Pharmacogenomics Variation in Drug Metabolizing Enzymes and Transporters in Relation to Docetaxel Toxicity in Lebanese Breast Cancer Patients: Paving the Way for OMICs in Low and Middle Income Countries

Abstract

We investigated the association of genetic polymorphisms in drug metabolizing enzymes (DMEs) and transporters in patients with docetaxel-induced febrile neutropenia, by a new high-throughput DMEs and transporters (DMETPlus) microarray platform, characterizing 1936 single nucleotide polymorphisms (SNPs) in 225 genes. We recruited 100 Lebanese breast cancer patients from a consecutive cohort of 277 patients who received docetaxel either alone, or in combination with trastuzumab. Out of 100 patients, 18 had developed febrile neutropenia (cases). They were age- and treatment- matched with 18 patients who did not develop febrile neutropenia on docetaxel (controls). We found that 12 SNPs in seven genes (ABCC6, ABCG1, ABCG2, CYP1A2, CYP2D6, FMO2, and FMO3) were significantly associated with febrile neutropenia after docetaxel treatment. Many of these SNPs have not been previously reported to be associated with toxicity due to docetaxel treatment. Interestingly, one SNP in the FMO3 gene (rs909530) was significantly associated with three clinical endpoints: febrile neutropenia, reduced absolute neutrophil count, and hemoglobin reduction. To the best of our knowledge, this is the first study that evaluated the effect of a large array of nearly 2000 polymorphisms in DMEs and transporters on docetaxel toxicity in breast cancer patients, and in a previously understudied population. Additionally, it attests to the feasibility of genomics research in low- and middle-income countries (LMICs). In light of the current global epidemic of noncommunicable diseases (NCDs) such as breast cancer impacting LMICs, we suggest pharmacogenomics is considered as an integral part of the global health research agenda for NCDs and personalized therapeutics.

Introduction

Docetaxel (Taxotere) is an antineoplastic agent that is prescribed, in addition to anthracycline-based therapy, for breast cancer patients with poor prognosis. The latter includes young patients, patients with hormone receptor-negative breast cancer, patients with nodal involvement, or patients with poorly differentiated tumor cells (Green and Hortobagye, 2007; Kobayashi et al., 2012). Docetaxel is also approved for treatment of other solid tumors, including non-small cell lung cancer, ovarian cancer, gastric cancer, and prostate cancer (Cui et al., 2012; Petrani et al., 2012; Sorbe et al., 2012; Milanowski et al., 2013). In patients with locally advanced or metastatic breast cancer, the recommended dosage is 60 mg/m² to 100 mg/m² given intravenously over 1 hour every 3 weeks. Most common adverse reactions associated with docetaxel therapy include hypersensitivity, neutropenia, febrile neutropenia, anemia, thrombocytopenia, and fluid retention. Grade 4 neutropenia (<500/mm³) is dose dependent, and occurs in 85% of patients given docetaxel 100 mg/m² and 75% of patients given 60 mg/m². Eventually, dose adjustment is recommended in cases of severe toxicity (Docetaxel FDA label, 2012).

The bulk of docetaxel is metabolized in the liver and excreted through the bile, with minimal urinary excretion (Docetaxel FDA label, 2012). Docetaxel enters the liver cell via the hepatocellular influx carrier, solute carrier organic anion transporter family member 1B3 (SLCO1B3), also known as organic anion-transporting polypeptide 1B3 (OATP1B3) (Smith et al., 2005). Inside the hepatocytes, docetaxel is oxidized by the cytochrome P450 enzymes, CYP3A4 and CYP3A5, to four inactive metabolites (Sparreboom et al., 1996; Shou et al., 1998). The primary route of elimination of the parent drug and hydroxylated metabolites is hepatobiliary excretion by the membrane localized energy-dependent drug efflux transporters: ABCB1, ABCG2, ABCC1, and ABCC2 (Oshiro et al., 2010) (Fig. 1). Intestinal P-glycoproteins dispose the remaining docetaxel left after circulation in the liver into the feces (Zuylen et al., 2000).

FIG. 1.

Pharmacokinetic pathway of docetaxel.

An important limitation associated with docetaxel treatment is the inter-patient heterogeneity, with up to 10-fold variability in clearance, even in patients with normal liver function (Longo et al., 2010). This enormous variation in docetaxel clearance may contribute to the variable drug response and toxicity. It has been reported that a 50% decrease in docetaxel clearance increases the odds of experiencing docetaxel-induced grade 4 neutropenia by 3-fold (Baker et al., 2009). Causes of these inter-individual variations may include age, gender, ethnicity, hepatic impairment, and genetic polymorphisms (Longo et al., 2010).

Several studies have attempted to identify genetic polymorphisms in genes encoding drug metabolizing enzymes (DMEs) and transporters accounting for the remarkable inter-individual variation in response to docetaxel. The results are conflicting and nonconclusive. For example, some reported an increase in docetaxel clearance in carriers of CYP3A4 polymorphisms (rs2740574) (Baker et al., 2009; Tran et al., 2006), a finding that was not supported by others (Bosch et al., 2006). Differences in sample size, various tumor types, different treatment regimens (single agent vs. combination of drugs interacting with docetaxel) and various ethnic groups may account for these conflicting results. In addition, most of these studies investigated the role of only few SNPs in genes involved in docetaxel disposition.

Recent understanding of drug metabolism and drug transport, and the realization that drug disposition is a complex multienzymatic process, have challenged the “candidate gene” approach (Deeken, 2009). For instance, many times SNPs in DMEs or drug transporters that are not involved in the elimination of a certain drug are found to be linked to toxicity and/or response of that drug. Few examples include the fact that although allelic variants in ABCB1 are only marginally involved in docetaxel clearance variability, few articles do report an association between these variants and docetaxel toxicity and outcome (Tsai et al., 2009). Another example includes the fact that although docetaxel is not known to be metabolized by CYP1B1 nor GSTP, CYP1B1*3 and GSTP1 A/B polymorphisms have been found to be linked with a poorer treatment outcome and febrile neutropenia respectively (Tran et al., 2006; Sissung et al., 2008). Therefore, it is now recommended to use genomic microarray technologies that do not require a priori gene identification, provide comprehensive genotyping information, and provide information on associations not previously expected (Deeken, 2009).

Although several genotyping platforms are available in the market, the relatively newly released DMETPlus microarray from Affymetrix is a much more suitable platform for pharmacogenomic studies, whereby it scans 1936 variants (1931 single nucleotide polymorphisms “SNPs” and 5 copy number variations “CNVs”) in 225 genes involved in the absorption, distribution, metabolism, and elimination of US FDA approved drugs (Sissung et al., 2010). To our knowledge, no investigators have used the DMET arrays in breast cancer patients. We are aware of only two publications in the setting of prostate cancer (Deeken et al., 2010) and uterine, ovarian, or peritoneal cancer (Uchimaya et al., 2012).

This is an hypothesis generating study that aimed at studying the association of a large number of genetic polymorphisms in DMEs and transporters with toxicity after docetaxel treatment in breast cancer patients. The questions posed were:

• What are the variants associated with the development of febrile neutropenia and/or myelosuppression upon treatment with docetaxel?

• Are there haplotypes associated with docetaxel-induced febrile neutropenia and/or myelosuppression?

Materials and Methods

Study sample

This is a nested case-control study within a larger retrospective cohort that included 277 Lebanese breast cancer patients who were admitted for chemotherapy at the American University of Beirut (AUB) Medical Center (AUBMC) between May 2009 and January 2012. Out of this consecutive sample, we were able to collect chemotherapy information for 210 patients, 100 of whom received 3–6 docetaxel cycles with or without trastuzumab (Herceptin). The dose of docetaxel was 100 mg/m², except one patient who received 75 mg/m². These were prescribed without prior chemotherapy, or after three to four cycles of cyclophosphamide with an anthracycline and/or 5-fluorouracil. Patients who received docetaxel concomitantly with cyclophosphamide and anthracyclines, and patients who were switched to paclitaxel for different reasons such as cost, were not included in the analysis (Fig. 2 and Table 1). This research project was approved by the AUB Institutional Ethics Review Board (IRB).

FIG. 2.

Flowchart of study participants.

Table 1.

