Genotyping Cytochrome P450 3A4*1B and 3A5"3 for 25 Fentanyl

Journal of Analytical Toxicology, Vol. 29, October 2005

Pharmacogenomicsas Molecular Autopsyfor Forensic Toxicology:Genotyping Cytochrome P450 3A4*1Band 3A5"3 for 25 Fentanyl Cases Ming Jin1,2, Susan B. Gock 1,2, Paul J. Jannetto1, Jeffrey M. Jentzen1,2, and Steven H. Wong 1,2,* IDepartmentof Pathology,Medical College of Wisconsin and 2MilwaukeeCountyMedical Examiner'sOffice, Milwaukee, Wisconsin

Abstract ] Pharmacogenomics, the study on genetic contributions to drug action may help in certifying fentanyl toxicity. Fentanyl is used clinically as an adjunct to surgical anesthesia and for chronic pain management. Its toxicity may be partially due to cytochrome P450 (CYP) 3A4" 1B and 3A5"3 variant alleles, resulting in variable fentanyl metabolism. In this study, we examined 25 fentanyl-related deaths (22 Caucasians, 1 African-American, and 2 NativeAmericans) from the Milwaukee County Medical Examiner's Office and referral cases. Fentanyl and norfentanyl in postmortem blood samples were analyzed by radioimmunoassayand liquid chromatography-mass spectrometry-mass spectrometry. The samples were then genotyped for CYP3A4*IB and 3A5"3 using Pyrosequencing TM. Genotyping showed: 1 CYP3A4*IB homozygous and CYP3A5*3 heterozygous, 1 compound CYP3A4*IB and CYP3A5*3 heterozygous, 22 CYP3A4*IB wild type and CYP3A5*3 homozygous, and 1CYP3A5*3 and CYP3A4*IB wild type. CYP variant allelic frequencies of the 25 cases were 6% for CYP3A4*IB and 92% for CYP3A5*3, compared with normal Caucasian CYP3A4*IB, 3-8%, and CYP3A5*3, 85-95%. The mean fentanyl

concentration and metabolic ratio of fentanyl to norfentanyl of the 2 cases with CYP3A4*IB and CYP3A5*3 variants were 12.8 and 1.4 pg/L, respectively, lower than those of 22 cases with wild type CYP3A4*IB and CYP3A5*3 homozygous variants, 16.7 and 7.3 pg/L, respectively. The postmortem/in vivo data provided the first scientific evidence that CYP3A5 is involved in the fentanyl metabolism, and homozygous CYP3A5 *3 causes impaired metabolism of fentanyl, and genotyping CYP3A4*IB and 3A5"3 variants may help to certify the fentanyl toxicity.

Introduction Pharmacogenomics is the study of genetic contributions to drug action. Most drugs are metabolized by cytochrome P450 (CYP) enzymes, and the variant of CYP gene may cause adverse * Author to whom correspondenceshould be addressed:Steven H. Wong, Ph.D., Department of Pathology, Medical College of Wisconsin, 870! W. Wate~'town Plank Road, Milwaukee, Wl 53226. E-mail: [email protected].

