Study to identify the genetic variations associated with phantom limb pain

BACKGROUND AND SIGNIFICANCE

Literature Review. Almost immediately after the loss of a limb, 90-95%
of all patients with major limb amputations experience a vivid phantom limb
sensation, such as warmth, cold, itching, pressure, and sense of position (1).
When the sensations become intense enough to define as painful, they are
referred to as phantom limb pain (PLP). PLP occurs in 80-90% of amputees and
usually appears immediately following awakening from anesthesia, but may be
delayed up to a few days or weeks in 25% of patients. Onset is not affected by
the limb amputated or site of amputation (2). In most cases the phantom is
present for a few days or weeks, then gradually fades from consciousness.
However, 30-70% of amputees have pain persisting for years or decades, with
pain persisting greater than six months becoming more difficult to treat (2,
3). The causes of PLP and non-painful phantom phenomena are not known; both
peripheral and central processes have been implicated (4). Some patients are
able to recall a phantom limb at will after its disappearance by intense
concentration or rubbing the residual limb (5). Memories of the limb*s posture
and form prior to amputation often survive in the phantom (6, 7). The study
principal investigator observed a patient injured by an improvised explosive
device (IED) who was holding his rifle prior to the injury. This patient
reported his amputated index finger stuck in the firing position on the
trigger, giving him cramping pain for six weeks afterwards. After a period of
several weeks, a patient*s phantom limb sometimes fades from consciousness and
disappears completely. Additionally, in approximately 50% of cases, and
especially in upper limb amputees, the arm progressively becomes shorter until
the subject is left with just the phantom hand alone, dangling from the stump,
a phenomenon that has been termed *telescoping* (8, 9). Ultimately, PLP is
remarkably difficult to treat, with several reports of failed drug trials in
the clinical literature (3, 4).
Human Subjects Justification. A more general problem in new analgesic
development is that many drugs that have appeared to relieve pain in rats have
failed in humans. Some researchers believe technical improvements in animal
models of pain will lead to a better understanding of the etiology of pain in
humans. However, there may be interspecies differences in the neurochemical
mechanisms of pain, and a more fundamental solution may require the integration
of human data into the early drug development triage process. Clinical genetic
association studies are one such approach. For example, a polymorphism that
decreases the function of a molecule was associated with reduced pain in
patients exposed to similar injuries. It might be reasonable to expect that a
drug inhibiting that molecule would prevent, or even relieve, pain in humans.

Genetic variations: We believe that identifying single nucleotide polymorphisms
(SNPs) between patients with phantom limb pain and those without may indicate
why human pain perception is highly variable from one individual to the next.
SNPs are DNA sequence variations that are usually of no consequence; however,
some SNPs may predispose individuals to certain diseases or alter the intended
effect of a particular drug. There is previous research that suggests a link
between phantom limb pain and differences in the genetic code. A group of
researchers led by Ze*ev Seltzer examined amputees in the Israeli army and
found several SNPs that appear to have a correlation with the presence or
absence of phantom pain (Seltzer, 2008 personal communication). Our proposed
study expands upon that pilot investigation and includes more thorough data
collection on the nature and characteristics of the subjects* phantom pain.

The Affymetrix Whole Genome Human SNP Microarray will be used to
identify all SNPs between the two study groups (patients with phantom limb pain
and patients who have little to no experience with phantom limb pain). This
study will further examine whether identifiable SNPs play a role in the
perception of painful phantom limb sensations.

Gene expression: Many underlying causes for neuropathic pain involve
changes in mRNA levels because of altered gene expression or transcript
stability (10, 11). RNA is pertinent to our genetic association study because
mRNA transcripts can vary widely according to changing genetic conditions (such
as the emergence of PLP), unlike the genome itself which is relatively fixed in
comparison. By extracting and quantifying RNA, we will create a gene expression
profile. Recently published articles have shown that there are specific gene
expression profiles in the peripheral blood of patients with various types of
neurological diseases, including dementia, Parkinson*s, and other behavioral
conditions (12-14).