Demographics and Clinical Characteristics of Initial Cohort of 100 Patients Receiving Treatment with Docetaxel

Demographics		All (N=100)^#	Cases (N=18)	Controls (N=18)	P value
Age at diagnosis (years)	Mean±SD	48±10.7	44.2±9.8	47.2±8.8	0.34^Ω
BMI (kg/m²)	Mean±SD	27.2±4.9	25.3±4.8	25.4±3.0	0.961^Ω
Type of breast cancer	Infiltrating ductal carcinoma	95	17 (94.4)^£	17 (94.4)	1.000^€
	Infiltrating lobular carcinoma	4	1 (0.06)	1 (0.06)
Stage	I	8	2 (11.1)	4 (23.5)	0.472^€
	II	50	9 (50.0)	10 (58.8)
	III	37	5 (27.8)	3 (17.6)
	IV	4	2 (11.1)	0
Grade	1	11	2 (11.1)	3 (16.7)	0.812^€
	2	37	5 (27.8)	6 (33.3)
	3	45	11 (61.1)	9 (50)
Estrogen receptor positive		74	11 (61.1)	14 (77.8)	1.000^€
Progestin receptor positive		60	12 (66.7)	11 (61.1)	1.000^€
Her2 overexpression		48	15 (83.3)	10 (55.6)	1.000^€
Menopausal status	Premenopausal	42	9 (50)	5 (27.8)	1.000^€
	Postmenopausal	42	7 (38.9)	11 (61.1)
	Menopause secondary to chemotherapy	16	2 (11.1)	2 (11.1)
Chemotherapy regimen^*	AC cycles followed by T cycles	27	2 (11.1)	2 (11.1)	1.000^€
	AC cycles followed by TH cycles	12	0	0
	FAC cycles followed by T cycles	31	4 (22.2)	5 (27.8)
	FAC cycles followed by TH cycles	24	10(55.6)	9 (50)
	TH cycles	6	2 (11.1)	2 (11.1)

Case patients developed febrile neutropenia after docetaxel treatment. Control patients did not develop febrile neutropenia after docetaxel treatment.

Numbers may not add up to 100 because of missing data, ^£N (%)

A, anthracyclines; C, cyclophosphamide; F, 5-fluorouracil; H, trastuzumab (herceptin); T, docetaxel (taxotere).

P value obtained using t-test and ^€p value obtained using Fisher-exact two sided test.

The charts of in- and out-patients were reviewed to collect baseline demographic data and details on the breast cancer patients at diagnosis. Clinical data such as chemotherapy regimen, doses of cytotoxic drugs, and number of cycles were recorded. White blood cell count (WBC), neutrophil and platelet counts, as well as hemoglobin (Hb) values, were collected 7–10 days after the first chemotherapy cycle and at each chemotherapy cycle when available. Treatment with granulocyte colony stimulating factor (GCSF), erythropoietin, iron, or packed red blood cell transfusion was also collected. Data on toxicities such as allergy, GI symptoms, and others were not reliably available and consequently were not collected.

Eighteen patients developed febrile neutropenia while on docetaxel alone or docetaxel and trastuzumab (cases). These cases were matched with 18 patients of similar ages who received the same treatment but did not develop febrile neutropenia (controls) (Tables 1 and 2). Although docetaxel is associated with many side effects including hematologic, neurosensory, and gastrointestinal (Docetaxel FDA label, 2012), febrile neutropenia was chosen as the primary endpoint because it is the most objective endpoint that has clinical implications and that was reliably retrieved from the patients' charts. Trastuzumab is not usually associated with neutropenia or febrile neutropenia (trastuzumab FDA label, 2000) and is not known to be eliminated by drug metabolizing enzymes and transporters. Patients who received prophylactic GCSF injections with every docetaxel dose were deliberately not chosen as controls, as they might have developed febrile neutropenia if no GCSF injection had been given. Other endpoints were chosen throughout the study. A more direct endpoint is the white blood cell count at nadir (7–10 days post-treatment). Nevertheless, clinical data were not available for all patients. White blood cell count 7–10 days after the first docetaxel cycle was available for 16 cases and 7 controls. Similarly, absolute neutrophil count was available for 13 cases and 7 controls. Change in neutrophil count and hematological data between the first two cycles of treatment with docetaxel were adopted as sharp endpoints.

Table 2.

Chemotherapy Regimen, Docetaxel Dose, Age, Tumor Stage and Grade

Case	Chemotherapy regimen	Docetaxel dose (mg)	Age (yrs)	Stage	Grade	Control	Chemotherapy regimen	Docetaxel dose (mg)	Age (yrs)	Stage	Grade
1	TH6	100	40	IIIB	3	1	TH6	100	46	IIIA	3
2	FEC3TH3	100	36	IV	3	2	FEC3TH3	100	48	IIA	3
3	TH4	100	58	IIIA	3	3	TH6	100	66	I	3
4	AC4T4	75	38	IIA	1	4	AC4T4	75	45	IIB	2
5	FEC3TH3	100	43	I	2	5	FEC3TH3	100	49	I	1
6	FEC3TH3	100	31	IIB	3	6	FEC3TH3	100	40	IIB	3
7	FEC3T3	100	38	IIA	1	7	FEC3T3	100	39	IIA	2
8	FEC3T3	100	51	IV	2	8	FEC3T3	100	51	IIA	3
9	FEC4T4	100	54	IIB	3	9	FEC3T4	100	51	IIIA	1
10	EC4T4	100	34	IIIA	2	10	AC4T4	100	37	IIB	2
11	FEC3TH3	100	54	IIA	2	11	FEC3TH3	100	50	IIIA	3
12	FEC3TH3	100	45	IIB	3	12	FEC3TH3	100	52	I	3
13	FEC3TH3	100	40	IIB	3	13	FEC3TH3	100	33	I	3
14	FEC3TH3	100	48	III	3	14	FEC3TH3	100	50	IIA	1
15	FEC3TH3	100	69	I	3	15	FEC3TH3	100	65	IIB	3
16	FEC3T3	100	39	IIIB	3	16	FEC3T3	100	43	N/A	2
17	FEC3TH3	100	41	IIA	2	17	FEC3TH3	100	48	IIA	2
18	FEC3TH3	100	36	IIA	3	18	FEC3T3	100	36	IIB	2

Case patients developed febrile neutropenia on docetaxel. Age- and treatment matched- control patients did not develop febrile neutropenia on docetaxel.

A, anthracyclines; C, cyclophosphamide; E, epirubicin;F, 5-fluorouracil; H, trastuzumab (herceptin); T, docetaxel (taxotere).

Genotyping using the drug metabolizing enzymes and transporters (DMETPlus) array

Two milliliters of whole blood were collected. DNA was isolated using a DNA isolation kit from Qiagen (Germantown; MD, USA) according to the manufacturer's guidelines and stored at −20°C until analysis. All genomic DNA samples were normalized to a single concentration of 60 ng/μL using Tris EDTA.

Samples were then genotyped using DMETPlus arrays from Affymetrix. The Affymetrix DMETPlus protocol was followed (DMET user guide, 2012). Scanning was performed with the genechip scanner 3000 7G and base calls were generated by the DMET console software, which is based on the Bayesian Robustly fitted, Linear Model with the Mahalanobis distance classification (BRLMM) algorithm (Cawley et al., 2006). A base call was made by the DMET console software if the SNR is ≥3 (Hardenbol et al., 2003). Three genomic DNA controls were run with the samples. They served as positive controls to detect assay performance in case the processed samples were of marginal DNA quality. Two random samples were run in duplicates to assess for assay reproducibility.

Concordance and reproducibility

The DMETPlus data were compared with 6 SNPs that were done as part of the larger cohort. Three CYP2B6 single nucleotide polymorphisms (SNPs) were screened using LightSnip® kits (TibMolbiol; Berlin, Germany) on the LightCycler® real time PCR platform (Hoffmann-La Roche; Basel, Switzerland). GSTP1 6624 A>G (rs947894/rs1695) was screened using LightSnip® kits (TibMolbiol; Berlin, Germany) on the LightCycler® real time PCR platform (Hoffmann-La Roche; Basel, Switzerland), while genotyping for GSTT1 and GSTM1 473 base pair and 210 base pair deletions, respectively, was performed using a polymerase chain reaction (PCR) followed by visualization of the bands on an agarose gel (Chen et al., 1996). Ten percent of the samples were run twice for assay reproducibility (Table 3).

Table 3.