590

drug reactions (ADRs),which is one of the major leading causes of death in the United States (1). ADRs may be caused by poor metabolism as well as ultrarapid metabolism of drugs due to CYP gene variant. An opioid intoxication case was due to three or more functional alleles of CYP2D6 gene (2). Pharmacogenomics may play a role in the prevention of ADRs, and it may also help in the certification of drug toxicity cases in forensic pathology. Previous studies of pharmacogenomics in postmortem blood samples showed the CYP2D6 *3, *4, and *5 prevalence of poor and intermediate metabolizers in oxycodone or methadone-related deaths were higher than those of the control group. These data suggested that genotyping of CYP2D6 genes may be useful or serve as an adjunct to certify oxycodone or methadone toxicity (3,4). Fentanyl, a synthetic opioid analgesic, has been used clinically as an adjunct to surgical anesthesia and for chronic pain management in cancer patients. Fentanyl is 50 to 100 times more potent than morphine and may be administrated intravenously, transdermally, and transmucosally. Because of its lipophilicity and low molecular weight, fentanyl is readily absorbed through the skin. The fentanyl transdermal system, the Duragesic | patch designed by Janssen Pharmaceutical, is widely used in pain management. The amount of fentanyl release from the Duragesic patch transdermal system is proportional to the surface area of the patch (25 I~g/h/10 cm2), and four dose sizes are available (25, 50, 75, and 100 lag/h). An ethylene-vinyl acetate copolymer membrane controls the rate of fentanyl release and therefore the adsorption through the skin surface. The pathways of fentanyl metabolism include piperidine N-dealkylation to norfentanyl, amide hydrolysis to despropionylfentanyl, and terminal methyl hydroxylation to hydroxyfentanyl. Hydroxyfentanyl can further be dealkylated to hydroxynorfentanyl. Among these metabolism pathways, norfentanyl is the major metabolite accounting for greater than 99% of fentanyl metabolism (5). CYP3A subfamilies are the most abundant of all of the CYP isoenzymes, and they mediate the metabolism of a broad range of drugs. CYP3A4and CYP3A5 are the two major important enzymes to metabolize many drugs in liver (6). CYP3A4 enzyme

Reproduction(photocopying)of editorialcontentof thisjournal is prohibitedwithoutpublisher'spermission.


mediates the dealkylation of fentanyl to norfentanyl in liver as well as in intestinal epithelium. An in vitro study showed that CYP3A5is also a catalyst of fentanyl oxidation (7). CYP3A4and CYP3A5 are polymorphic. The most frequent and common single nucleotide CYP3A4 polymorphism is -392A > G (3A4*1B), and the polymorphism of CYP3A5 is 6986A > G (3A5"3), and these result in variable fentanyl metabolism. CYP3A4 mediates fentanyl metabolism, showing large interindividual variability (5). Although fentanyl abuse results in toxicity,the fentanyl toxicity may partially be due to CYP3A4 or CYP3A5variant alleles, resulting in impaired fentanyl metabolism. Therefore understanding the correlation of CYP3A4and 3.45with fentanyl toxicity and genotyping CYP3A4 and CYP3A5individually may aid in the death certification of fentanyl toxicity in postmortem forensic toxicologycases. In this study,we hypothesizedthat CYP3A4and/or 3/15variant alleles resulting in variable fentanyl metabolism may partially contribute to fentanyl toxicity. We investigated the correlation of CYP3A4and 3A5 genotype with fentanyl toxicity by studying 25 fentanyl-related death cases from the Milwaukee County MedicalExaminer'sOfficeand referral cases and assess the possibility to use pharmacogenomicsas a supplementarytool in certifying the cause and manner of fentanyl-related deaths.

Materials and Methods Case selection The study protocol was approved by the Institutional Review Board of the MedicalCollege of Wisconsin-Froedtert Memorial Lutheran Hospital. The cases in this study were selected from the time period between 2002 and 2004. Cases containing fentanyl were identified from cases investigated by the Milwaukee County Medical Examiner's Office, State of Wisconsin, with some referrals from other counties. The criteria used in case selection included cases with fentanyl identified in toxicology screen, cases certified with fentanyl toxicity, and mixed drug toxicity containing fentanyl. Deaths that were due to gunshot wounds, carbon monoxide poisoning, and fire were excluded. Further, the postmortem interval, medical history, history of suicidal ideation, previous suicide attempt, concomitant drug administration with potential drug-drug interactions, autopsy findings, medication history as a basis for acute versus chronic ingestion, and death scene investigations were reviewed.

