128(1), 57–71 (2012)
doi:10.1093/toxsci/kfs149
Advance Access publication April 26, 2012
toxicological sciences
Mode of Action Associated With Development of Hemangiosarcoma in
Mice Given Pregabalin and Assessment of Human Relevance
Kay A. Criswell,*,1 Jon C. Cook,* Zbigniew Wojcinski,† David Pegg,‡ James Herman,§ David Wesche,¶ John Giddings,||
Joseph T. Brady,* and Timothy Anderson*
*Pfizer Worldwide Research & Development, Drug Safety Research & Development, Groton, Connecticut 06340; †Drug Development Preclinical Services,
LLC, Ann Arbor, Michigan 48103; ‡Michigan Technology and Research Institute, Ann Arbor, Michigan 48104; §Integrated Nonclinical Development Solutions,
Ann Arbor, Michigan 48103; ¶Ann Arbor, Michigan 48105; and ||University of Wales College of Medicine, Heath Park Cardiff, CF14 4XN, United Kingdom
1
To whom correspondence should be addressed at Pfizer Worldwide Research & Development, 2400 Eastern Point Road, Groton, CT 06340.
Fax: (860) 441-7049. E-mail: [email protected].
Received November 8, 2011; accepted April 3, 2012
Pregabalin increased the incidence of hemangiosarcomas in carcinogenicity studies of 2-year mice but was not tumorigenic in rats.
Serum bicarbonate increased within 24 h of pregabalin administration in mice and rats. Rats compensated appropriately, but mice
developed metabolic alkalosis and increased blood pH. Local tissue
hypoxia and increased endothelial cell proliferation were also confirmed in mice alone. The combination of hypoxia and sustained
increases in endothelial cell proliferation, angiogenic growth factors, dysregulated erythropoiesis, and macrophage activation is
proposed as the key event in the mode of action (MOA) for hemangiosarcoma formation. Hemangiosarcomas occur spontaneously
in untreated control mice but occur only rarely in humans. The
International Programme on Chemical Safety and International
Life Sciences Institute developed a Human Relevance Framework
(HRF) analysis whereby presence or absence of key events can be
used to assess human relevance. The HRF combines the MOA with
an assessment of biologic plausibility in humans to assess human
relevance. This manuscript compares the proposed MOA with Hill
criteria, a component of the HRF, for strength, consistency, specificity, temporality, and dose response, with an assessment of key
biomarkers in humans, species differences in response to disease
conditions, and spontaneous incidence of hemangiosarcoma to
evaluate human relevance. Lack of key biomarker events in the
MOA in rats, monkeys, and humans supports a species-specific
process and demonstrates that the tumor findings in mice are not
relevant to humans at the clinical dose of pregabalin. Based on
this collective dataset, clinical use of pregabalin would not pose an
­increased risk for hemangiosarcoma to humans.
Key Words: hemangiosarcoma; mice; pregabalin; α2δ subunit;
mode of action; human relevance; clinical trials.
Pregabalin is effective for the treatment of neuropathic pain
(associated with diabetic peripheral neuropathy, postherpetic
Disclaimer: The authors certify that all research involving human subjects was
done under full compliance with all government policies and the Helsinki Declaration.
neuralgia, and following spinal cord injury), as an adjunctive
therapy in the treatment of partial seizures, and in the treatment of generalized anxiety and fibromyalgia. An increased incidence of a single tumor type, hemangiosarcomas, was seen
in male and female mice when fed with the diet for approximately 2 years (Pegg et al., unpublished data). These tumors
were located predominantly in liver, spleen, and bone marrow,
which are hematopoietic tissues in mice. In contrast, there was
no increased incidence of hemangiosarcomas, or any other
tumor type, in rats fed with dosages that achieved equivalent
concentrations of pregabalin in the feed for 2 years. In a battery
of tests, pregabalin was found to be nongenotoxic (Pegg et al.,
unpublished data).
Hemangiosarcomas are endothelial cell–derived tumors that
comprise poorly differentiated, proliferating endothelial cells
(Mendenhall et al., 2006). These tumors form spontaneously
and in response to many different compounds in mice but
are rare in humans. To our knowledge, there are only two
examples of hemangiosarcoma induction by a chemical
stimulus that occur in both humans and rodents (i.e., mouse
and rat). Both examples are genotoxic carcinogens (vinyl
halides and Thorotrast), and both produce primarily liver
hemangiosarcoma (Boivin-Angele et al., 2000; Lipshutz et al.,
2002). In contrast, numerous commercial pharmaceutical
products and chemicals that produce hemangiosarcomas solely
in rodents act by nongenotoxic, proliferative mechanisms
(Cohen et al., 2009).
Previous manuscripts in this series have provided the proposed mode of action (MOA) for the species-specific, nongenotoxic formation of hemangiosarcomas in mice (Criswell
et al., unpublished data), whereas this manuscript provides an
evaluation of human relevance including assessment of several
biomarkers of the early events in humans. The Hill Criteria
(Hill, 1965) have long been utilized by epidemiologists to evaluate data and identify causality. The International Programme
© The Author 2012. Published by Oxford University Press on behalf of the Society of Toxicology. All rights reserved.
For permissions, please email: [email protected]
58
CRISWELL ET AL.
on Chemical Safety (IPCS) has recommended this rigorous
process and the International Life Sciences-Risk Sciences
Institute (ILSI-RSI) outlined a methodology referred to as the
Human Relevance Framework (HRF) to evaluate human risk
from animal hazard assessment studies. The Hill criteria form
a key component of the HRF process. An excellent example
and model for using this evaluation approach was provided by
Pastoor et al. (2005), which described thiamethoxam-related
mouse tumors and their relevance to human health. This manuscript follows the ILSI-RSI HRF approach outlined by Meek
and coworkers (2003) where the MOA in rodents is described
and then three fundamental questions are answered to make
a HRF assessment: (1) Is the weight of evidence sufficient to
establish the MOA in animals? (2) Are the key events in the
animal MOA plausible in humans? and (3) Taking into account
kinetic and dynamic factors, is the animal MOA plausible in
humans? The Hill criteria are embedded in Question 1.
SUMMARY OF PROPOSED MOA IN RODENTS
Figure 1 summarizes the key events in the pregabalin MOA
and contrasts the findings between those seen in mice versus
rats. Mice react inadequately to metabolic alkalosis following
pregabalin administration, thus creating a cascade of events
starting with an increase in serum bicarbonate and leading to
sustained depression of respiratory rate and minute volume
(Criswell et al., unpublished data). The sustained increase in
blood pH hampers oxygen release from hemoglobin due to the
Bohr effect, creating conditions conducive to hypoxia. Persistent perturbation of acid-base balance in mice is consistent with
the identification of chronic tissue hypoxia in this species (Fig. 1,
Criswell et al., unpublished data). One method of compensating for hypoxia is an increased production of erythrocytes to
provide greater oxygen-carrying capacity. An appropriate circulating erythrocytosis is seen in rats but is lacking in mice,
highlighting an additional key species-specific event that contributes to tissue hypoxia in mice alone (Criswell et al., unpublished data). The inability of the mouse to increase erythrocytes
is due to dysregulated erythropoiesis that is apparent in bone
marrows.
Spontaneous hemangiosarcomas in mice occur primarily
in hematopoietic tissues (liver, spleen, bone marrow) (Cohen
et al., 2009), and pregabalin increased the incidence of hemangiosarcomas in these tissues as well (Pegg et al., unpublished data). In mice, pregabalin treatment increased platelet
and megakaryocyte counts, decreased myeloid to erythroid
ratio (M:E) up to 49%, increased the number of bone marrow macrophages/erythrophages, and was associated with the
congestion of bone marrow and spleen and with extramedullary splenic hematopoiesis (Criswell et al., unpublished data).
Bone marrow changes occurred early, within 1–3 months
of the drug in diet, so were not the result of tumor formation (Criswell et al., unpublished data) and were consistent
with a temporal association. These findings demonstrate that
dysregulated erythropoiesis is occurring in mouse bone marrow where enhanced red blood cell (RBC) production is occurring, but there is a failure of release. This failure of release
explains the decreased RBC response of the mouse when
compared with the rat and results in macrophage activation
and increased hemosiderin production that increases oxidative
stress (Fig. 1, Criswell et al., unpublished data).
In contrast to mice, pregabalin administration decreased
overall bone marrow cellularity in rats, including megakaryocytes (up to 24% decrease, Criswell et al., unpublished data).