Drug Metabolizing Enzymes and Transporters (DMETPlus) Quality Control (QC) and Concordance with RT-PCR and RFLP

Polymorphism	Assay	rs number	Nucleotide change	QC of the variant	Concordance
CYP2B6^4*	RT-PCR	rs2279343	785A>G	100%	100%
CYP2B6^5*	RT-PCR	rs3211371	1459C>T	43.9%	92.1%
CYP2B6^6*	RT-PCR	rs3745274	516G>T	92.7%	100%
GSTP1	RT-PCR	rs1695	6624A>G	100%	100%
GSTM1	RFLP	N/A	210bp deletion	100%	100%
GSTT1	RFLP	N/A	473bp deletion	95.1%	100%

RFLP, restriction fragment length polymorphism; RT-PCR, Real time polymerase chain reaction;

N/A, not applicable.

Data analysis

Of the 1936 markers, we performed genotype translational analysis using Affymetrix supplied translation and annotation files (DMET Plus v1.20120731), and subsequent analyses were limited to these markers. The visualizations and analysis were conducted in R statistical environment (v2.14.1); a p value of less than 0.05 was considered statistically significant. In addition to association with febrile neutropenia, other endpoints that indicate myelosupression were analyzed (Tables 4, 5, 6, and 7). The genotype calls reported by the Affymetrix DMET console for each SNP site across all patients were segregated into homozygous wild-type (Ref/Ref), heterozygous (Ref/Var), homozygous variant (Var/Var), or ‘No Call’. Samples were included in the analysis only if they had more than 85% of total variants successfully genotyped. Alleles that were not in Hardy-Weinberg Equilibrium (HWE) were excluded from analysis.

Table 4.

Myelosuppression Variables after Docetaxel Injection

Treatment outcome	Cases ( N =18)	Controls ( N=18)	P value^#
NCI grade of white blood cell count 7–10 days after first docetaxel injection
1	3/16 (18.8)^*	2/7 (28.6)	0.097
2	1/16 (6.2)	3/7 (42.9)
3	7/16 (43.8)	2/7 (28.6)
4	5/16 (31.2)	0/7 (0)
NCI grade of absolute neutrophil count 7–10 days after first docetaxel injection
1	2/13 (15.4)	1/7 (14.3)	0.075
2	0 (0)	1/7 (14.3)
3	0 (0)	2/7 (28.6)
4	11/13 (84.6)	3/7 (42.9)
Need for GCSF	15/18 (83.3)	2/18 (11.1)	0.000
Need for red blood cell transfusion or erythropoietin
	3/18 (16.7)	1/18 (5.6)	0.603
NCI grade of hemoglobin 7–10 days after first docetaxel injection
1	13/16 (81.2)	6/7 (85.7)	1.00
2	3/16 (18.8)	1/7 (14.3)
NCI grade of platelet count 7–10 days after first docetaxel injection
1	15/15 (100)	7/7 (100)	1.00

“Case” patients developed febrile neutropenia; “control” patients did not develop febrile neutropenia.

N/available of the 18 (%); ^#P values obtained using Fisher-exact two sided test.

GCSF, granulocyte colony stimulating factor.

Grading according to National Cancer Institute (NCI) hematological toxicity criteria.

Table 5.

Frequency Distribution and Characteristics of Statistically Significant SNPs in Drug Metabolizing Enzymes and Transporters (DMETs)

Gene	SNP	rs number	Genotype	Case/Control	HWE	Call rate (%)	Function ^*	Predicted function ^*	Risk allele	P value
ABCC6	c.3803G>A	rs2238472	Ref/ Ref	14/7	0.588	72	Missense	Tolerated	Ref	0.012
			Ref/Var	0/5
			Var/Var	0/0
ABCG2	c.421C>A	rs2231142	Ref/ Ref	18/13	0.654	100	Missense	Tolerated	Ref	0.0455
			Ref/Var	0/5
			Var/Var	0/0
CYP1A2	5347T>C	rs2470890	Ref/ Ref	9/2	0.516	100	Synonymous	N/A	Ref	0.0288
			Ref/Var	7/9
			Var/Var	2/7
CYP2D6	100C>T	rs1065852	Ref/ Ref	16/9	0.081	100	Missense	Damaging	Ref	0.036
			Ref/Var	2/6
			Var/Var	0/3
CYP2D6	−2178G>A	rs28360521	Ref/ Ref	16/9	0.081	100	Intronic	N/A	Ref	0.036
			Ref/Var	2/6
			Var/Var	0/3
FMO2	c.337delG	rs28369860	Ref/ Ref	18/13	0.654	100	Missense	Damaging	Ref	0.0455
			Ref/Var	0/5
			Var/Var	0/0
FMO2	c.1239T>G	rs2020865	Ref/ Ref	18/13	0.654	100	Missense	Tolerated	Ref	0.0455
			Ref/Var	0/5
			Var/Var	0/0
FMO2	c.210_211insGAC	rs2020868	Ref/ Ref	18/13	0.654	100	Synonymous	N/A	Ref	0.0455
			Ref/Var	0/5
			Var/Var	0/0
FMO2	c.242T>C	rs2020860	Ref/ Ref	18/13	0.654	100	Missense	Damaging	Ref	0.0455
			Ref/Var	0/5
			Var/Var	0/0
FMO2	c.941A>G	rs2020863	Ref/ Ref	15/8	0.469	83	Missense	Damaging	Ref	0.0309
			Ref/Var	1/6
			Var/Var	0/0
ABCG1	c.973+672G>A	rs3788007	Ref/ Ref	7/16	0.701	100	Intronic	N/A	Var	0.0045
			Ref/Var	10/2
			Var/Var	1/0
FMO3	c.855C>T	rs909530	Ref/ Ref	11/17	0.453	100	Synonymous	N/A	Var	0.0408
			Ref/Var	7/1
			Var/Var	0/0

Case patients developed febrile neutropenia; control patients did not develop febrile neutropenia on docetaxel.

Based on SNPnexus database (Chelala et al., 2009).

HWE, Hardy-Weinberg equilibrium; N/A, Not applicable; Ref, reference allele; Var, variant allele.

Table 6.

Mean and 95% Confidence Interval (CI) of Change in Absolute Neutrophil Count from Cycle 1 to Cycle 2 of Docetaxel Treatment among Different Genotypes