Autopsy Autopsy of all the selected cases was performed in the Milwaukee County Medical Examiner's Office.A complete autopsy included dissection of the thoracic, abdominal, cranial, and neck compartments. Routine samples of subclavian blood and vitreous fluid were attempted upon initial body examination. Samples obtained during the autopsy includedvitreous fluid (if not collectedat initial body examination),bile, urine, peripheral blood (such as iliacvein), not central blood (such as aorta), or heart blood (in the absence of peripheral blood), a portion of liver, stomach contents, and pulled head hair. A dried blood sample on the absorbed paper was also collectedfor future DNA

analysis. The blood samples were preserved in 2% sodium fluoride to inhibit decompositionof drugs such as cocaine.All samples were refrigerated immediately followingthe autopsy.

Toxicological analysis All cases were subject to a full toxicological examination. When drugs were present at the initial scene investigation or further case investigation and medical history indicated the use of therapeutic or abused drugs that would not otherwise be routinely detected and identified by the described analytical techniques, tests were conducted either onsite or sent to an outside toxicologyreference laboratory to include these drugs as part of the comprehensive toxicological examination. Comprehensive toxicological examination for each case included screening tests for alcohols, carboxyhemoglobin,as well as therapeutic and abused drugs by a combination of color tests, immunoassays, and chromatographic techniques [gas chromatography (GC) and GC-tandem mass spectrometry (MS)]. Acid-dichromate test was used for detection of methyl, ethyl, isopropyl alcohol, acetone, and acetaldehyde. Blood was analyzed by the IL 482 or 682 Co-oximeter for the presence of carboxyhemoglobin. Therapeutic and abused drugs were extracted from blood using a 300-mg Clean Screen| solid-phase extraction column (UnitedChemicalTechnologies,Inc., Bristol, PA) and established protocol with subsequent detection by GC or GC-MS (UCT).ImmunalysisTM radioimmunoassay (RIA)kits (Immunalysis Corp., San Dimas, CA) were used for the detection of amphetamines, benzodiazepines,cannabinoids, cocaine metabolite, and opiates in blood. Analytes (drugs) detected by initial screening methodologies were then confirmed by an alternate analyticalmethodologyand quantified for assessment of drug toxicity.Blood chromatographic screening, confirmation, and quantification techniques utilized established GC or GC-MS protocols (3,4). In cases in which the fentanyl use was suspected based on the scene investigation or medical history, but negative in drug screen, postmortem blood samples were screened for fentanyl and norfentanyl by RIA (DiagnosticProducts Corp., Los Angeles,CA).The samples with positive screen were subsequently confirmed and quantified for both fentanyl and norfentanyl using liquid chromatography-MS-MS by National MedicalService.

DNA purification DNAwas extracted from the whole blood using PUREGENE| DNA isolation kit (Gentra Systems, Minneapolis, MN) with a modified procedure. Briefly,2 mL of red blood cells were lysed with 6 mL RBC LysisSolution. Aftercentrifuge supematant was removed, leaving behind the white cells pellet about 100-200 IJL of the residual liquid. Remaining white blood cells pellet were first vortex mixed vigorously and then lysed in the presence of 2 mL of cell lysissolution. AfterRNaseA treatment, cellular proteins were precipitated by adding 0.67 mL of protein precipitation solution and discardedprior to DNAprecipitation. DNAwas then precipitated with 2 mL of isopropanol. The resuspended DNAwas quantified by spectrophotometric methods (absorbance reading at 260 and 280 nm). The quality and purity were checked by A260/A280ratio, and the samples were stored at -20~

591


DNA amplification by polymerase chain reaction (PCR)

The two common variant alleles of CYP3A4*IBand 3.45*3 were genotyped using Pyroseqencing (Biotage). Briefly, a sequencing primer (A738FS for 3A4*lB and A542FS for 3A5"3) was hybridized to a single-stranded, previously PCR-amplified DNA template and incubated with a series of enzymes, dNTPs were added consecutively to the reaction. If the nucleotide added was complementary to the base on the template strand, the incorporation releases pyrophosphate in a quantity equimolar to the amount of incorporated nucleotide. ATP sulfurylase quantitatively converts PPi to ATP in the presence of adenosine 5' phosphosulfate. This ATP drives the conversion of luciferin to oxyluciferin that generates visible light detectable by a CCD camera and is evident by a peak in the Pyrogram. Each light signal is proportional to the number of nucleotides incorporated. Apyrase degrades the unincorporated dNTPs and excessATP prior to the addition of another dNTP.As the process continues, the complementary DNA strand is built up and the sequence determined from the signal peak in the Pyrogram TM.

