of histone acetylation by hydralazine derivatives. Journal
of Hepatology 46:322–329.
Mushin, W. W., M. Rosen, and E. V. Jones. 1971. Post-halothane jaundice in relation to previ-
ous administration of halothane. British Medical Journal 3(5765):18–22.
Nolan, C. M., S. V. Goldberg, and S. E. Buskin. 1999. Hepatotoxicity associated with isoniazid
preventative therapy: A 7-year survey from a public health tuberculosis clinic. Journal of
the American Medical Association 282(23):2207–2208.
Olanow, C. R., and P. B. Watkins. 2007. Tolcapone: An efficacy and safety review. Clinical
Neuropharmacology 30(5):287–294.
Olson, H., G. Betton, D. Robinson, K. Thomas, A. Monro, G. Kolaja, P. Lilly, J. Sanders,
G. Sipes, W. Bracken, M. Dorato, K. Van Deun, P. Smith, B. Berger, and A. Heller. 2000.
Concordance of the toxicity of pharmaceuticals in humans and in animals. Regulatory
Toxicology and Pharmacology 23:56–57.
Olsson, R., C. Wesslau, T. W. Olsson, and L. Zettergren. 1989. Elevated aminotransferases
and alkaline phosphatases in unstable diabetes mellitus without ketoacidosis or hypo-
glycemia. Journal of Clinical Gastroenterology 11:541–545.
Park, B. K., N. R. Kitteringham, J. L. Maggs, M. Pirmohamed, and D. P. Williams. 2005. The
role of metabolic activation in drug-induced hepatotoxicity. Annual Review of Pharma-
cology and Toxicology 45:177–202.
Pauls, L. 2004 (February 5). Can we predict hepatotoxicity based on MO reviews of submitted
data? Presentation by Center for Drug Evaluation and Research. FDA Hepatotoxicity
Steering Committee. http://www.fda.gov/cder/livertox/presentations2004/LP5Feb04.ppt
(accessed October 5, 2008).
Rautou, P. E., D. Cazals-Hatem, R. Moreau, C. Francoz, G. Feldmann, D. Lebrec, E. Ogier-
Denis, P. Bedossa, D. Valla, and F. Durand. 2008. Acute liver cell damage in patients
with anorexia nervosa: A possible role of starvation-induced hepatocyte autophagy.
Gastroenterology 135:840–848.
Roth, R. A., J. P. Luyendyk, J. F. Maddox, and P. E. Ganey. 2003. Inflammation and drug
idiosyncrasy—is there a connection? The Journal of Pharmacology and Experimental
Therapeutics 307(1):1–8.
Sanderson, J. P., D. J. Naisbitt, and B. K. Park. 2006. Role of bioactivation in drug-induced
hypersensitivity reactions. The AAPS Journal 8(1):E55–E64.
OCR for page 57
BIOMARKERS OF ACUTE IDIOSYNCRATIC HEPATOCELLULAR INJURY
Sayuk, G. S., J. E. Elwing, and M. Lisker-Melman. 2007. Hepatic glycogenosis: An under-
recognized source of abnormal liver function tests? Digestive Diseases and Sciences
52:936–938.
Temple, R. 2001. Hepatotoxicity through the years: Impact on the FDA. www.fda.gov/cder/
livertox/meeting2001.htm#Drug-Induced (accessed October 17, 2008).
Thulin, P., I. Rafter, K. Stockling, C. Tomkiewicz, E. Norjavaara, M. Aggerbeck, H. Hellmold,
E. Ehrenborg, U. Andresson, I. Cotgreave, and B. Glinghammar. 2008. PPARalpha
regulates the hepatotoxic biomarker alanine aminotransferase (ALT1) gene expression in
human hepatocytes. Toxicology and Applied Pharmacology 231(1):1–9.
Uetrecht, J. 2007. Idiosyncratic drug reactions: Current understanding. Annual Review of
Pharmacology and Toxicology 47:513–539.
Watkins, P. B. 1998. Hepatic dysfunction associated with troglitazone. New England Journal
of Medicine 338:916–917.
Watkins, P. B., H. J. Zimmerman, M. J. Knapp, S. I. Gracon, and K. W. Lewis. 1994. Hepato-
toxic effects of tacrine administration in patients with Alzheimer’s disease. Journal of the
American Medical Association 271(13):992–998.
Watkins, P. B., P. J. Seligman, J. S. Pears, M. I. Avigan, and J. R. Senior. 2008. Using con-
trolled clinical trials to learn more about acute drug-induced liver injury. Hepatology
48(5):1680–1689.
Wilke, R. A., D. W. Lin, D. M. Roden, P. B. Watkins, D. Flockhart, I. Zineh, K. M. Giacomini,
and R. M. Krauss. 2007. Identifying genetic risk factors for serious adverse drug reactions:
Current progress and challenges. Nature Reviews Drug Discovery 7:904–916.
Xu, J. J., P. V. Henstock, M. C. Dunn, A. R. Smith, J. R. Chabot, and D. de Graaf. 2008.
Cellular imaging predictions of clinical drug-induced liver injury. Toxicological Sciences
105:97–105.
Zimmerman, H. J. 1968. The spectrum of hepatotoxicity (Kober lecture). Perspectives in
Biology and Medicine 12(1)135–161.