Decreases in megakaryocytes in rats correspond with decreased
peripheral platelet counts, which are in contrast to elevations in
platelet counts and highly proliferative bone marrow in mice.
There was no evidence of macrophage or erythrophage accumulation in rats, and M:E ratios were similar to those in the
controls (Criswell et al., unpublished data). The effects seen in
mice and absent in rats and monkeys suggest an association between hematopoietic stimulus and selective endothelial tumor
induction in mice. In addition, they provide potential sentinel
biomarkers for assessing the potential risk to humans.
Hematopoiesis and angiogenesis are closely linked. Endothelial
cells, the cell type of origin for hemangiosarcomas, are believed
to be derived from a common hematopoietic precursor, the
hemangioblast (Keller, 2001). Two hypotheses have been described
for the cellular ontogeny of spontaneous hemangiosarcomas:
either a series of mutations occur in fully differentiated vascular
endothelial cells, giving them malignant potential or the malignant
endothelial cells arise from bone marrow/hematopoietic-derived
stem cells near the time of endothelial cell commitment (LameratoKozicki et al., 2006). Therefore, a highly proliferative, reactive
bone marrow is relevant to hemangiosarcoma formation because
bone marrow endothelial cells may migrate and repopulate other
tissues such as the liver and spleen where pregabalin increased the
incidence of hemangiosarcoma.
An inflammatory component is also considered important in the MOA (Cohen et al., 2009) and is evidenced by
a dose- and time-dependent increase in macrophages in the
bone marrow, spleen, and liver of pregabalin-treated mice
(Criswell et al., unpublished data). Macrophages can release
reactive oxygen species that can damage endothelial cells as
well as release angiogenic cytokines such as interleukin-6 to
stimulate endothelial cell proliferation. Using immunohistochemistry, pregabalin was shown to increase tissue angiogenic growth factors (vascular endothelial growth factor [VEGF]
and basic fibroblast growth factor [bFGF]) (Criswell et al.,
unpublished data) and increase endothelial cell proliferation
as early as 2 weeks and sustained out to 12 months (Criswell
et al., unpublished data). Therefore, mice treated with pregabalin demonstrate all of the components proposed by the
unified MOA for nongenotoxic hemangiosarcoma formation
(Cohen et al., 2009).
Angiogenesis is an important physiologic process. It
­occurs monthly in the uterus of menstruating women, during pregnancy with the formation of the placenta, and in
HUMAN RELEVANCE OF PREGABALIN-INDUCED HEMANGIOSARCOMA
59
FIG. 1. Mode of action of hemangiosarcoma formation in pregabalin-treated mice.
wound ­healing. Because angiogenesis is so important, it is
tightly regulated with over 20 angiogenic growth factors and
more than 300 angiogenic inhibitors (http://www.angio.org/
understanding/fact.php). Although angiogenesis is a complex
process, under regulated conditions the concentration of angiogenic inducers approximately equals the concentration of
inhibitors (Fig. 2). During physiologic angiogenesis, hypoxia
stimulates paracrine release of angiogenic growth factors.
These growth factors bind to their specific receptors on endothelial cells, initiating a cascade of events including endothelial cell activation and endothelial cell proliferation, which
ultimately leads to new blood vessel formation (Fig. 2). Once
oxygenation is reestablished, the angiogenic process returns
to a balanced state. Hence, these sequence of events characterize regulated angiogenesis.
In contrast, the production of hemangiosarcomas in
pregabalin-treated mice is believed to occur due to pathologic
angiogenesis that is characterized by dysregulated angiogenesis
(Fig. 3). Dysregulated angiogenesis refers to the condition
where normal blood vessel formation does not occur in
response to a hypoxia signal. Hence, there is a continued
hypoxia signal leading to angiogenic growth factors exceeding
inhibitors coupled with macrophage activation and release
of reactive oxygen species that damage endothelial cells,
all of which promote increased endothelial cell activation,
proliferation, and damage of endothelial cell DNA (Fig. 3).
Figure 4 illustrates a hemangiosarcoma in mouse liver where
there is a pooling of blood from the blind vessels surrounded
by the hemangiosarcoma.
Hypoxia has been demonstrated in the liver of mice treated with pregabalin as well as time-related increases in multiple angiogenic growth factors (Criswell et al., unpublished
data). Additionally, the discontinuous endothelium found in
liver, spleen, and bone marrow is supported by numerous macrophages (Kupffer cells in the liver), an additional source of
angiogenic growth factors. Angiogenesis is controlled by two
different components, paracrine and autocrine. The paracrine
component is driven by nonendothelial expression of angiogenic growth factors such as VEGF and bFGF. In the autocrine
component, the endothelial cells themselves are induced to
express VEGF (Fong, 2009). In pregabalin-treated mice, paracrine VEGF and bFGF were demonstrable in bone marrow
and spleen and were particularly noteworthy in macrophages,
megakaryocytes, and early erythroid cells (Criswell et al.,
60
CRISWELL ET AL.
FIG. 2. Regulated angiogenesis. Regulated angiogenesis leads to new blood vessel formation with reoxygenation of tissues.
unpublished data). In the liver of these animals, vascular
­endothelial growth factor receptor 2 expression was increased.
This suggests that both paracrine and autocrine increases in
angiogenic growth factors are involved in pregabalin-treated
mice, which drive an increase in the proliferation rate of endothelial cells (Fig. 1) (Criswell et al., unpublished data).
The increase in endothelial cell proliferation can lead to an
­enhanced potential to fix spontaneous mutations because more
cells are undergoing mitosis, resulting in growth dysregulation,
clonal expansion of altered cells, and ultimately hemangiosarcoma formation (Cohen et al., 2009). Based on the above, the
MOA for pregabalin has been established in mice and is consistent with the proposed unified MOA for hemangiosarcoma
by other nongenotoxic compounds (Cohen et al., 2009).
On the basis of the above MOA described for pregabalin
in mice, we address the three fundamental questions to make
a HRF assessment: (1) Is the weight of evidence sufficient to
establish the MOA in animals? (2) Are the key events in the
animal MOA plausible in humans? and (3) Taking into account
kinetic and dynamic factors, is the animal MOA plausible in
humans?
Question 1: Is the Weight of Evidence Sufficient to Establish
the MOA in Animals?
The Hill criteria are a list of nine characteristics that help distinguish association from causation. These criteria include (1)
analogy (considerations of alternate explanations), (2) strength
of association, (3) consistency of association, (4) specificity (a
single putative cause produces a specific effect), (5) temporality
(exposure precedes outcomes), (6) dose-response relationships,
(7) biologic plausibility, (8) coherence (association is compatible with existing theory), and (9) experiment (the condition
can be prevented or ameliorated by an appropriate experimental
regimen). Numerous studies ranging in length from 24 h to 2
years have amassed the data necessary to develop the proposed
MOA and evaluate these criteria in mice, rats, monkeys, and
humans administered with pregabalin.
Analogy (Consideration of Alternate Explanations). Rodent nongenotoxic carcinogens generally act by modifying the
normal physiology of specific tissues, resulting in increased
cell proliferation. Increased cell proliferation can result from
increased cell replication due to cytotoxicity/regenerative hy-
HUMAN RELEVANCE OF PREGABALIN-INDUCED HEMANGIOSARCOMA
61
FIG. 3. Dysregulated angiogenesis. Dysregulation of angiogenesis may lead to enhanced endothelial cell proliferation with enhanced potential to fix spontaneous mutations (because more cells are undergoing mitosis), resulting in growth dysregulation, clonal expansion of altered cells, and ultimately hemangiosarcoma formation.
perplasia, direct or indirect (e.g., hormonally induced) mitogenesis, decreased apoptosis, or combinations of these effects. Cell
proliferation is required for fixation of spontaneous mutations
secondary to replicative DNA synthesis and/or may facilitate
selective clonal growth of spontaneously initiated cells. Neoplasms ultimately develop from these spontaneously initiated
or preneoplastic cells.
Data indicate that pregabalin increases endothelial cell
proliferation in mice treated for at least 2 weeks and up
to 1 year at a dose that resulted in an increased incidence
in hemangiosarcomas, a key feature in nongenotoxic
carcinogenesis (Criswell et al., unpublished data). There was
no histopathologic evidence of cytotoxicity in endothelial
cells of mice in repeat-dose toxicity studies with pregabalin
(Pegg et al., unpublished data); thus, regenerative hyperplasia
is unlikely to be responsible for increased endothelial cell
proliferation. In vivo, pregabalin did not affect apoptosis of
hepatic endothelial cells (unpublished data), and it did not
induce proliferation in vitro of splenic endothelial cells isolated
from B6C3F1 mice, indicating that pregabalin is not a direct
endothelial cell mitogen (unpublished data). Therefore, these
alternate explanations are not operational, and the most likely
explanation is dysregulated angiogenesis.