Gene	SNP	rs number	Predicted function ^*	Effect on protein function ^*	Genotype	Mean	95% CI	HWE	Call rate (%)	Risk allele	P value
ABCB11	c.-12519C>T	rs4148768	Intronic	N/A	Ref/Ref	3.774	−4.386; 11.934	0.102	100	Ref	0.0212
					Ref/Var	6.25	2.638; 9.862
					Var/Var	70	(N/A ; N/A)
ABCB11	c.-8583G>A	rs3770603	Intronic	N/A	Ref/Ref	3.7	−4.735; 12.135	0.21	100	Ref	0.0211
					Ref/Var	6.2	3.401; 8.999
					Var/Var	70	(N/A ; N/A)
ABCC6	c.2835C>T	rs2856585	Synonymous	N/A	Ref/Ref	4.057	−3.17; 11.285	0.933	100	Ref	0.0053
					Ref/Var	70	(N/A ; N/A)
					Var/Var	N/A	N/A
CYP1A1	2452C>A	rs1799814	Nonsynonymous	Tolerated	Ref/Ref	1.5	−6.382; 9.382	0.649	97	Ref	0.0140
					Ref/Var	29.8	6.797; 52.803
					Var/Var	N/A	N/A
CYP2C8	11041C>G	rs1058930	Nonsynonymous	Damaging	Ref/Ref	3.062	−4.59; 10.715	0.724	100	Ref	0.0453
					Ref/Var	28.5	−2.423; 59.423
					Var/Var	N/A	N/A
CYP3A5	6986A>G	rs776746	Intronic	N/A	Ref/Ref	N/A	N/A	0.585	100	Ref	0.0372
					Ref/Var	−12.667	−35.206; 9.872
					Var/Var	9.6	1.748; 17.452
SLCO1B3	c.334T>G	rs4149117	Nonsynonymous	Tolerated	Ref/Ref	9.5	3.438; 15.562	0.585	100	Ref	0.0430
					Ref/Var	−12.167	−47.48; 23.147
					Var/Var	N/A	N/A
CYP2B6	64C>T	rs8192709	Nonsynonymous	Damaging	Ref/Ref	8.394	1.081; 15.707	0.862	97	Var	0.0193
					Ref/Var	−32.5	−102.079; 37.079
					Var/Var	N/A	N/A
CYP2B6	12740G>C	rs2279341	Synonymous	N/A	Ref/Ref	8.394	1.081; 15.707	0.794	100	Var	0.0369
					Ref/Var	−21.667	−67.104; 23.771
					Var/Var	N/A	N/A
CYP2C8	11054A>T	rs11572103	Nonsynonymous	Damaging	Ref/Ref	8	1.089; 14.911	0.933	100	Var	0.0010
					Ref/Var	−68	(N/A ; N/A)
					Var/Var	N/A	N/A
CYP4B1	c.964A>G	rs45467195	Nonsynonymous	Damaging	Ref/Ref	7.786	−0.473; 16.044	0.777	86	Var	0.0253
					Ref/Var	−26	−67.546; 15.546
					Var/Var	N/A	N/A
CYP4F2	2099C>T	rs8110714	Synonymous	N/A	Ref/Ref	8.037	−0.887; 16.961	0.923	78	Var	0.0049
					Ref/Var	−66	(N/A ; N/A)
					Var/Var	N/A	N/A
FMO3	c.855C>T	rs909530	Synonymous	N/A	Ref/Ref	12.429	6.041; 18.817	0.453	100	Var	0.0013
					Ref/Var	−17	−38.897; 4.897
					Var/Var	N/A	N/A
FMO3	c.923A>G	rs2266780	Nonsynonymous	Tolerated	Ref/Ref	10.185	1.483; 18.887	0.847	81	Var	0.0373
					Ref/Var	−29.5	−101.039; 42.039
					Var/Var	N/A	N/A
SLCO1B3	c.1557G>A	rs2053098	Synonymous	N/A	Ref/Ref	9.5	3.438; 15.562	0.585	100	Var	0.0430
					Ref/Var	−12.167	−47.48; 23.147
					Var/Var	N/A	N/A
SULT1A1	c.638G>A	rs9282861	Nonsynonymous	Damaging	Ref/Ref	7.348	−2.828; 17.524	0.701	100	Var	0.0067
					Ref/Var	9.083	2.63; 15.536
					Var/Var	−66	(N/A ; N/A)
VKORC1	c.173+525C>T	rs17708472	Intronic	N/A	Ref/Ref	12.192	5.133; 19.252	0.693	97	Var	0.0006
					Ref/Var	−7.125	−24.249; 9.999
					Var/Var	−66	(N/A ; N/A)
UGT2B15	c.168C>T	rs3100	3’-UTR	N/A	Ref/Ref	11.4	−4.09; 26.89	0.825	100	N/A	0.0075
					Ref/Var	−7.375	−19.338; 4.588
					Var/Var	18.2	8.657; 27.743
UGT1A1	UGT1A1^28*	rs8175347	Intronic	N/A	Ref/Ref	21	−8.399; 50.399	0.092	75	N/A	0.0450
					Ref/Var	−1.647	−10.258; 6.964
					Var/Var	20.25	2.322; 38.178

Based on SNPnexus database (Chelala et al., 2009).

N/A, Not applicable; Ref, reference allele; Var, variant allele.

Table 7.

Mean and 95% Confidence Interval (CI) of Change in Hemoglobin from Cycle 1 to Cycle 2 of Docetaxel Treatment Among Different Genotypes

Gene	SNP	rs number	Predicted function ^*	Effect on protein function ^*	Genotype	Mean	95% CI	HWE	Call rate (%)	Risk Allele	P value
ABCC1	c.1385G>T	rs8056298	3’-UTR	N/A	Ref/Ref	−0.753	−0.948; −0.558	0.932	97	Ref	0.0002
					Ref/Var	1.7	N/A; N/A
					Var/Var	N/A	N/A
ABCC1	c.1512T>C	rs212091	3’-UTR	N/A	Ref/Ref	−0.797	−0.998; −0.595	0.72	97	Ref	0.0059
					Ref/Var	0.2	−0.865; 1.265
					Var/Var	N/A	N/A
ABCC4	c.2712G>A	rs1678339	Synonymous	N/A	Ref/Ref	−0.685	−0.912; −0.458	0.848	81	Ref	0.0406
					Ref/Var	0.4	−2.148; 2.948
					Var/Var	N/A	N/A
ABCC4	c.951A>G	rs2274406	Synonymous	N/A	Ref/Ref	−0.87	−1.093; −0.647	0.505	100	Ref	0.0382
					Ref/Var	−0.869	−1.205; −0.533
					Var/Var	−0.22	−0.713; 0.273
ABCC6	c.1890C>G	rs8058696	Synonymous	N/A	Ref/Ref	−0.777	−1.015; −0.539	0.65	81	Ref	0.0153
					Ref/Var	−0.683	−1.045; −0.322
					Var/Var	0.3	−0.683; 1.283
CYP1A1	2454A>G	rs1048943	Nonsynonymous	A/T Damaging;	Ref/Ref	−0.713	−0.948; −0.479	0.855	89	Ref	0.0178
				A/C Tolerated	Ref/Var	0.5	−0.676; 1.676
					Var/Var	N/A	N/A
CYP2C18	c.1154C>T	rs2281891	Nonsynonymous	Damaging	Ref/Ref	−0.816	−0.985; −0.647	0.654	100	Ref	0.0046
					Ref/Var	0.1	−1.01; 1.21
					Var/Var	N/A	N/A
CYP2C19	19154G>A	rs4244285	Synonymous	N/A	Ref/Ref	−0.8	−0.972; −0.628	0.585	100	Ref	0.0300
					Ref/Var	−0.133	−1.149; 0.882
					Var/Var	N/A	N/A
FMO3	c.769G>A	rs1736557	Nonsynonymous	Tolerated	Ref/Ref	−0.744	−0.96; −0.528	0.864	100	Ref	0.0480
					Ref/Var	0.25	−1.416; 1.916
					Var/Var	N/A	N/A
NAT1	c.459G>A	rs4986990	Synonymous	N/A	Ref/Ref	−0.726	−0.961; −0.491	0.858	92	Ref	0.0493
					Ref/Var	0.3	−1.268; 1.868
					Var/Var	N/A	N/A
SLC22A1	c.113G>A	rs35888596	Synonymous	N/A	Ref/Ref	−0.753	−0.948; −0.558	0.864	100	Ref	0.0205
					Ref/Var	0.4	−2.148; 2.948
					Var/Var	N/A	N/A
UGT1A1	c.211G>A	rs4148323	Nonsynonymous	Tolerated/Damaging	Ref/Ref	−0.757	−0.946; −0.568	0.933	100	Ref	0.0002
					Ref/Var	1.7	N/A; N/A
					Var/Var	N/A	N/A
FMO3	c.855C>T	rs909530	Synonymous	N/A	Ref/Ref	−0.55	−0.81; −0.29	0.453	100	Var	0.0228
					Ref/Var	−1.175	−1.466; −0.884
					Var/Var	N/A	N/A

Based on SNPnexus database (Chelala et al., 2009).

N/A, not applicable; Ref, reference allele; Var, variant allele.

The counts for Ref/Ref, Ref/Var and Var/Var genotypes were computed for each SNP site across the cases and control groups and compared using Fishers Exact test with Mehta's adjustment (Mehta and Patel, 1986; Clarkson et al., 1993). SNPs having counts with less than 7 calls for either of the two groups were eliminated from subsequent comparison due to lack of statistical power (Table 5). Haplotype calls that were generated by the software were also tabulated and analyzed in relation with febrile neutropenia, using Fishers Exact test with Mehta's adjustment. Counts of cases and controls were adjusted against total counts of all other haplotypes within the same gene.

The change in the absolute neutrophil count and Hb were assessed for association with underlying genotypes. One-way analysis of variance (ANOVA) was applied to assess potential association between the given genotype categories and corresponding change in the hematological variables between the first two cycles of docetaxel (Tables 6 and 7). We did not assess for change in platelet count as all patients had a platelet count >75000 at 7–10 days post treatment (i.e., at nadir), and reduction in platelet count is not known to be an early toxicity of docetaxel.

We were unable to analyze the change in hematological variables over additional time points, as most patients who developed febrile neutropenia received GCSF. Furthermore, we did not analyze “nadir” data due to incomplete data (Table 4). As this was a hypothesis generating and feasibility study of genomics research, no sample calculation was performed.