BDNF analysis: The protein brain-derived neurotrophic factor (BDNF)
plays a key role in the modulation of pain nociception (15) and therefore may
be implicated in phantom limb pain. After peripheral nerve injury, BDNF
expression is dramatically increased in pain receptors of the brainstem. This
upregulation signals neurons that peripheral nerve damage has occurred (16-18).
Previous research has shown that continuous signaling between BDNF and its
neuronal receptor is necessary and sufficient to maintain the sensation of
neuropathic pain (16, 19). Blocking this pathway in mice inhibits the
development of allodynia, a condition in which a painful response occurs to a
normally innocuous stimulus (19). Conversely, inducing BDNF release from
microglia promotes the development of allodynia and contributes to long term
synaptic enhancement of the pain pathways (20). The reinforcement of chronic
pain by BDNF in the central nervous system suggests to us that BDNF may be
relevant to the presence of phantom limb pain. We expect that BDNF levels will
be markedly higher in subjects with phantom limb pain than in those without
phantom limb pain. By studying the levels of BDNF in the serum of both subject
groups, we hope to confirm this hypothesis and elucidate the specific role that
BDNF plays in phantom pain signaling.
PLP in non-Caucasians: There is no data at this time to suggest any differences
in the potential development of phantom limb pain based on gender or
race/ethnicity. However, ethnicity does affect genotype frequency and haplotype
pattern significantly. By comparing SNPs, gene expression and BDNF across
different ethnic populations, we hope to determine whether there is a
relationship between PLP associated genes and ethnicity. This comparison will
allow us to generalize our findings for non-Caucasian populations.

MILITARY RELEVANCE

Almost immediately after the loss of a limb, 90-95% of all patients experience
a vivid phantom. Case reports suggest that the incidence may be higher
following traumatic limb loss than after a planned surgical amputation of a
non-painful limb. The presence of phantom limb pain not only impairs recovery
by delaying rehabilitation but also causes continued discomfort and associated
depression. Over 800 amputees have entered the military care system since the
beginning of the Iraq conflict (personal communication, Mr. Giovani Ortega). By
current estimates, 85-90% will at some time experience PLP (21). Understanding
the genetic component of PLP may help in predicting which patients will
experience PLP and which amputees will respond to the various treatment options
available. Furthermore, the discovery of a genetic predisposition or link to
PLP may guide researchers in developing new treatments for PLP. Because of the
large number of amputee patients in the military healthcare system, optimizing
care and finding new, potentially more effective treatments is of great
military relevance.

The primary goals of this study are to:
i. Identify genetic variations responsible for PLP across the whole genome
More than 90% of SNPs are found in introns or intergenic regions and most of
them are probably of little functional consequence. However, SNPs in promoter
regions can alter the affinity of DNA binding proteins and modify the level of
gene expression. Other SNPs in exon / intron boundaries result in intron
retention or exon skipping, thus profoundly changing the structure of the
resulting protein. It is also possible that any genomic region where the SNP is
located can function as an RNA interference element. The existence of SNPs in
any portion of genomic DNA can be meaningful for the phenotype because of the
complicated dynamics of DNA structure and gene expression. Genome wide
association study along with multi-layer designed follow up study has been
applied to clinical pain research. The advantage of this approach is that no
specific functional genetic hypotheses are required prior to undertaking
analysis.
ii. Identify the gene expression profiles between individuals with and without
PLP using microarray technology.
It is suggested that analyzing gene expression profile will have clinical
applications to neurological diseases in humans. Recently published studies
have shown that there are specific gene expression profiles in the peripheral
blood of patients with various types of neurological diseases. The working
hypothesis is that differences in gene expression profiles after major limb
amputation produce a unique phenotype through molecular interactions.
Characteristics of PLP such as presence and severity of phantom pain may be
modulated by changes in gene expression over the time in the post-amputation
period. We will test the hypothesis that the gene expression profiles are
different between PLP and non-PLP amputee patients.