Strength of Association and Dose-Response Relationships. Endothelial cell proliferation is essential for the initiation of
angiogenesis. It has been demonstrated in multiple models that
the complex process of angiogenesis begins when cells respond
to hypoxia, allowing accumulation of hypoxia-inducible ­factors
and secondarily increasing paracrine and autocrine angiogenic
growth factors. Hypoxia is a commonly accepted trigger for endothelial cell proliferation and angiogenesis (Fong, 2008, 2009;
Pugh and Ratcliffe, 2003; Schäfer et al., 2003; Semenza, 2001).
Tissue-specific hypoxia was demonstrable in mice treated with
62
CRISWELL ET AL.
FIG. 4. This slide illustrates a hemangiosarcoma in a B6C3F1 mouse liver.
The red region is the pooling of blood that occurs when the endothelial cells do
not form the normal tube-like blood vessel structure.
pregabalin but not in rats (Criswell et al., unpublished data).
This correlates with a mouse-specific increase in endothelial cell
proliferation and increased incidence of h­ emangiosarcomas.
Pregabalin treatment at exposures comparable to those
achieved in the mouse carcinogenicity studies did not induce hemangiosarcoma or tumors of any type in rats (Pegg et al., unpublished data), a species also known to develop hemangiosarcoma
spontaneously but at a much lower frequency than mice (Poteracki and Walsh, 1998; Walsh and Poteracki, 1994). Although
dose-response relationships were not evaluated for all endpoints
of experiments conducted over a 2-year period, a dose-response
relationship was evident in endothelial cell proliferation, bone
marrow changes, increases in macrophage numbers, and platelet
counts and function in mice (Criswell et al., unpublished data).
Additionally, B6C3F1 mice exposed to pregabalin demonstrated higher levels of respiratory depression than CD-1 mice with
equivalent exposures (Criswell et al., unpublished data). Strain
sensitivity of response to respiratory depression correlates with
higher levels of hemangiosarcoma formation in B6C3F1 mice
compared with CD-1 mice (Pegg et al., unpublished data).
Consistency of Data. Pregabalin produces hemangiosarcoma in two strains of mice (B6C3F1 and CD-1) but not in two
bioassays conducted with Wistar rats. Effects on bicarbonate,
bone marrow, respiration, metabolic changes, and endothelial
cell proliferation were reproducible at several time points and
in repetitive studies. Consistent with the criteria that hypoxia is
an accepted trigger for endothelial cell proliferation, evidence
of hypoxia in the mouse is accompanied by endothelial cell
proliferation, whereas rats do not demonstrate tissue hypoxia
and do not increase endothelial cell proliferation (Criswell
et al., unpublished data).
Specificity (Single Putative Cause Produces Specific Effect). To our knowledge, there is no specific disease model
that produces the exact effects observed with pregabalin administration in mice. However, loss of von Hippel-Lindau (VHL)
tumor suppressor protein represents an extreme condition that
mimics a sustained hypoxic state in which hypoxia inducible
factor (HIF) is constitutively stabilized, leading to the inappropriate activation of genes associated with endothelial cell
proliferation, increased VEGF, and erythropoietin. Mutation
in VHL tumor suppressor gene in mice produces life-long elevation of HIF with associated elevations in VEGF and other
angiogenic growth factors and predisposes VHL+/– heterozygote mice to hemangiosarcoma formation (Kleymenova et al.,
2004). This demonstrates the relationship between chronic
life-long hypoxia and hemangiosarcoma formation in mice.
In humans, VHL disease is an autosomal dominant disorder
that is associated predominantly with clear cell renal carcinomas and benign retinal hemangioblastomas, but there is no
evidence that transformation to malignant hemangiosarcomas
occurs in people (Gordeuk et al., 2004; Kleymenova et al.,
2004; ­Nordstrom-O’Brien, 2010), suggesting a species-specific
sensitivity for this tumor type. Although the susceptibility is
undoubtedly different between species, it should not be oversimplified because benign angiomatous proliferative lesions
can also be identified in VHL knockout mice. Interestingly,
however, the Kleymenova (2004) VHL mutation model failed
to produce any effects on renal cancer in mice, further demonstrating species differences. Additionally, even in VHL patients, there is a strong genotype-phenotype correlation in VHL
disease. The location of the mutation determines the type of
tumor that arises (Kaelin, 2002) or if a predisposition to tumor
formation exists at all. Individuals with a 598C→T mutation
in exon 3 of VHL are not predisposed to tumor formation of
any type; instead, these individuals have a disease known as
Chuvash polycythemia (Hickey et al., 2007). Importantly, individuals with Chuvash polycythemia also demonstrate life-long
elevations of HIF, with increased levels of VEGF and erythropoietin without evidence of increased risk of tumor formation,
further suggesting a decreased sensitivity in people compared
with mice for this tumor type.
Multiple other compounds that produce hemolysis also
produce hemangiosarcomas in mice (2-butoxyethanol [BE],
phenylhydrazine, aniline). Hemolysis produces tissue hypoxia through active destruction of erythrocytes and results
in decreased availability of red blood cells to carry adequate
oxygen to tissues. Hemolysis is also classically associated with
upregulation of bone marrow erythroid production and splenic
increases in erythrocytes in rodents. Additionally, an inflammatory component accompanies hemolytic processes due to the
release of cytotoxic products of hemoglobin degradation when
it occurs within the vascular system. Therefore, these hemolytic compounds, as per the data from VHL gene mutations and
pregabalin, support the specificity of hypoxia to induce hemangiosarcoma in mice.
HUMAN RELEVANCE OF PREGABALIN-INDUCED HEMANGIOSARCOMA
63
Temporality (Exposure Precedes Outcome). Pregabalin
produces increases in bicarbonate on day 1, and local tissue
hypoxia and macrophage increases and bone marrow changes
are seen as early as 2 weeks (Criswell et al., ­unpublished data).
Bone marrow changes are progressive with time. All of these
changes occur well before the first occurrence of hemangiosarcoma. Endothelial cell proliferation was increased as early as 2
weeks in the liver and 12 weeks in the bone marrow (Criswell
et al., unpublished data). Hemangiosarcomas were not evident
until after 12 months, and most ­occurred after 18 months or
longer of pregabalin treatment.
(Cunningham, 2002; Siesky et al., 2002). A vitamin E intervention study blocked hepatic pregabalin-induced endothelial cell
proliferation in mice (Criswell et al., unpublished data). These
data are consistent with the proposed MOA and also with the
species specificity as vitamin E levels equivalent to those found
in rats were sufficient to block endothelial cell proliferation in
mice. Humans have even higher vitamin E levels than rats or
mice, and so they have greater antioxidant protective reserve
and presumably would be at lower risk for induction of hemangiosarcoma.
Plausibility and Coherence (Association Should Be Compatible With Existing Theory). Hemangiosarcomas are vascular
tumors. Physiologic angiogenesis is well characterized, and
the hypothesis that dysregulated angiogenesis is responsible
for induction of ­hemangiosarcoma is a logical extension based
on disruption of a normal physiologic process. Pregabalin acts
through a nongenotoxic mechanism to produce hemangiosarcomas in mice. Increased proliferation of the target cell type
or hyperplasia is an accepted hallmark of nongenotoxic tumorigenesis. It should be anticipated that increases in endothelial
cell tumors would be driven by biological drivers that increase
proliferation rate of endothelial cells. Disruption of normal
physiologic processes is well accepted for rodent thyroid tumors and Leydig cell tumors, which are driven by the growth
factors specific for these cell types. For example, elevated
thyroid stimulating hormone induces hyperplasia and thyroid
tumors in rats, and increased luteinizing hormone concentrations result in stimulation of Leydig cells, which also produce
an increased incidence of hyperplasia and adenomas in rats. In
both cases, rodents have been characterized as containing endocrine tissue that is “highly sensitive” and not characteristic of
the response in humans (Capen, 2000). Hypoxia is a commonly
accepted driver for angiogenesis and endothelial cell proliferation. Hypoxia is known to stimulate angiogenesis, and sustained
life-long hypoxia is known to induce hemangiosarcoma in mice
as evidenced by mice with the VHL mutation (Kleymenova et
al., 2004). Furthermore, tissue hypoxia has been proposed as
the key event in the formation of mouse-specific hemangiosarcomas with a diversity of agents including peroxisome proliferator activated receptor (PPARγ), 2-BE (a hemolytic industrial
solvent), and pregabalin (Cohen et al., 2009), demonstrating
a common mechanism among diverse compounds. Therefore,
dysregulated angiogenesis as a response to a sustained hypoxia
signal is biologically plausible for hemangiosarcoma formation
in mice.