Results

Assay performances

Two samples were repeated on the DMETPlus array as they had less than 85% of total variants successfully genotyped. After that, all samples had a call rate of greater than 85%; 32 had a call rate of more than 90%. Duplicate samples showed 99% reproducibility. Genomic controls showed high call rates, indicating that the samples were successfully run on Affymetrix. Of all the analyzed SNPs, 96% of the SNPs passed HWE, and 93% of the SNPs were successfully called in greater than 70% of the samples. There was 100% concordance for CYP2B6*4, CYP2B6*6, GSTP1, GSTM1, and GSTT1, and 92.1% concordance for CYP2B6*5 (Table 3). Noteworthy, CYP2B6*5 polymorphism, in contrast to CYP2B6*4, CYP2B6*6, GSTP1, GSTM1, and GSTT1 was not successfully genotyped by the DMETPlus platform (QC i.e., % of samples having valid genotypes for CYP2B6*5 was 43.9%).

Sample characteristics

Tables 1 and 2 show the characteristics of the subjects. As expected, there were no significant differences in the age distribution and treatment regimens between the cases and controls. There were also no significant differences in the breast cancer subtypes, histological grade, molecular markers, and the clinical stage of disease. Furthermore, the incidence of febrile neutropenia in our sample population was 18%, a risk comparable to that reported by others (Vanderberg et al., 2010).

As expected, the need for GCSF was statistically significantly different in the cases vs. control groups since GCSF injection is usually given after the onset of febrile neutropenia. In contrast to GCSF, the need for red blood cell transfusion or erythropoietin was not statistically significantly different between the two groups. Since only 3 patients received erythropoietin or red blood cell transfusion, it was not relevant to study the association of genetic polymorphisms in drug metabolizing enzymes and transporters and these treatments (Table 4).

Association of DMETPlus genotypes with febrile neutropenia after docetaxel therapy

There were no prominent differences in the genotype distribution of the SNPs between cases and control groups (Fig. 3). Fourteen SNPs in 7 genes (ABCC6 c.3803G>A, ABCG1 c.973+672G>A, ABCG2 c.421C>A, CYP1A2*1F 163C>A, CYP1A2 5347T>C, CYP2D6 100C>T and −2178G>A, DPYD c.496A>G, FMO2 c.337delG, c.1239T>G, c.210_211insGAC, c.242T>C, and c.941A>G, and FMO3 c.855C>T) revealed a statistically significant association with febrile neutropenia from docetaxel treatment (p<0.05) (Fig. 4). Twelve passed the HWE test and were successfully genotyped in >70% of the patients. Table 5 lists the basic information of these 12 significant SNPs. ABCG1 c.973+672G>A and FMO3 c.855C>T polymorphic alleles were more frequent in the cases, whereas the remaining SNPs were more frequent in the controls. Consequently, ABCG1 c.973+672G>A or FMO3 c.855C>T polymorphic alleles may increase the risk of developing febrile neutropenia after docetaxel treatment, whereas the polymorphic alleles ABCC6 c.3803G>A, ABCG2 c.421C>A, CYP1A2*1F 163C>A, CYP1A2 5347T>C, CYP2D6 100C>T, −2178G>A, DPYD c.496A>G, FMO2 c.337delG, c.1239T>G, c.210_211insGAC, c.242T>C, or c.941A>G may decrease the risk of developing febrile neutropenia after docetaxel treatment. Haplotype analysis revealed no association between haplotypes and febrile neutropenia with docetaxel.

FIG. 3.

Distribution of the 1936 SNPs in 225 genes among cases (patients who developed febrile neutropenia after docetaxel treatment), and controls (patients who did not develop febrile neutropenia after docetaxel treatment).

FIG. 4.

P values of the association of different SNPs with the development of febrile neutropenia after treatment with docetaxel. The gray line represents the threshold of significance (p=0.05).

Association of DMETPlus genotypes with myelosuppression after docetaxel therapy

Twenty three SNPs in 18 genes were associated with reduction in absolute neutrophil count after one cycle of docetaxel treatment. Of these 23 SNPs, 19 were in HWE and successfully called in >70% of the patients. The p values, mean, and confidence intervals of the change in absolute neutrophil count, and details of these 20 SNPs are presented in Table 6. Twenty-two SNPs in 18 genes were associated with reduction in Hb after one cycle treatment with docetaxel. Thirteen SNPs passed HWE test and were successfully called by >70% of the patients (Table 7).

Note that one SNP, FMO3 c.855C>T, was significantly associated with febrile neutropenia, reduction in absolute neutrophil count, and reduction in Hb with p values of 0.001, 0.023, and 0.041 respectively. Therefore, C855T variant in FMO3 gene may exacerbate docetaxel-induced myelosuppression.

Discussion

Docetaxel is known to be associated with wide unpredictable inter-individual variability in efficacy and toxicity (Baker et al., 2009). It has been proposed that inherited differences in metabolism and excretion could explain the variable pharmacokinetics and pharmacodynamics of docetaxel (Longo et al., 2010). Although multiple studies were done to determine possible associations between genetic polymorphisms in some pathway genes and docetaxel treatment, none fully identified the genetic determinant(s) of docetaxel treatment outcome. The results of these studies are summarized in Tables 8 and 9. Most of these studies evaluated the effect of only few functional variants in some pathway genes on docetaxel clearance or toxicity. In order to obtain a comprehensive understanding of the pharmacogenomics of docetaxel, we used the DMETPlus genotyping platform, which enables us to genotype a large number of SNPs in genes related to drug absorption, distribution, metabolism, and elimination.

Table 8.

Data Representing Various Articles that Studied Effects of Genetic Polymorphisms in Drug Metabolizing Enzymes and Transporters on Pharmacokinetic Parameters and/or Treatment Outcomes of Docetaxel

Study	Type of cancer	Treatment regimen	Sample size	Genes studied
Bosch et al., 2006¹	Breast, prostate, lung, other cancers	Docetaxel alone or in combination^*	92	CYP3A4, CYP3A5 and ABCB1
Tran et al., 2006²	Breast, prostate, lung, other cancers	Docetaxel^#	58	CYP3A4, CYP3A5, ABCB1, GSTM1, GSTT1, GSTM3 and GSTP1
Baker et al., 2009³	Breast, prostate, lung, other cancers	Docetaxel either alone or with Capecitabine, Cyclophosphamide, Doxorubicin, Trastuzumab or other medications	92	SLCO1B3, ABCB1, ABCC2, CYP3A4 and CYP3A5
Tsai et al., 2009⁴	Breast cancer	TEC for 6 cycles	59	CYP3A4, CYP3A5 and ABCB1
Kiyotani et al., 2008⁵	Lung, breast, esophageal, other cancers	Docetaxel alone or in combination^*	84	CYP3A4, CYP3A5, ABCB1, ABCC2, SLCO1B3, NR1I2, NR1I3
Deeken et al., 2010⁶	Prostate cancer	Docetaxel alone or with thalidomide	47	1229 variants in 169 DMETs
Sissung et al., 2008⁷	Prostate cancer	Docetaxel alone or with thalidomide and estramustine or with prednisone	52	CYP1B1
Uchiyama et al., 2012⁸	Uterine, peritoneal, ovarian cancers	Docetaxeland carboplatin	42	1936 variants in 225 DMETs

TEC, docetaxel (taxotere), epirubicin, and cyclophosphamide.

The combination regimen was not specified in the study.

The study did not mention whether docetaxel was given alone or with other chemotherapeutic agents.

Table 9.