the death in the 25 cases were 17 accidents, 1 suicide, 6 natural, and 1 undetermined. Genotyping of CYP3A4*IBand 3.45*3showed 1 CYP3A4*IB homozygous and CYP3AS*3 heterozygous, 1 compound CYP3A4*IBand CYP3AS*3heterozygous, 22 CYP3A4*IBwild type and CYP3AS*3 homozygous, and 1 CYP3AS*3 and CYP3A4HB wild type. A1Mic frequencies of the 25 cases were CYP3A4*IB, 0.06, and CYP3AS*3, 0.92, compared with variant allelic frequencies of normal Caucasians CYP3A4*IB, 0.033-0.08 (8,9), and CYP3AS*3,0.85--0.95 (10,11). The mean of fentanyl concentration of the 2 cases with CYP3A4*IBand 3.45*3 variant was 12.8 lag/L (6.6 and 19.0 lag/L), and it was lower, but not statistically significant (P > 0.05), than the mean (16.9 IJg/L,range 0.6 to 72.0 IJg/L)of the 22 cases with wild type CYP3A4*IB and CYP3AS*3 homozygous variant. MRs of the 2 cases with both CYP3A4*IB and 3A5"3 variant, and the 22 cases with wild type CYP3A4*IBand CYP3A5*3homozygous variant were 1.4 and 7.3, respectively. Furthermore, these 25 cases may be divided into 3 groups according to their genotypes: (1) compound variants carriers for both CYP3A4/5as in cases 1 and 2; (2) wild type CYP3A4and homozygous CYP3A5 as in cases 3 to 24; and (3) wild-type for both CYP3A4/5as in case 25. Selected case histories are included to illustrate the possible genetic effects of compound variants as in cases 1 and 2; the lack of genetic effect of CYP3A5variants in cases 4, 12, and 16 with either low or high metabolite norfentanyl concentrations; the possible effect of CYP3A5variants in cases 23 and 24 with high fentanyl concentrations and metabolite ratios; and the lack of genetic effect but a highly toxic acute ingestion in case 25. These cases are described in more detail.

Statistical analysis

Case 1

Student T-test and descriptive means were calculated using Microsoft Excel Program.

The decedent was an African-Americanfemale with a medical history significant for insulin-dependent diabetes, congestive heart failure, and renal failure. She apparently suffered anoxic brain injury during the perioperative period of a surgical procedure to amputate her right leg was belowthe knee about four years previously. The family of the decedent was involved in legal action against the anesthesiologist, alleging that the brain injury resulted from malpractice. Since the injury, the decedent had a change in both personality and mental status. She was admitted to a nursing home for a hospice care program six days prior to her death. She was comatose the whole time and was agitated, thrashing about in the bed. The doctor did not know the details of her past and assumed her coma stemmed from a brain injury. The decedent was given comfort measures only, which included the administration of pain medications as she might hurt herself in the bed. The decedent did not smoke, drink alcohol, or abuse drugs.

Fifty microliters of PCR reaction included 5 laL of 10x Qiagen PCR buffer, 1 laL of 10mM deoxynucleotide triphosphate bases (dNTP), 0.5 IJL of each 201JM primer (Primers A733FP and A742RPB in 3A4*lB reaction, and Primers A539FP and A540RPB in 3.45*3 reaction), 0.25 laL of 5 U/I~LQiagen HotStar Taq enzyme, 0.4 IJL of 50 IJg/IJLDNA sample, and 42.35 laL of Qiagen PCR water. The PCR reactions were first run at 95~ for 10 min, and then run 45 cycles of 40 s each at 95~ 55~ and 72~ After the 45 cycles, the reactions were held at 72~ for 5 rain.