Question 2: Are the Key Events in the Mouse MOA Plausible
in Humans?
The basic premise is that nongenotoxic compounds induce
hemangiosarcoma by dysregulated angiogenesis, which is
to say, dysregulation of physiologic angiogenesis. Although
physiologic angiogenesis is similar across species and therefore plausible in humans, there are several reasons why this is
unlikely to occur in humans.
Treatment of mice for 2 years with pregabalin induces an
increased incidence of hemangiosarcoma, but the latency and
tissue distribution of the tumor are not altered from those that
arise spontaneously (Pegg et al., unpublished data). Thus, in
mice, pregabalin appears to exacerbate extant processes associated with the spontaneous occurrence of hemangiosarcoma.
Increases in serum bicarbonate and tissue hypoxia are events
that could occur in any species. Therefore, a simple assumption
might be that the proposed MOA may be operational in other
species. However, there is strong evidence that this is not the
case. The most convincing evidence is that although rats also
demonstrate the same increase in bicarbonate as mice, pregabalin does not induce hemangiosarcomas in rats (Pegg et al., unpublished data). Although there are many biological processes
that can result in tissue hypoxia, there are an equally diverse set
of compensatory processes that typically prevent the development of long-term hypoxia. The ability of pregabalin to sustain
respiratory depression over the lifetime of the mouse (assessed
up to 22 months) with an associated metabolic alkalosis without adequate compensatory mechanisms is highly unique and
has been identified only in the mouse (Criswell et al., unpublished data). In contrast, rats exposed to pregabalin demonstrate
appropriate compensatory processes.
Evaluation of species-specific responses helps to frame
the relevance of any toxicological finding to human health.
Similar duration studies were conducted in the mouse and rat
to compare the toxicologic response in a species sensitive
to hemangiosarcoma formation with long-term pregabalin
administration and in a species where there is no evidence of this
tumor (Criswell et al., unpublished data). Table 1 summarizes the
species-specific changes in key elements of the MOA, including
studies conducted in monkeys for up to 1 year. The key difference
between mice and rats is that mice respond inadequately to an
acid-base imbalance with subsequent development of chronic
Experiment (Condition Can Be Prevented or Ameliorated
by an ­Appropriate Experimental Regimen). Vitamin E is a
fat-soluble antioxidant that has been shown to inhibit angiogenic growth factors (Woodson et al., 2002) and inflammatory
cytokines (Huey et al., 2008). Additionally, it has been shown
with 2-BE that the level of oxidative injury and associated cell
proliferation in liver is inversely related to vitamin E levels
64
CRISWELL ET AL.
TABLE 1
Species Specificity of Pregabalin-Induced Changes in Parameters
Associated With Hemangiosarcoma in Mice
Parameter
Increased bicarbonate
Decreased respiration
Acid-base imbalance
Bone marrow congestion
Macrophage/erythrophage
proliferation
Extramedullary hematopoiesis
Increased platelets and/or
megakaryocytes
Increased platelet activation
Altered platelet morphology
Altered platelet aggregation
Increased circulating PDGF
Increased VEGF (bone marrow
and spleen)
Increased bFGF (bone marrow
and spleen)
Increased VEGFR2 (liver)
Chronic tissue hypoxia
Increased endothelial cell
proliferation
Increased incidence of
hemangiosarcoma
B6C3F1 mice Wistar rat Monkey Human
+
+
+
+
+
+
+
–
–
–
–
–
–
–
–
–
–
N.T.
N.T.
N.T.
+
+
–
–
–
–
N.T.
–
+
+
+
+
+
–
–
–
–
–
–
–
–
N.T.
N.T.
–
–
–
N.T.
N.T.
+
–
N.T.
N.T.
+
+
+
–
–
–
N.T.
N.T.
N.T.
N.T.
N.T.
–a
+
–
–
–
Note. + = effect; – = no effect.
N.T., not tested; PDGF, platelet-derived growth factor; VEGFR2, vascular
endothelial growth factor receptor 2.
a
As assessed by negative effects on thrombomodulin.
tissue hypoxia. These differences can be summarized as follows:
(1) Rats respond appropriately to acute metabolic alkalosis. They
effectively decrease respiration rate. In contrast to mice, rats
normalize minute volumes; (2) Rats also show the anticipated
compensatory response to increased serum bicarbonate, which
returns the HCO3:pCO2 ratios to control levels. This results in
the expected acid balance and no impact on blood pH; (3) Rats
respond to acute changes with a peripheral erythrocytosis such
that circulating erythrocyte counts are nearly 20% higher than
those in untreated control rats by the end of the 2-year bioassay.
The presence of increased erythrocytes provides another
appropriate and expected compensatory response to provide
additional oxygenation to tissues; (4) Rats did not demonstrate the
key components of the proposed MOA (i.e., hypoxia, increased
endothelial cell proliferation, dysregulated erythropoiesis). There
was no evidence of tissue hypoxia in liver when assessed with
an immunohistochemistry stain for hypoxia (Hypoxyprobe,
NPI, Burlington, MA) or a transcriptomic/pathway analysis.
There was no evidence of increases in macrophages, angiogenic
growth factors, the production of a reactive, proliferative bone
marrow, and most importantly, no evidence of endothelial cell
proliferation; and (5) Rats have higher levels of vitamin E than
mice, which may provide an additional protective mechanism
against hemangiosarcoma formation likely via multiple
protective effects on angiogenesis. Importantly, several of these
endpoints provide potential biomarkers that can be used to assess
risk in humans.
Scientific contributions to the literature surrounding hemangiosarcomas in mice have also added to the ability to evaluate
human relevance. For the first time, tissue hypoxia and, perhaps
more importantly, how mice uniquely respond to that metabolic
alkalosis leading to hypoxia, have been proposed as the key
event in the formation of mouse-specific hemangiosarcomas
with a diversity of agents including PPARγ agonists, 2-BE (a
hemolytic industrial solvent), and pregabalin (Cohen et al.,
2009). A recent systems biology study of 2-BE demonstrates
that hypoxia is a contributing factor for hemangiosarcoma
formation in mice by this industrial solvent (Laifenfeld et al.,
2010).
Hypoxia is a well-accepted signal for increased endothelial
cell proliferation and normal angiogenesis. Further suggestive
evidence of the lack of relevance of hypoxia as a sole inducing
effect of hemangiosarcoma in humans is that there is no epidemiological evidence that people living at high altitudes have a
greater incidence of hemangiosarcoma, whereas Mori-Chavez
et al. (1970) report that exposure of mice to high altitude does
increase the incidence of spontaneous ovarian angiomas.
Many of the compounds reported to produce hemangiosarcomas in mice by nongenotoxic mechanisms are hemolytic in
nature (Cohen et al., 2009). Chronic hemolysis would drive hematopoietic compensatory mechanisms increasing proliferative
activity of bone marrow and spleen and accumulating iron—attributes that have also been identified with pregabalin administration even though pregabalin does not produce ­hemolysis.
Recent advances in the understanding of tumor angiogenesis
highlight the important role of circulating, bone marrow–­
derived progenitor cells, including both endothelial and monocyte/macrophage lineage cells, in pathological angiogenesis
(Ding et al., 2008; Lamerato-Kozicki et al., 2006). These findings suggest that hemangiosarcomas may arise not only from
transformation of tissue-resident endothelial cell populations
but also from circulating progenitors or adult stem cells recruited from bone marrow or possibly also from extramedullary sites such as the spleen. The marked differences in bone
marrow and spleen of pregabalin-treated mice compared with
bone marrow of rats and monkeys further highlight the speciesspecific components of the proposed MOA. There are no reports of increased hemangiosarcomas in humans with chronic
hemolytic disease or bone marrow proliferative ­disorders.
Question 3: Taking into Account Kinetic and Dynamic
Factors, Is the Animal MOA Plausible in Humans?