The secondary goal of this study is to:
Identify predictive factors for the prognosis of PLP among the interactions
between genetic variations, gene expressions, protein levels and other
physiologic variables from the integrative genomic-phenomic analyses.
The analyzed data from primary goals will contribute to the analysis for
secondary goal. This integrative genomic-phenomic analysis can generate
hypothesis for underlying pathways suggesting how gene expression changes in
the serum can be linked to neurological diseases such as PLP. Phenotypic data
will be determined from elements of the patient questionnaire used as potential
covariates to classify PLP. The tertiary goal of this study is to:
Verify candidate SNPs that are associated with PLP severity or duration in a
group of non-Caucasian subjects.
After identifying SNPs that are associated with PLP severity or duration in an
ethnically homogenous sample, we will determine whether these findings can be
generalized across different ethnic populations by comparing our genomic and
proteomic findings from Caucasian and non-Caucasian subjects. The DNA and RNA
analysis will be the same for both groups, and we will use principal component
analysis to determine if SNPs identified in the Caucasian group are similar for
non-Caucasian subjects. While previous research has suggested that frequency of
genomic markers vary among different ethnic groups, there has not been any
evidence to suggest an association between PLP and ethnicity. We therefore
hypothesize that there will be no difference between PLP associated genes in
the Caucasian group and the non-Caucasian group.

This observational, case-controlled, cross-sectional study will evaluate
whether genetic differences correlate with phantom limb pain. There will be no
blinding since both patient groups will need to report their level of PLP to
the investigators. This will not affect genetic analysis. There will not be
randomization, as the placement of subjects into groups will depend on degree
of PLP.

Subjects will be males or females at least 18 years of age who have sustained a
major limb amputation at least three months ago. PLP phenotype will be
determined at screening with the inclusion/exclusion criteria screening form.

The study sample will be comprised of one thousand (1200) amputees (950 with a
history of PLP; 250 with no history of PLP). Each participant*s one study visit
will consist of the following: (1) a questionnaire to assess the nature of PLP,
(2) a current medication listing, (3) a routine 30 mL blood draw from which
DNA, RNA, and BDNF will be harvested, and (4) a DEXA scan to account for the
effect of body fat composition on BDNF levels.

All data and blood samples will be de-identified. The de-identified patient
questionnaire, medication listing, and DEXA scan results will remain at WRNMMC,
while de-identified blood samples will be shipped to and analyzed by
collaborators at the NIH. Volunteers recruited outside of military treatment
facilities (MTFs) will not be listed on the master list since the only link to
their data from blood tubes and CRF*s would be the consent form. This is
discusse din more detail in sections 7 and 8.

Blood draws pose minimal risk to the subject, but participants undergoing a
DEXA may be exposed to a small amount of radiation, the effects of which are
too small to be determined. To minimize these risks, the blood draws and DEXA
scans will be performed by experienced practitioners in a controlled setting.

Actions to Minimize Risks

Confidentiality Protection: Each subject will be identified throughout this
study by coded identifiers (i.e. P1, P2, NP1, NP2) assigned by the lead site*s
research team. All data collection forms will be pre-labeled with the
subject*s code. Only the subject*s code (not the subject*s name) will appear on
data collection forms and blood samples.

For subjects recruited outside of the military clinical setting, no link will
be established between the subject*s study code and the subject*s personal
identifying information (i.e. name, age, etc). Signed consent forms will be
stored in a portable locked box for the duration of the off-site recruitment.
The locked box will be kept with a member of the study team during transport
across state lines. Only members of the local site study team will have access
to this box. At the local site, the signed consent forms will be stored in a
locked filing cabinet that only members of the local study team will have
access to.
For subjects recruited within the military clinical setting, a link between
the subject*s code and personal identity will be kept in a concealed
binder/folder in a locked filing cabinet housed in the an office supervised by
the local PI for three years. Only the local PI, not the lead site PI, will
have access to this linkage information. After three years, the link between
the subject*s name and the study will be destroyed in a manner that is
irreversible and that renders the data unreadable. For non-digital data,
disposal will involve shredding of all documents. All data reported in
publications will be aggregated, de-identified data.

Certificate of Confidentiality: N/A

Reporting Adverse Events

Please refer to performance site protocols for information about the reporting
of adverse events or unanticipated problems. Each performance site will adhere
to local guidelines and policies.