Available Literature and Data on Effect of Genetic Polymorphisms in Drug Metabolizing Enzymes and Transporters on Docetaxel Pharmacokinetics and Pharmacodynamics

			Pharmacokinetics		Pharmacodynamics
Gene	Polymorphism	rs number	Effect	Risk allele	Effect	Risk allele
CYP3A4	−A392G	rs2740574	No effect¹		No effect²
			Higher clearance^2;3	Var
	T15713C	rs55785340	No effect¹		N/A
	T23171C	rs4986910	No effect¹		N/A
	A352G	rs55951658	N/A		No effect⁴
	C653G	rs55901263	N/A		No effect⁴
	C21896T	rs12721629	No effect¹		N/A
	T20070C	rs28371759	N/A		No effect⁴
CYP3A5	C27289A	rs28365083	No effect¹		N/A
	A6986G	rs776746	No effect¹		No effect²
			Higher clearance³	Ref	Febrile neutropenia⁴	Ref
			No effect²
ABCB1	−41 A>G	N/A	N/A		No effect⁴
	−145 C>G	rs34976462	N/A		No effect⁴
	C3435T	rs1045642	No effect^1;2;3		Grade3 neutropenia²	Var
					Neutropenia⁴	Ref
	G2677T/A	rs2032582	No effect^1;2;3		No effect²
					Febrile neutropenia⁴	Ref
	C1236T	rs1128503	No effect³		No effect⁴
			Lower clearance¹	Var
ABCC2	−1019A>G	rs2804402	No effect³		N/A
	−24C>T	rs717620	No effect³		N/A
	G1249A	rs2273697	No effect³		N/A
	IVS26 −34T>C	rs8187698	No effect³		N/A
	C3972T	rs3740066	No effect³		N/A
	G4544A	rs8187710	No effect³		N/A
	G101620771C	rs12762549	N/A		Leukopenia/neutropenia⁵	Var
ABCC6	G3803A	rs2238472	N/A		Toxicity⁶	N/A
GSTM1	Wild type	N/A	No effect²		No effect²
	Null allele
GSTM3	3bp deletion	rs1799735	No effect²		No effect²
GSTP1	A313G	rs1695	No effect²		Hematologic toxicity²	Var
	A313G+C341T		No effect²		No effect²
	C341T	rs1138272	No effect²		No effect²
SLCO1B3	T334G	rs4149117	No effect³		N/A
	A439G	rs57585902	No effect³		N/A
	G699A	rs7311358	No effect³		N/A
	G767C	rs60140950	No effect³		N/A
	A1599C	N/A	No effect³		N/A
	T1679C	rs12299012	No effect³		N/A
	IVS11-5676A>G	rs11045585	N/A		Leukopenia/neutropenia⁵	Var
CYP1B1	C4326G	rs1056836	N/A		Poor prognosis⁷	Var
PPAR-6	A48189G	rs6922548	N/A		Response⁶^*	N/A
	G73444A	rs2016520	N/A		Response⁶^*	N/A
	G35369806A	rs1883322	N/A		Response⁶^*	N/A
	C89676T	rs3734254	N/A		Response⁶^*	N/A
	G57191A	rs7769719	N/A		Response⁶^*	N/A
SULT 1C2	G108994808C	rs1402467	N/A		Response⁶^*	N/A
CHST3	C50388T	rs4148943	N/A		Response⁶^*	N/A
	T50998C	rs4148947	N/A		Response⁶^*	N/A
	G53895A	rs12418	N/A		Response⁶^*	N/A
	A53643G	rs730720	N/A		Response⁶^*	N/A
	G52587A	rs4148950	N/A		Toxicity⁶^*	N/A
	G52895A	rs1871450	N/A		Toxicity⁶^*	N/A
	C50471T	rs4148945	N/A		Toxicity⁶^*	N/A
spastic paraplegia 7 (SPG7)	T43319C	rs2292954	N/A		Toxicity⁶^*	N/A
	G50524A	rs12960	N/A		Toxicity⁶^*	N/A
CYP2D6	2539-2542del	rs72549353	N/A		Toxicity⁶^*	N/A
NAT2	G590A	rs1799931	N/A		Toxicity⁶^*	N/A
ATP7A	G2299C	rs2227291	N/A		Toxicity⁶^*	N/A
CYP4B1	C517T	rs4646487	N/A		Toxicity⁶^*	N/A
SLC10A2	C26469T	rs2301159	N/A		Toxicity⁶^*	N/A
CYP39A1	T56503A	rs7761731	Higher AUC	Var	Grade 4 neutropenia⁸	Var

No information available; N/A, not applicable.

AUC, area under the curve; Ref, reference allele; Var, variant allele.

Numbers refer to references that are listed in Table 8.

Our results show that 12 SNPs in 7 genes (ABCC6, ABCG1, ABCG2, CYP1A2, CYP2D6, FMO2, and FMO3) were significantly associated with febrile neutropenia after docetaxel treatment. Furthermore, since myelosuppression is known to be an early toxicity of docetaxel (Docetaxel FDA label, 2012), we studied the association between genetic polymorphisms in DMETs and reduction in Hb and absolute neutrophil count after the first cycle of docetaxel. Several SNPs appeared to be involved (Tables 6 and 7). Out of all results, 6 SNPs are located in 4 genes potentially involved in docetaxel metabolism and transport (Table 10).

Table 10.

SNPs in Genes Involved in Docetaxel Metabolism and Transport on Docetaxel that Were Associated with Docetaxel-Induced Febrile Neutropenia, Reduction in Neutrophils, or Reduction in Hemoglobin

Gene	rs number	Risk allele for febrile neutropenia	Risk allele for reduction in absolute neutrophil count	Risk allele for reduction in hemoglobin
ABCG2	rs2231142	Ref	N/A	N/A
CYP3A5	rs776746	N/A	Ref	N/A
SLCO1B3	rs2053098	N/A	Var	N/A
	rs4149117	N/A	Ref	N/A
	rs3764006	N/A	N/A	N/A
ABCC1	rs8056298	N/A	N/A	Ref
	rs212091	N/A	N/A	Ref

N/A, not applicable; Ref, reference allele; Var, variant allele.

SNP rs2231142 is present in ABCG2 that belongs to the subfamily G of ATP binding cassette of transporters. ABCG2 is involved in the efflux of docetaxel form the hepatocytes into the bile canaliculi (Oshiro et al., 2005). ABCG2 C421A polymorphism is a missense mutation in the fifth exon of the gene ABCG2, and it results in the substitution of glutamine for lysine (Campa et al., 2011). Investigation of the functional effect of this amino acid change suggests that this polymorphism is correlated with a lower ABCG2 expression and hence increased drug accumulation (Imai et al., 2002; Kondo et al., 2004; Morisaki et al., 2005). Therefore, carriers of the polymorphic allele are expected to have decreased docetaxel efflux and more toxicity from docetaxel. This is inconsistent with our results. Nevertheless, these functional studies were done in cell lines that might not reflect the human physiology (Imai et al., 2002), and some showed inconsistencies regarding the effect of the polymorphism on the transporter ATPase activity (Kondo et al., 2004; Mizuarai et al., 2004; Morisaki et al., 2005). Moreover, some studies did not show any effect of the amino acid change on ABCG2 protein or mRNA expression in the intestine or heart (Zamber et al., 2003; Meissner et al., 2006). Finally, the SIFT program suggests a “tolerated” effect of this nonsynonymous polymorphism (Chelala et al. 2009). Hence, further research is warranted.

Other SNPs, rs8056298 and rs212091, are present in the 3’-UTR locus of the ABCC1 gene involved in the efflux of docetaxel from the liver to the bile (Oshiro et al., 2010). In our study, these SNPs were associated with a reduction in hemoglobin, a finding that was not previously reported (Deeken et al., 2010). These conflicting results may be attributed to differences in gender, tumor types, chemotherapy regimen, and toxicity endpoints among the studies.

We also observed in our study that SNP rs776746 in CYP3A5 gene that catalyzes metabolism of docetaxel to its inactive 4-hydroxy metabolite is associated with docetaxel-induced risk of reduction of neutrophil count, an observation also reported in the study of Tsai et al. (2009) done on breast cancer patients. This SNP is located in the intronic region of CYP3A5, hence suggesting that it may be in linkage disequilibrium (LD) with other variants that affect protein activity and docetaxel clearance.

SLCO1B3 c.1557G>A and c.334T>G variants were associated with reduction in neutrophil count after docetaxel treatment. SLCO1B3 transports docetaxel from the blood to the liver (Smith et al., 2005). SLCO1B3 c.1557G>A is a synonymous polymorphism, whereas SLCO1B3 c.334T>G is a nonsynonymous variant that results in amino acid change from serine to alanine at position 112. Some studies reported a reduced protein activity associated with the variant genotype (Nassif et al., 2012). However, according to the SNPnexus database, this structural variation is associated with a “tolerated” change in protein activity (Chelala et al., 2009). No other study reported an association between SLCO1B3 polymorphisms and docetaxel-induced myelosuppression (Baker et al., 2009). Research on the function of this variant is recommended to understand the findings in our study.