This phase of the ILSI-RSI HRF decision-making process requires a quantitative and kinetic analysis to determine
whether the MOA observed in animals could plausibly occur in
humans. Species-specific susceptibility to spontaneous occurrence of hemangiosarcoma constitutes a major element in determining human relevance. Hemangiosarcoma is a relatively
common tumor in mice, with a spontaneous incidence as high
HUMAN RELEVANCE OF PREGABALIN-INDUCED HEMANGIOSARCOMA
as 10–12% in some cases (Chandra and Frith, 1992a,b; Eiben,
2001; Giknis and Clifford, 2005; Haseman et al., 1988, 1999;
Tamano et al., 1988; Ward et al., 1979). It follows that mice are
highly susceptible to endogenous factors perturbing normal homeostatic mechanisms and inducing target endothelial cell proliferation. In contrast, the spontaneous incidence of hemangiosarcoma in humans has recently been estimated at 0.00021%,
indicating that humans are substantially less susceptible to
spontaneous hemangiosarcoma development than mice (U.S.
National Cancer Institute SEER Database).
Recent publications provide further evidence that the mouse
is more susceptible to the induction of hemangiosarcoma than
humans:
• Spontaneous cell proliferation rates (labeling index) of hepatic endothelial cells in mice (0.84–1.14), rats (0.28–0.36),
and humans (0.20–0.25) correlate with the spontaneous incidence of hemangiosarcoma (Ohnishi et al., 2007).
• Troglitazone, a compound that produces hemangiosarcomas
in mice but not in rats, has been shown to stimulate endothelial cell proliferation in mouse, but not human microvascular
endothelial cells (Kakiuchi-Kiyota et al., 2009).
• Mice have lower antioxidant levels (e.g., vitamin E) than
rats (2.5-fold higher than mice) and humans (100-fold higher), which may contribute to enhanced susceptibility to developing hemangiosarcoma in mice but not in rats treated
with 2-BE (Siesky et al.,2002). By supplementing the diet of
mice with vitamin E to levels typically found in rats, pregabalin-induced increases in endothelial cell proliferation were
inhibited.
• Humans who have mutations in the VHL do not develop
hemangiosarcoma, yet mice do develop tumors (Gordeuk
et al., 2004; Kleymenova et al., 2004).
• There are no reports of increased hemangiosarcomas in humans with chronic hemolytic disease or bone marrow proliferative disorders.
• Further suggestive evidence of the lack of relevance of hypoxia as a sole inducing effect of hemangiosarcoma in humans is that there is no epidemiological evidence that people
living at high altitudes have a greater incidence of hemangiosarcoma, whereas Mori-Chavez et al. (1970) report that
exposure of mice to high altitude does increase the incidence
of spontaneous ovarian angiomas.
Based on the above evidence, humans do appear to be less
susceptible to hemangiosarcoma induction than mice. Beyond
species-specific differences in the spontaneous rate of hemangiosarcomas, carcinogenicity studies in rats and mice achieved
similar exposure levels, yet hemangiosarcomas were found only
in mice. The negative finding in rats was confirmed by a second
negative 2-year carcinogenicity study in Wistar rats. Although
most of the mechanistic experiments in rodents were conducted
at doses resulting in exposures greater than those achieved in
humans, effects on platelet number, platelet function, platelet
morphology, and bone marrow were seen in mice at 200 mg/kg
65
with exposures equivalent to a human dose of 600 mg. Therefore, exposures comparable to those in mice at which platelet
and bone marrow effects were elicited have been achieved in
humans but without similar effects on platelets, which further
supports lack of human evidence for the mechanism.
To test the human relevance, several key biomarkers identified in the mouse MOA experiments were assessed in humans.
Use of these biomarkers allows the assessment of the question
remaining, namely, whether results in humans are more similar to results in mice or rats. This evaluation was conducted
through quantitative evaluation of key components of the MOA
in humans in clinical trials measuring several biomarkers.
Assessment of Key Elements of the MOA in Humans Compared With the Effects Seen in Mice. Potential effects of pregabalin on key elements of the MOA established in mice were
investigated in a human clinical t­rial. The clinical trial study
design is described in the Supplemental Section. It was not
ethically possible to duplicate endothelial cell proliferation and
hypoxia assays in humans as current methods would have required obtaining liver sections. Therefore, selective biomarkers were evaluated in a clinical trial to assess whether pregabalin treatment at the maximum therapeutic dose altered key
elements of the MOA established in mice. Bicarbonate was
measured because this is hypothesized to be the initial event in
the mouse for the induction of hemangiosarcoma. White blood
cell (WBC) counts, platelet numbers, and platelet function were
measured as biomarkers of dysregulated erythropoiesis. The
data supporting platelets as a biomarker have been previously
described (Criswell et al., unpublished data). Because increased
endothelial cell proliferation is paramount in the production of
hemangiosarcomas by nongenotoxic mechanisms and direct assessment of endothelial cell proliferation in humans was not
possible, thrombomodulin and the generation of Annexin Vlabeled circulating endothelial microparticles, markers of endothelial cell activation or injury, were evaluated as alternate
biomarkers of endothelial cell proliferation. Because endothelial cell activation is a prerequisite step for endothelial cell proliferation to occur in normal angiogenesis, it was considered a
relevant biomarker for endothelial cell proliferation.
Bicarbonate. Serum bicarbonate values were assessed in
humans receiving 600 mg/day pregabalin in the 1-month study
at 24 h, 15 days, and 29 days. There was no change in bicarbonate in pregabalin-treated subjects compared with controls
(Table 2; Fig. 5). This demonstrates that at maximum recommended dose, pregabalin does not increase bicarbonate as it
does in mice and rats.
Respiratory depression in humans of a magnitude comparable to that in mice would most certainly have been evident
clinically in volunteers. Additionally, although blood pH was
not measured in the clinical study of platelet function, there
was no evidence that venous bicarbonate levels were affected
by pregabalin treatment in humans. Therefore, humans are either not susceptible to the elevations of bicarbonate at a daily
66
CRISWELL ET AL.
TABLE 2
Bicarbonate Concentrations in Plasma in Healthy Volunteers
Receiving Placebo or Pregabalin
Bicarbonate concentration (mmol/L ± S.D.)
Dose group
Placebo
Screening
Day 1 –
Day 15
Day 29
Predose
28.1 ± 1.91 24.8 ± 2.05 25.4 ± 2.37 24.3 ± 1.97
Pregabalin 27.5 ± 1.92 24.9 ± 2.11 25.5 ± 1.85 25.2 ± 1.55
(300 mg/
BID)
Close-out
27.8 ± 1.74
27.8 ± 2.33
Note. n = 18–22 subjects per time point.
maximum dose of 600 mg or, in contrast to mice, are able to
compensate adequately. This demonstrates that the first identifiable event in the mouse MOA, elevation in serum bicarbonate
levels, with a resulting metabolic alkalosis does not occur in
humans.
WBC and Platelets. There were no effects on WBC or
platelet count at day 15 or 29 in this clinical trial (results not
shown). In mice, pregabalin treatment is associated with increased WBC (approximately twofold), platelet count (approximately 20–30%), changes in morphology (giant platelets,
agranular platelets, aggregates), and increase in platelet activation based on P-selectin expression (Criswell et al., unpublished data).
Although direct examination of bone marrow has not been
done in humans, lack of peripheral changes in humans and lack
of bone marrow changes in rats (up to 2 years) and monkeys
(up to 69 weeks) supports a species-specific difference in a key
FIG. 5. HCO3 concentrations assessed on pretest, day 15, and day 29 in
subjects receiving placebo (blue circles) or 600 mg/kg pregabalin/day (red
triangles). There was no difference in HCO3 between placebo and pregabalintreated subjects. Closed symbols represent arithmetic means.
component of the proposed MOA. Therefore, the absence of
effects on WBC and platelet count, morphology, and function
demonstrates that dysregulated erythropoiesis is not occurring
in humans treated for 1 month with pregabalin.
Platelet Function Assays. There were no statistically significant (p < 0.05) differences between placebo and pregabalintreated groups on days 15 or 29 for the primary (nonparametric)
analysis of P-selectin (a marker for platelet activation; Fig. 6).
Pregabalin had no effect on platelet or erythrocyte morphology
as visualized by light microscopy. Furthermore, in patients given
pregabalin at up to 600 mg/day for up to 3 years, there were no
clinically significant changes in platelet morphology, considered a surrogate marker of platelet activation (data not shown).