Genotyping for CYP3A4 and 3A5 by pyrosequencing

Results We identified 25 fentanyl-related death cases, and 24 of them occurred in residence. Comprehensivetoxicologicalexamination for each case included testing for alcohols, carboxyhemoglobin, and therapeutic and abused drugs by a combination of color tests, immunoassays, and chromatographic techniques (GC and GC-MS). The age, sex, race, genotype, the toxicology results (fentanyl and norfentanyl, metabolic ratio of fentanyl to norfentanyl, alcohol, and other drugs), the cause of death, and manner of death are listed in Table I. Among the 25 cases, there were 16 females and 9 males and 22 Caucasians, 1 AfricanAmerican, and 2 Native-Americans.The average age of the decedents was 48. Fentanyl concentrations ranged from 0.6 to 100 I~g/Lwith a mean of 19.9 lag/L,and norfentanyl concentration ranged from 0.5 to 67 IJg/Lwith a mean of 10.8 lag/L.In 15 cases, the metabolic ratio (MR) of fentanyl to norfentanyl was greater than 1, with the range of 1.1 to 38.9. The causes of death in the 25 cases were 7 fentanyl toxicity, 9 mixed-drug toxicity, and 16 non-drug-toxicity related, and the manners of

592

Case 2

The decedent had a history of drug abuse and psychiatric problems. She used crack cocaine and marijuana and abused pain medications. In addition, she once cut her arms in an attempt to obtain medications and had expressed suicidal ideation. One day, she and her neighbors had a rummage sale. After the sale, she spoke with her boyfriend and complained

Journal of Analytical Toxicology, Vol. 29, O c t o b e r 2005

0

-~

~

z

.~

=

,_

o

~.~ 0

o:=

o

~

= o~

~ ~=~ o

..~ ~

~

"~_

"~ .__

~

~

-=

~..~.~ >

"-o

~a . ~

-~

o

"~ -.-

"~ ~_oo

-~-0 x a.

>.~

"('~~ E "I: ~

~-~-

~

~0.}

z

Z

D

.~ o

~_~ ~

_ "~' ~ _._m. --

'~z~

.~

o -ell~ u ~

._

~ ~

~-~

o

.,~..~ ~ ~

~a-.m =

~..._u ~

~ o

o

~

~

.,~ ~

~_~

~

0

.a .~r~ < Z t

593

Journal o f Analytical Toxicology, Vol. 29, O c t o b e r 2005

g

O

g

.~

~
. O

9- -

--~

~.--_o

:_~ ~.._= ._ .-~ .c~0 . - ~

.~

E

"~

o> ",5

=>r~

" ' ~ - ._= - ~ _~ E ~"E

~h~ ~

m

O

N

"G

.-~ a.-~

I

~

_>.

~ m._~-

}'~

_>=

x~-o

~.-~-8 o o~

O

"-5

N

-8

s

O

-85 .~

o

~

-~

&

u ~-~---~

..s~X

._~'~

o . ,.~.~

-

o

~o~

~

o

ov.~

~

--~'~

o-&-~q.~-~

~.~-~=~.~._>_

._

-

~

,~

m ~

"

,.~

m

.::...~_~

_ 8 " ~ u ~ N =_. oE

E ,.~ ~..~ -'=

E~

~,.. ~E"

E

~.~,~ =

~

,~

.-

.~

~.~

~D

.-[

~

"

.8o~

.~_ ~ . ' ~ ,

~

L.)

N

m ~

"~ r'~

r~ z

tm z

o

=

tm z

rm z

-8

-8 o

Z

-8 _~o

o

._@o

E E m.

z

...i o

~.

%

o.

,--:

,--=. I--..

%

o.

,--:.

o~

25 "5

,1.1

o,

E

o.

...i

o.

o.

o. "6

"I-

-r-

-r-"

I

-r"

d O

L;

Z

.o

.