Interestingly, one SNP in FMO3 gene (rs909530) was found to be significantly associated with the three endpoints: febrile neutropenia, absolute neutrophil count, and Hb reduction. FMO3 encodes a member of flavin monooxygenase enzymes that metabolize a number of foreign chemicals, including tamoxifen and thiacetazone. Preferred substrates contain at the site of oxygenation a nucleophilic moiety such as nitrogen, sulfur, selenium, or phosphorous (Phillips et al., 2008). Though FMO3 is not known to be involved in the metabolism of docetaxel, an association between FMO3 rs909530 and large docetaxel area under the curve (AUC) was reported in a study by Uchiyama et al. (2012) conducted on 10 patients who received docetaxel for the treatment of ovarian or uterine cancer. Carriers of the mutant genotype c.855TT had a 67% increase in docetaxel clearance than carriers of the wild genotype c.855CC, with no increased risk of docetaxel-induced grade 4 neutropenia. However, a strong correlation between docetaxel pharmacokinetic parameters and febrile neutropenia is known whereby a decrease in docetaxel clearance by 50% increased the odds of experiencing docetaxel-induced grade 4 neutropenia by 3 (Baker et al., 2009; Longo et al., 2010). This SNP is a synonymous SNP that does not alter the amino acid sequence. It is hence possible that it is in LD with another one that may exacerbate docetaxel-induced myelosuppression.

This study is hypothesis generating in nature, and provides a number of important leads for future candidate pharmacogenomics biomarkers of docetaxel toxicity in a previously understudied population from Lebanon. It is noteworthy that our study and the approach utilized herein are, however, comparable to, and consistent with, other studies reported in the literature. For example, Deeken et al. (2010) evaluated an older version of the DMET platform with 47 prostate cancer patients, 33 of whom were treated with docetaxel and thalidomide and 14 with docetaxel alone. Furthermore, Tsai et al. (2009) reported an association of ABCB1 2677G/C polymorphism with febrile neutropenia in a sample of 59 Taiwanese women with breast cancer receiving a combination of docetaxel, epirubicin, and cyclophosphamide treatment, of whom 6 had developed febrile neutropenia. In addition, Tran et al. (2006) reported an association of GSTP1*A/*B with febrile neutropenia in a sample of 58 French patients with different types of solid tumors, 6 of whom developed febrile neutropenia. Note that in our study cases and controls were matched by treatment type and docetaxel dose. Also, comparison of the myelosuppression variables after docetaxel injection in cases and controls revealed that patients who developed febrile neutropenia (cases) developed a lower white blood cell count and absolute neutrophil count than patients who did not develop febrile neutropenia on docetaxel (controls), with p=0.097 and p=0.075, respectively, hence asserting that febrile neutropenia was an adequate toxicity outcome for evaluation (Table 2). Because of the small sample size, many SNPs involved in the analysis were found to be monomorphic or with a very low minor allele frequency.

We wish to underscore for the reader that the selected study population is not necessarily representative of all subjects with febrile neutropenia after docetaxel. Our study was based on a retrospective chart review; so we were not able to have complete data for every patient regarding several variables such as liver function tests and other docetaxel toxicities, including gastrointestinal or neurosensory adverse events. In addition, although baseline characteristics such as tumor type, grade, and disease stage (Table 1) were similar between both groups, additional confounders may have interfered with the onset of febrile neutropenia to include different epidemiological factors (socioeconomic status and diet) and in some patients, prior chemotherapy cycles of cyclophosphamide with an anthracycline and/or 5-fluorouracil.

Indeed, larger cross sectional or prospective studies are needed to evaluate the genetic factors that are linked to the development of febrile neutropenia after docetaxel in a sample representative of all subjects with febrile neutropenia after docetaxel. It is possible though that further genetic studies may fail to replicate our findings due to ethnic differences, in addition to the limitations mentioned above. Further, since docetaxel-induced febrile neutropenia is likely to be caused by the nonlinear interactions of numerous genetic and environmental risk factors, a multi-factorial approach is necessary to study risk interactions.

Conclusions

This is a study in a fairly large sample of 100 patients from which we identified a case control group for a low and moderate frequency side effect for pharmacogenomics research. This study identified SNPs that may be important determinants of docetaxel toxicity. Twelve single nucleotide polymorphisms (SNPs) in 7 genes (ABCC6, ABCG1, ABCG2, CYP1A2, CYP2D6, FMO2, and FMO3) were significantly associated with febrile neutropenia after docetaxel treatment, and one SNP in the FMO3 gene (rs909530) was found to be significantly associated with three endpoints: febrile neutropenia, absolute neutrophil count, and Hb reduction on cycle 2 of docetaxel treatment. Importantly, many of these SNPs have not been previously reported to be associated with toxicity attendant to docetaxel treatment. Further studies are required to characterize the functions of these genetic variants and to confirm our findings in a larger and more representative number of patients.

Genomics research is playing an important role in low and middle income countries (LMICs) (Sustainable Sciences Institute, 2013). Studies such as association of genetic variants with hypertension or type 2 diabetes in African Americans (Ehret et al., 2011; Palmer et al., 2012), prevalence of tropical neglected diseases (Harris et al., 1998; Balmaseda et al., 2003; Hotez et al., 2012) and development of vaccines against them (Goud et al., 2012; Groat-Carmona et al., 2012) provide a better understanding of noncommunicable diseases and reveal novel pathways for further investigation. Our study also lends evidence for the promise of data-intensive genomics research in developing countries.

To the best of our knowledge, this is the first study that evaluates the effect of a large array of nearly 2000 polymorphisms in DMEs and transporters on docetaxel toxicity in breast cancer patients. We report a number of genetic variants that may be associated with myelosuppression in Lebanese breast cancer patients treated with docetaxel. In light of the current global epidemic of noncommunicable diseases (NCDs) such as breast cancer impacting LMICs in addition to infectious diseases, we suggest pharmacogenomics should be considered as an integral part of the global health research agenda for NCDs and personalized medicine (Editorial, 2011).

Footnotes

Acknowledgments

The present study was funded by the American University of Beirut Medical Practice Plan. We are grateful to Dr. Joseph Simaan for reviewing our manuscript.

Author Disclosure Statement

No competing financial interests exist.

References

Baker

, Verweij

, Cusatis

et al. 2009. Pharmacogenetic pathway analysis of docetaxel elimination. Clin Pharmacol Therapeut, 85:155–163.

Balmaseda

, Guzmán

, Hammond

et al. 2003. Diagnosis of dengue virus infection by detection of specific immunoglobulin M (IgM) and IgA antibodies in serum and saliva. Clin Diag Lab Immunol, 10:317–322.

Bosch

, Huitema

, Doodeman

et al. 2006. Pharmacogenetic screening of CYP3A and ABCB1 in relation to population pharmacokinetics of docetaxel. Clin Cancer Res, 12:5786–5793.

Campa

, Butterbach

, Slager

et al. 2011. A comprehensive study of polymorphisms in the ABCB1, ABCC2, ABCG2, NR1I2 genes and lymphoma risk. Intl J Cancer, 131:803–812.

Cawley

, Di

, Hubbell

et al. 2006. BRLMM: An improved genotype calling method for Affymetrix® Mapping arrays. http://www.ashg.org/genetics/abstracts/abs06/f1577.htm. 2013 January 7.

Chelala

, Khan

, Lemoine

. 2009. SNPnexus: A web database for functional annotation of newly discovered and public domain single nucleotide polymorphisms. Bioinformatics, 25:655–661.

Chen

, Liu

, Relling

. 1996. Simultaneous characterization of glutathione S-transferase M1 and T1 polymorphisms by polymerase chain reaction in American whites and blacks. Pharmacogenetics, 6:187–191.

Clarkson

, Fan

, Joe

. 1993. A remark on algorithm 643: FEXACT: An algorithm for performing Fisher's exact test in r x c contingency tables. ACM Trans Math Software, 19:484–488.

Cui

, Li

, Yu

et al. 2013. Combination of low-dose docetaxel and standard-dose S-1 for the treatment of advanced gastric cancer: Efficacy, toxicity, and potential predictive factor. Cancer Chemother Pharmacol, 71:145–152.

10.

Deeken

. 2009. The Affymetrix DMET platform and pharmacogenetics in drug development. Curr Opin Mol Therapeut, 11:260–268.

11.

Deeken

, Cormier

, Price

et al. 2010. A pharmacogenetic study of docetaxel and thalidomide in patients with castration-resistant prostate cancer using the DMET genotyping platform. Pharmacogenom J, 10:191–199.

12.

DMET user guide. http://media.affymetrix.com/support/downloads/manuals/dmet_user_guide.pdf. 2012 February 2.