For comparison, platelet activation was increased 46% within
1 month at the tumorigenic dose of 1000 mg/kg in mice, and
platelet counts were increased from 32 to 58% above the control range after 2 years of treatment in the carcinogenicity studies
with mice.
Pregabalin had no effect on maximum platelet aggregation
(Fig. 7a) and no effect on ADP threshold aggregation (Fig. 7b).
Similarly, pregabalin had no clinically relevant effect on platelet
function analyzer (PFA)-100 closure time stimulated by either
ADP or epinephrine (Figs. 8a and b). The results in humans differ markedly from what was observed in mice, where there was a
28–46% decrease in maximum aggregation in pregabalin-treated animals, consistent with activation and depletion of dense
granules. Thus, in contrast to the marked effects displayed in
mice, platelet changes are not observed in humans administered
with the maximum recommended clinical dose of pregabalin.
Thrombomodulin and Annexin V Microparticles. Activation of endothelial cells is a prerequisite step that precedes endothelial cell proliferation in normal angiogenesis
(Fig. 2). If there is no evidence of endothelial activation,
FIG. 6. Mean platelet activation assessed by flow cytometric evaluation
of platelet P-selectin on pretest, day 15, and day 29 in subjects receiving placebo (blue circle) or 600 mg/kg pregabalin/day (red triangle). There was no
evidence of platelet activation in subjects receiving pregabalin. The dotted line
at 2% represents the investigator-defined limit for subject inclusion in the study.
HUMAN RELEVANCE OF PREGABALIN-INDUCED HEMANGIOSARCOMA
67
FIG. 7. Platelet aggregation assessed with a platelet aggregometer in human subjects. (a) Mean maximum percentage platelet aggregation in response to
300 µM ADP and (b) mean threshold concentration required to produce 75% aggregation of platelets was unchanged at days 15 and 29 in subjects receiving
600 mg/kg pregabalin (red triangles) compared with placebo (blue circles). Data offset for clarity. Dotted lines represent laboratory normal ranges.
p­ rogression to endothelial cell proliferation would be unlikely. Therefore, because endothelial cell proliferation could not
feasibly be assessed in humans, thrombomodulin and Annexin
V-labeled microparticles were assessed as clinically relevant
markers of direct endothelial cell damage and/or activation.
Thrombomodulin is an endothelial cell transmembrane
glycoprotein that plays an important role as an anticoagulant
and is predominantly expressed on vascular endothelial cells
of arteries, veins, and capillaries (Maruyama et al., 1985).
Small amounts of thrombomodulin may also be detected
on platelets, monocytes, and neutrophils (Fink et al., 1993).
Soluble thrombomodulin is found in serum or urine and is
probably released during endothelial cell damage or activation.
Soluble thrombomodulin is considered to be a specific marker
of endothelial cell injury (Ishii et al., 1991), and it has been
used widely as a marker of microvascular endothelial injury
and/or activation in various clinical disease states, including
disseminated intravascular coagulation (Endo et al., 1995),
multiple sclerosis (Tsukada et al., 1995), rheumatic diseases
(Ohdama, et al., 1994), systemic lupus erythematosus, lupus
nephritis (Frijns et al., 2001), renal transplant (Keven et al.,
2010), congestive heart failure (Chong et al., 2009),
preeclampsia, and uteroplacental insufficiency associated
with intrapartum hypoxia (Moussa, 2010; http://www.
obgyn.net/pregnancy-birth/pregnancy-birth.asp?page=/pb/
articles/moussa-preeclampsia). Therefore, assessment of
thrombomodulin in subjects receiving pregabalin was utilized
to assess direct endothelial cell injury and/or activation.
FIG. 8. PFA-100 closure times in human subjects. The mean time required for hemodynamic closure when stimulated with (a) ADP or (b) epinephrine was
similar in subjects receiving 600 mg/kg pregabalin (red triangles) for 15 or 29 days compared with placebo (blue circles). Data offset for clarity. Dotted lines
represent laboratory normal ranges.
68
CRISWELL ET AL.
Microparticles are fragments of membrane that are shed
during apoptosis or activation of several cell types including erythrocytes, platelets, and endothelial cells (Martinez
et al., 2005; Tesse et al., 2006). Increased levels of circulating ­microparticles serve as a marker for endothelial activation,
endothelial dysfunction, or platelet activation (Martin et al.,
1995). Annexin V is commonly used in flow cytometry as an
early marker of apoptosis. Annexin V has the ability to bind to
phosphatidylserine, which is exposed on the cell surface early in the process of apoptosis or cellular activation (Martin et
al., 1995; Reutelingsperger and van Heerde, 1997). Increased
­levels of circulating endothelial Annexin V positive microparticles have been used as a valid predictor of endothelial effects in
a broad variety of vascular disease states including atherosclerosis, renal failure (Amabile, et al., 2005; Shehata, et al., 2010),
pulmonary hypertension (Amabile et al., 2008), and metabolic
syndrome (Agouni et al., 2008). As such they serve as another
valid marker of possible drug-induced effects on endothelial
cells and vascular ­function.
Pregabalin had no effect on endothelial cell activation
as measured by soluble thrombomodulin. There were no
statistically significant differences in thrombomodulin levels in
humans receiving pregabalin from placebo observed on either
days 15 or 29 (Fig. 9). Additionally, there were no differences
in Annexin V labeling (Fig. 10a) or the number of endothelial
microparticles (Fig. 10b) on either day 15 or 29, providing
further evidence that pregabalin did not activate or directly
injure endothelial cells in humans. Increases in endothelial cell
FIG. 9. Comparison of mean serum soluble thrombomodulin assessed
on pretest, day 15, and day 29 in subjects receiving placebo (blue circles) or
600 mg/kg pregabalin/day (red triangles). Pregabalin had no effect on soluble
thrombomodulin: least-squares means and 95% confidence intervals from the
parametric analysis. Data are offset for clarity. Dotted lines represent mean 
standard deviation reported in healthy males aged 21–55 years.
proliferation were observed in mice as early as 2 weeks in the
liver (Criswell et al., unpublished data).
In summary, there was no evidence that pregabalin-treated
patients or volunteers had changes in bicarbonate (the first
event in the mouse for hemangiosarcoma formation), dysregulated erythropoiesis (changes in WBC or platelet function), or
FIG. 10. (a) Comparison of Annexin V binding assessed on pretest, day 15, and day 29 in subjects receiving placebo or 600 mg/kg pregabalin/day. Pregabalin
treatment in humans had no effect on Annexin V binding. Least-squares means with 95% confidence intervals from the parametric analysis. Triangles represent
pregabalin-treated subjects and circles represent placebo; data are offset for clarity. (b) Comparison of endothelial microparticle formation assessed on pretest, day
15, and day 29 in subjects receiving placebo or 600 mg/kg pregabalin/day. Pregabalin treatment in humans had no effect on microparticle formation. Least-squares
means with 95% confidence intervals from the parametric analysis (lower panel). Triangles represent pregabalin-treated subjects and circles represent placebo;
data are offset for clarity.
HUMAN RELEVANCE OF PREGABALIN-INDUCED HEMANGIOSARCOMA
activated endothelial cells (thrombomodulin, Annexin V, or
microparticles), which are surrogates for endothelial cell proliferation. These data support the conclusion that humans do
not exhibit the specific responses to pregabalin treatment that
were observed in the mouse and hence do not appear to be at
increased risk for hemangiosarcoma at the maximum daily dosage of pregabalin.
CONCLUSION
This paper assessed the human relevance of pregabalininduced hemangiosarcoma in mice using the IPCS/ILSI-Health
and Environmental Sciences Institute HESI HRF. The first step
of this process is developing a data-driven MOA in animals and
assessing the weight of evidence for the MOA using the Hill
criteria. The proposed MOA for pregabalin is consistent with
four other compounds previously reported as a unified MOA
frame work proposed for induction of hemangiosarcoma in mice
by nongenotoxic compounds (Cohen et al., 2009). Those key
elements include hypoxia, macrophage activation, increased
angiogenic growth factors, dysregulated angiogenesis/
erythropoiesis, and increased endothelial cell proliferation.
The second step in the IPCS/ILSI-HESI HRF is assessing
whether the key events in the animal MOA are plausible in
humans. Because hemangiosarcoma is due to disruption of a
physiologic process (i.e., angiogenesis), which is shared across
all species, this MOA is theoretically plausible for humans.
Therefore, one must address the third question of the HRF.
The third step in the IPCS/ILSI-HESI HRF, taking into account kinetic and dynamic factors, asks if the animal MOA
is plausible in humans. This was assessed by comparing susceptibility across species, disease states, and measuring key
biomarkers in a clinical trial with pregabalin. Several lines of
evidence were presented that supported the conclusion that
the mouse is more susceptible to the induction of hemangiosarcoma than humans. Additionally, key biomarkers identified from the mouse studies were assessed in patients treated
with the maximum recommended human dose of pregabalin.
There was no evidence that pregabalin-treated patients or
volunteers had changes in bicarbonate (temporally the first
change identified in mice treated with pregabalin and likely
responsible for elevation in blood pH in mice) or biomarkers for dysregulated erythropoiesis or endothelial cell proliferation. Collectively, these data support the conclusion that
humans do not exhibit the specific responses to pregabalin
treatment that were observed in the mouse and hence are not
at increased risk for hemangiosarcoma at the maximum daily
dose of pregabalin.
FUNDING
All studies were funded by Pfizer Drug Safety Research &
Development. There were no external sources of funding.
69
REFERENCES
Agouni, A., Lagrue-Lak-Hal, A. H., Ducluzeau, P. H., Mostefai, H. A.,
Draunet-Busson, C., Leftheriotis, G., Heymes, C., Martinez, M. C., and
­Andriantsitohaina, R. (2008). Endothelial dysfunction caused by circulating
microparticles from patients with metabolic syndrome. Am. J. Pathol. 173,
1210–1219.
Amabile, N., Guerin, A. P., Leroyer, A., Mattat, Z., Nguyen, C., Boddaert,
J., London, G. M., Tedgui, A., and Boulanger, C. M. (2005). Circulating endothelial microparticles are associated with vascular dysfunction
in patients with end-stage renal failure. J. Am. Soc. Nephrol. 16, 3381–
3388.
Amabile, N., Heiss, Ch. Real, W. M., Minasis, P., McGlothlin, D., Rame, E.
J., Grossman, W., De Marco, T., and Yeghiazarians, Y. (2008). Circulating
endothelial microparticle levels predict hemodynamic severity of pulmonary
hypertension. Am. J. Respir. Crit. Care Med. 177, 1268–1275.
Angiogenesis Foundation. Available at: http://www.angio.org/understanding/
fact.php. Accessed October 31, 2011.
Boivin-Angèle, S., Lefrançois, L., Froment, O., Spiethoff, A., Bogdanffy, M. S.,
Wegener, K., Wesch, H., Barbin, A., Bancel, B., Trépo, C., et al.
(2000). Ras gene mutations in vinyl chloride-induced liver tumours are
carcinogen-specific but vary with cell type and species. Int. J. Cancer 85,
223–227.
Capen, C. C. (2000) Toxic responses of the endocrine system. In Casarett &
Doull’s Toxicology: The Basic Science of Poisons. (C. D. Klaassen, Ed.) pp.
711–759. McGraw-Hill Publishing Company, NY.
Chandra, M., and Frith, C. H. (1992a). Spontaneous neoplasms in B6C3F1
mice. Toxicol. Letters 60, 91–8.
Chandra, M., and Frith, C. H. (1992b). Spontaneous neoplasms in aged CD-1
mice. Toxicol. Letters 61, 67–74.
Chong, A., Freestone, B., Patel, J., Lim, H., Hughes, E., Blann, A., and Lip, G.
(2009). Endothelial activation, dysfunction and damage in congestive heart
failure and the relation to brain natriuretic peptide outcomes. Am. J. Cardiol.
97, 671–675.
Cohen, S. M., Storer, R. D., Criswell, K. A., Doerrer, N. G., Dellarco, V. L.,
Pegg D. G., Wojcinski, Z. W., Malarkey, D. E., Jacobs A. C., Klaunig, J. E.,
et al. (2009). Hemangiosarcoma in rodents: Mode-of-action evaluation and
human relevance. Toxicol. Sci. 111, 4–18.
Cunningham, M. L. (2002). A mouse is not a rat is not a human: Species differences exist. Toxicol. Sci. 70, 157–158.
Ding, Y., Kumar, S., and Yu, D. (2008). The role of endothelial progenitor cells
in tumour vasculogenesis. Pathobiology 75, 265–273.
Eiben, R. (2001). Frequency and time trends of spontaneous tumors found in
B6C3F1 mice oncogenicity studies over 10 years. Exp. Toxic. Pathol. 53,
399–408.
Endo, S., Inada, K., Nakae, H., Takakuwa, T., Kasai, T., Yamada, Y., Suzuki,
T., Taniguchi, S., and Yoshida, M. (1995). Blood levels of endothelin-1 and
thrombomodulin in patients with disseminated intravascular coagulation and
sepsis. Res. Commun. Mol. Pathol. Pharmacol. 90, 277–288.
Fink, L. M., Eidt, J. F., Johnson, K., Cook, J. M., Cook, C. D., Morser, J., Marlar, R., Collins, C. L., Schaefer, R., Xie, S. S., et al. (1993). Thrombomodulin
activity and localization. Int. J. Dev. Biol. 37, 221–226.
Fong, G. H. (2008). Mechanisms of adaptive angiogenesis to tissue hypoxia.
Angiogenesis 11, 121–140.
Fong, G. H. (2009). Regulation of angiogenesis by oxygen sensing mechanisms. J. Mol. Med. 87, 549–560.
Frijns, R., Fijnheer, R., Schiel, A., Donders, R., Sixma, J., and Derksen, R.
(2001). Persistent increase in plasma thrombomodulin in patients with a history of lupus nephritis: Endothelial cell activation markers. J. Rheumatol.
28, 514–9.
70
CRISWELL ET AL.
Giknis, M. L. A., and Clifford, C. B. (2005). Spontaneous neoplastic lesions in
the Crl:CD-1(ICR) mouse in control groups from 18 month to 2 year studies. Charles River Laboratories Research Models. Available at: http://criver.
com/research_models_and_services/research_models/RM_literature.html.
Accessed February 21, 2007.
Gordeuk, V. R., Sergueeva, A. I., Miasnikova, G. Y., Okhotin, D., Voloshin,
Y., Choyke, P. L., Butman, J. A., Jedlickova, K., Prchal, J. T., and Polyakova, L. A. (2004). Congenital disorder of oxygen sensing: Association of
the homozygous Chuvash polycythemia VHL mutation with thrombosis and
vascular abnormalities but not tumors. Blood 103, 3924–3932.
Haseman, J. K., Elwell, M. R., and Hailey, J. R. (1999). Neoplastic incidences
in B6C3F1 mice: NTP historical data. In Pathology of the Mouse (R. Maranpot, Ed.), pp. 679–689. Cache River Press, Vienna, IL.
Haseman, J. K., Hailey, J. R., and Morris, R. W. (1988). Spontaneous neoplasm
incidence in Fischer 344 rats and B6C3F1 mice in 2-year carcinogenicity
studies: A National Toxicology Program update. Toxicol. Pathol. 26, 428–441.
Hickey, M. M., Lam, J. C., Bezman, N. A., Rathmell, W.K., and Simon, M.C.
(2007). Von Hippel-Lindau mutation in mice recapitulates Chuvash polycythemia via hypoxia-inducible factor-2α signaling and splenic erythropoiesis. J. Clinical Invest. 117, 3879–3889.
Hill, A. B. (1965). The environment and disease: Association or causation?
Proc. Royal Soc. Med. 58, 295–300.
Huey, K. A., Fiscus, G., Richwine, A. F., Johnson R. W., and Meador B. M.
(2008). In vivo vitamin E administration attenuates interleukin-6 and interleukin-1B responses to an acute inflammatory insult in mouse skeletal and
cardiac muscle. Exp. Physiol. 93, 1263–1272.
Ishii, H., Uchijima E., and Kazama, M. (1991). Soluble thrombomodulin antigen in conditioned medium is increased by damage of endothelial cells.
Thromb. Haemost. 65, 618–623.
Kaelin, W. G., Jr. (2002). Molecular basis of the VHL hereditary cancer syndrome. Natl. Rev. Cancer 2, 673–682.
Kakiuchi-Kiyota, S., Vetro, J. A., Suzuki, S., Varney M. L., Han H. Y., Nascimento, M., Pennington, K. L., Arnold, L. L., Singh, R. K., and Cohen, S. M.