13.

Docetaxel FDA label. http://www.accessdata.fda.gov/drugsatfda_docs/label/2012/201525s002lbl.pdf. 2012 May 5.

14.

Editorial (no authors listed). 2011. Two days in New York: Reflections on the UN NCD summit. Lancet Oncol, 12:981.

15.

Ehret

, Munroe

, Rice

et al. 2011. Genetic variants in novel pathways influence blood pressure and cardiovascular disease risk. Nature, 478:103–109.

16.

Goud

, Gupta

, Hotez

et al. 2012. Expression, purification, and molecular analysis of the Necator americanus glutathione S-transferase 1 (Na-GST-1): A production process developed for a lead candidate recombinant hookworm vaccine antigen. Protein Express Purif, 83:145–151.

17.

Green

, Hortobagye

. 2007. Breast Cancer, 2nd. Springer: New York.

18.

Groat-Carmona

, Orozco

, Friebe

, Payne

, Kramer

, Harris

. 2012. A novel coding-region RNA element modulates infectious dengue virus particle production in both mammalian and mosquito cells and regulates viral replication in Aedes aegypti mosquitoes. Virology, 25:511–526.

19.

Hardenbol

, Banér

, Jain

et al. 2003. Multiplexed genotyping with sequence-tagged molecular inversion probes. Nature Biotechnol, 21:673–678.

20.

Harris

, Guy Roberts

, Smith

et al. 1998. Typing of dengue viruses in clinical specimens and mosquitoes by single-tube multiplex reverse transcriptase PCR. J Clin Microbiol, 10:317–322.

21.

Hotez

, Asojo

, Adesina

. 2012. Nigeria: “Ground Zero” for the prevalence neglected tropical diseases. PLoS Neglected Trop Dis, 6:e1600.

22.

Imai

, Nakane

, Kage

et al. 2002. C421A polymorphism in the human breast cancer resistance protein gene is associated with low expression of Q141K protein and low-level drug resistance. Mol Cancer Therap, 1:611–616.

23.

Kiyotani

, Mushiroda

, Kubo

, Zembutsu

, Sugiyama

, Nakamura

. 2008. Association of genetic polymorphisms in SLCO1B3 and ABCC2 with docetaxel-induced leukopenia. Cancer Sci, 99:967–972.

24.

Kobayashi

, Yokoyama

, Nakamura

, Nakajima

. 2012. A case of locally advanced breast cancer for which local control was obtained by combination therapy. Gan To Kagaku Ryoho, 39:2054–2056.

25.

Kondo

, Suzuki

, Itoda

et al. 2004. Functional analysis of SNPs variants of BRCP/ABCG2. Pharmaceut Res, 21:1895–1903.

26.

Longo

, D'Andrea

, Sarmiento

, Gasparini

. 2010. Pharmacogenetics in breast cancer: Focus on hormone therapy, taxanes, trastuzumab and bevacizumab. Expert Opin Invest Drugs, 19:41–50.

27.

Mehta

, Patel

. 1986. Algorithm 643: FEXACT: A Fortran subroutine for Fisher's exact test on unordered r*c contingency tables. ACM Transact Math Software, 12:154–161.

28.

Meissner

, Heydrich

, Jedlitschky

et al. 2006. The ATP-binding cassette transporter ABCG2 (BCRP), a marker for side population stem cells, is expressed in human heart. J Histochem Cytochem, 54:215–221.

29.

Milanowski

, Szmygin-Milanowska

. 2013. Treatment of non-small cell lung cancer—Where we are? Pneumonol Alergol Pol, 81:55–60.

30.

Montero

, Fossella

, Hortobagyi

, Valero

. 2005. Docetaxel for treatment of solid tumors: A systematic review of clinical data. Lancet Oncol, 6:229–239.

31.

Morisaki

, Robey

, Ozvegy-Laczka

et al. 2005. Single nucleotide polymorphisms modify the transporter activity of ABCG2. Cancer Chemother Pharmacol, 56:161–172.

32.

Nassif

, Jia

, Keiser

et al. 2012. Visualization of hepatic uptake transporter function in healthy subjects by using gadoxetic acid–enhanced MR imaging. Radiology, 264:741–750.

33.

National Cancer Institute toxicity grading. http://www.download.bham.ac.uk/bctu/AML16/Trial/DocsForExistingCentres/National%20Cancer%20Institution_NCI%20Toxicity%20Criteria%20formatted.pdf. 2012 December 22.

34.

Oshiro

, Marsh

, McLeod

, Carrillo

, Klein

, Altman

. 2010. Pharmacogenetics and Genomics Knowledge Base. www.pharmgkb.org/pathway/PA154426155. 2013 January 1.

35.

Palmer

, McDonough

, Hicks

et al. 2012. A genome-wide association search for type 2 diabetes genes in African Americans. Plos One, 7:e29202.

36.

Petranyi

. 2012. The treatment of castration-resistant prostate cancer. Hung Oncol, 56:219–228.

37.

Phillips

, Shephard

. 2008. Flavin-containing monooxygenases: Mutations, disease and drug response. Trends Pharmacol Sci, 29:294–301.

38.

Shou

, Martinet

, Korzekwa

, Krausz

, Gonzalez

, Gelboin

. 1998. Role of human cytochrome P450 3A4 and 3A5 in the metabolism of taxotere and its derivatives: Enzyme specificity, Interindividual distribution and metabolic contribution in human liver. Pharmacogenetics, 8:391–401.

39.

Sissung

, Danesi

, Price

et al. 2008. Association of the CYP1B1*3 allele with survival in patients with prostate cancer receiving docetaxel. Mol Cancer Therap, 7:19–26.

40.

Sissung

, English

, Venzon

, Figg

, Deeken

. 2010. Clinical pharmacology and pharmacogenetics in a genomic era: The DMET Platform. Pharmacogenomics, 11:89–103.

41.

Smith

, Acharya

, Desai

, Figg

, Sparreboom

. 2005. Identification of OATP1B3 as a high-affinity hepatocellular transporter of paclitaxel. Cancer Biol Ther, 4:815–818.

42.

Sorbe

, Graflund

, Horvath

et al. 2012. Phase II study of docetaxel weekly in combination with carboplatin every 3 weeks as first-line chemotherapy in stage IIB to stage IV epithelial ovarian cancer. Intl J Gynecol Cancer, 22:47–53.

43.

Sparreboom

, Van Tellingen

, Scherrenburg

et al. 1996. Isolation, purification and biological activity of major docetaxel metabolites from human feces. Drug Metab Dispos, 24:655–658.

44.

Sustainable Sciences Institute. http://www.sustainablesciences.org/History. 2013 Feb 21.

45.

Tran

, Jullien

, Alexandre

et al. 2006. Pharmacokinetics and toxicity of docetaxel: Role of CYP3A, MDR1, and GST polymorphisms. Clin Pharmacol Therap, 79:570–580.

46.

Trastuzumab FDA label. http://www.fda.gov/downloads/Drugs/DevelopmentApprovalProcess/HowDrugsareDevelopedandApproved/ApprovalApplications/TherapeuticBiologicApplications/ucm092760.pdf. 2012 December 12.

47.

Tsai

, Lin

, Wu

et al. 2009. Side effects after docetaxel treatment in Taiwanese breast cancer patients with CYP3A4, CYP3A5, and ABCB1 gene polymorphisms. Clin Chim Acta, 404:160–165.

48.

Uchiyama

, Kanno

, Ishitani

et al. 2012. An SNP in CYP39A1 is associated with severe neutropenia induced by docetaxel. Cancer Chemother Pharmacol, 69:1617–1624.

49.

Vandenberg

, Younus

, Al-Khayyat

. 2010. Febrile neutropenia rates with adjuvant docetaxel and cyclophosphamide chemotherapy in early breast cancer: Discrepancy between published reports and community practice—A retrospective analysis. Curr Oncol, 17:2–3.

50.

Zamber

, Lamba

, Yasuda

et al. 2003. Natural allelic variants of breast cancer resistance protein (BCRP) and their relationship to BCRP expression in human intestine. Pharmacogenetics, 13:19–28.

51.

Zuylen

, Verweij

, Nooter

, Brouwer

, Stoter

, Sparreboom

. 2005. Role of intestinal P-glycoprotein in the plasma and fecal disposition of docetaxel in humans. Clin Cancer Res, 6:2598–2603.