(2010). Effect of the PPARγ agonist troglitazone on endothelial cells in vivo
and in vitro: Differences between human and mouse. Toxicol. Appl. Pharmacol. 237, 83–90.
Keller, G. (2001). The hemangioblast. In Stem Cell Biology (R. L. Gardner, D.
R. Marshak and D. Gottlief, Eds.), pp. 329–348. Cold Spring Harbor, New
York, NY.
Keven, K., Elmaci, S., Sengul, S., Akar.N. , Egin, Y., Genc., V., Erturk, S., and
Erbay, B. (2010). Soluble endothelial cell protein C receptor and thrombomodulin levels after renal transplantation. Int. Urol. Nephrol. 42, 1093–1098.
Kleymenova, E., Everitt, J. L., Pluta, L., Portis, M., Gnarra, J. R., and Walker,
C. L. (2004). Susceptibility to vascular neoplasms but no increased susceptibility to renal carcinogenesis in Vhl knockout mice. Carcinogenesis 25,
309–315.
Laifenfeld, D., Gilchrist, A., Drubin, D., Jorge M., Eddy S. F., Frushour B.
P., Ladd B., Obert L. A., Gosink M. M., Cook J. C., et al. (2010). The role
of hypoxia in 2-butoxyethanol-induced hemangiosarcoma. Toxicol. Sci. 113,
254–266.
Lamerato-Kozicki, A. R., Helm, K. M., Jubala, C. M., Cutter, G. C., and Modiano, J. F. (2006). Canine hemangiosarcoma originates from hematopoietic
precursors with potential for endothelial differentiation. Exp. Hematol. 34,
870–878.
Lipshutz, G. S., Brennan, T. V., and Warren, R. S. (2002). Thorotrast-induced
liver neoplasia: A collective review. J. Am. Coll. Surg. 195, 713–718.
Martin, S. J., Reutelingsperger, C. P. M., McGahon, A. J., Rader, J. A., vanSchie,
R. C. A. A., LaFrace, D .M., and Green, D.R. (1995). Early redistribution
of plasma membrane phosphatidylserine is a general feature of apoptosis
regardless of initiating stimulus: Inhibition by overexpression of Bcl-2 and
Ab1. J. Exp. Med. 182, 1545–1556.
Martinez, M. C., Tesse, A., Zobairi, F., and Andriantsitohaina, R. (2005). Shed
membrane microparticles from circulating and vascular cells in regulating
vascular function. Am. J. Physiol. 228, H1004–H1009.
Maruyama, I., Bell, C. E., and Majerus, P. W. (1985). Thrombomodulin is
found on endothelium of arteries, veins, capillaries, and lymphatics, and on
syncytiotrophoblast on human placenta. J. Cell. Biol. 101, 363–371.
Meek, M., Bucher, J., Cohen, S., Dellarco, V., Hill, R., Lehman-McKeeman, L.,
Longfellow, D., Pastoor, T., Seed, J., and Patton, D. (2003). A framework for
human relevance analysis of information on carcinogenic modes of action.
Crit. Rev. Toxicol. 33, 591–653.
Mendenhall, W. M., Mendenhall, C. M., Werning, J. W., Reith, J. D., and Mendenhall, N.P. (2006). Cutaneous angiosarcoma. Am. J. Clin. Oncol. 29, 524–528.
Mori-Chavez, P., Upton, A.C., Salazar, M., and Conklin, J.W. (1970). Influence
of altitude on late effects of radiation in RF/Un mice: Observations on survival time, blood changes, body weight, and incidence of neoplasms. Cancer
Res. 30, 913–928.
Moussa, A.A. (2010). Thrombomodulin and antithrombin-III as evidence for
varied activation of the coagulation cascade in pre-eclampsia. Available at:
http://www.obgyn.net/pregnancy-birth/pregnancy-birth.asp?page=/pb/articles/moussa-preeclampsia.
Nordstrom-O’Brien, M., van der Luijt, R. B., van Rooijen, E., van den Ouweland, A. M., Majoor-Krakauer, D. F., Lolkema, M. P., van Brussel, A., Voest,
E. E., and Giles, R. H. (2010). Genetic analysis of von Hippel-Lindau disease.
Hum. Mutat. 31, 521–537.
Ohdama, S., Takano, S., Miyake, S., Kubota, T., Sato, K., and Aoki, N. (1994).
Plasma thrombomodulin as a marker of vascular injuries in collagen vascular
diseases. Am. J. Clin. Pathol. 101, 109–113.
Ohnishi, T., Arnold, L. L., Clark, N. M., Wisecarver, J. L., and Cohen, S. M.
(2007). Comparison of endothelial cell proliferation in normal liver and adipose
tissue in B6C3F1 mice, F344 rats, and humans. Toxicol. Pathol. 35, 904–909.
Pastoor, T., Rose, P., Lloyd, S., Peffer, R., and Green, T. (2005). Case study:
Weight of evidence evaluation of the human health relevance of thiamethoxamrelated mouse liver tumors. Toxicol. Sci. 86, 56–60.
Poteracki, J., and Walsh, K. M. (1998). Spontaneous neoplasms in control
Wistar rats: A comparison of reviews. Toxicol. Sci. 45, 1–8.
Pugh, C. W., and Ratcliffe, P. (2003). Regulation of angiogenesis by hypoxia:
Role of the HIF system. Nature Med. 9, 677–684.
Reutelingsperger, C. P. M., and van Heerde, W. L. (1997). Annexin V, the regulator of phosphatidylserine-catalyzed inflammation and coagulation during
apoptosis. Cell Mol. Life Sci. 53, 527–532.
Schäfer, M., Ewald, N., Schäfer, C., Stapler, A., Piper, H. M., and Noll, T.
(2003). Signaling of hypoxia-induced autonomous proliferation of endothelial cells. FASEB J. 17, 449–451.
Semenza, G.L. (2001). Regulation of hypoxia-induced angiogenesis: A chaperon escorts VEGF to the dance. J. Clin. Invest. 108, 39–40.
Shehata, M., El Abd, D., El Shanawani, F., Abdallah, E. A., Darwish, H.,
Moghazy, M. F., Metwaly, A., Ahmed, A., and Hadi A. (2010). Cadherin
5 and Annexin V as circulating endothelial microparticles: Markers for
artheroslcerotic vascular lesion in patients with chronic renal failure. K
­ idney
19, 307–315.
Siesky, A. M., Kamendulis, L. M., and Klaunig, J. E. (2002). Hepatic effects of
2-butoxyethanol in rodents. Toxicol. Sci. 70, 252–260.
Tamano, S., Hagiwara, A., Shibata, M. A., Kurata, Y., Fukushima, S., and
Ito, N. (1998). Spontaneous tumors in aging (C57BL/6N x C3H/HeN)F1
(B6C3F1) mice. Toxicol. Pathol. 16, 321–326.
Tesse, A., Martinez, M. C., Mexiani, F., Hugel, B., Panaro, M. A., Mitolo, V.,
Freyssinet, J. M., and Adnriantsitohaina, R. (2006). Origin and biological
significance of shed-membrane microparticles. Endocr. Metab. Immune Discord. Drug Targets 6, 287–294.
Tsukada, N., Matsuda, M., Miyagi, K., and Yanagisawa, N. (1995). Thrombomodulin in the sera of patients with multiple sclerosis and human
HUMAN RELEVANCE OF PREGABALIN-INDUCED HEMANGIOSARCOMA
lymphotropic virus type-1-associated myelopathy. J. Neuroimmunol. 56,
113–116.
U.S. National Cancer Institute’s Surveillance, Epidemiology, and End Results
(SEER) database. Available at: http://seer.cancer.gov/. Accessed February
21, 2007.
Walsh, K. M., and Poteracki, J. (1994). Spontaneous neoplasms in control
Wistar rats. Fund. Appl. Toxicol. 22, 65–72.
71
Ward, J. M., Goodman, D. G., Squire, R. A., Chu, K. C., and Linhart, M. S.
(1979). Neoplastic and nonneoplastic lesions in aging (C57BL/6N x C3H/
HeN)F1 (B6C3F1) mice. J. Natl. Cancer Inst. 63, 849–854.
Woodson, K., Triantos, S., Hartman, T., Taylor, P. R., Virtamo, J., and Albanes
D. (2002). Long-term alpha-tocopherol supplementation is associated with
lower serum vascular endothelial growth factors. Anticancer Res. 22, 375–
378.