Pharmacokinetics and pharmacogenetics of Gemcitabine as a mainstay in adult and pediatric oncology: an EORTC-PAMM perspective
Gemcitabine is an antimetabolite ranking among the most prescribed anticancer drugs worldwide. This nucleoside analog exerts its antiproliferative action after tumoral conversion into active triphosphorylated nucleotides interfering with DNA synthesis and targeting ribonucleotide reductase. Gemcitabine is a mainstay for treating pancreatic and lung cancers, alone or in combination with several cytotoxic drugs (nab-paclitaxel, cisplatin and oxaliplatin), and is an option in a variety of other solid or hematological cancers. Several determinants of response have been identified with gemcitabine, i.e., membrane transporters, activating and inactivating enzymes at the tumor level, or Hedgehog signaling pathway. More recent studies have investigated how germinal genetic polymorphisms affecting cytidine deaminase, the enzyme responsible for the liver disposition of gemcitabine, could act as well as a marker for clinical outcome (i.e., toxicity, efficacy) at the bedside. Besides, constant efforts have been made to develop alternative chemical derivatives or encapsulated forms of gemcitabine, as an attempt to improve its metabolism and pharmacokinetics profile. Overall, gemcitabine is a drug paradigmatic for constant searches of the scientific community to improve its administration through the development of personalized medicine in oncology.
KeywordsGemcitabine 2′,2′-difluoro-2′-deoxyuridine Cytidine deaminase Polymorphisms Pharmacogenetics Gemcitabine prodrugs
Among the pyrimidine analogs, gemcitabine (2′,2′-difluorodeoxycytidine, dFdC; Gemzar®) is one the most widely used drugs in clinical oncology and ranked the third anticancer agent prescribed worldwide. It is a cytidine analog, where two fluorine atoms have replaced the hydroxyl on the ribose. In particular, gemcitabine is a mainstay in pancreatic adenocarcinoma [1, 2] and is widely prescribed to treat a variety of other solid tumors such as breast, ovarian, bladder or non-small-cell lung (NSCLC) cancers [3, 4]. In addition to solid tumors, gemcitabine is indicated as well in several hematological disorders such as acute leukemia . Beyond adult patients, gemcitabine can be an attractive option in pediatric cancers because its toxic profile is usually considered as mild as compared with other cytotoxic drugs.
Because gemcitabine is the backbone of numerous regimens, several studies have tried to identify molecular or genetic determinants of response, both at the somatic and the constitutional levels . In addition, recent efforts have been made to improve the metabolism and pharmacokinetics (DM-PK) profile of gemcitabine, creating novel chemical derivatives, prodrugs or nanomedicine forms .
Gemcitabine pharmacokinetics and pharmacodynamics
Because of its hydrophilic nature, gemcitabine does not readily cross the membrane by diffusion, and it is transported into the cells by membrane nucleoside transporters . Following cellular uptake, gemcitabine is phosphorylated to its active diphosphate (dFdCDP) and triphosphate (dFdCTP) metabolites, which inhibit RR and DNA synthesis, respectively . dCK is the rate-limiting enzyme in the biotransformation of nucleoside analogs, and several studies have suggested that dCK is a limiting factor for gemcitabine activity, because its deficiency/modulation is critically involved in acquired resistance in different in vitro models [13, 14]. Moreover, pretreatment dCK expression level could be used as a predictive parameter of tumor sensitivity, as observed with a clear correlation between dCK activity and gemcitabine sensitivity in tumor cells and xenografts .
The dynamics of dFdCDP and dFdCTP formation and activity in vivo are complex; dFdCTP is incorporated into DNA followed by one or more deoxynucleotides masking gemcitabine and preventing DNA repair by 3′5′-exonuclease activity, a process designated as “masked DNA chain termination” . This causes an S-phase-specific cell cycle arrest and programmed cell death. dFdCTP is also incorporated into RNA, thus inhibiting RNA synthesis , while dFdCDP inhibits RR, inducing a depletion of the cellular pool of deoxynucleoside triphosphates, and blocks the de novo DNA synthesis pathway .
However, only a proportion of gemcitabine is converted into the active di- or triphosphate forms. The majority of gemcitabine is rapidly inactivated in the liver and to a lesser extent in blood by deamination into dFdU, through a reaction catalyzed by CDA. Additionally, 10 % of unchanged gemcitabine can undergo renal filtration, and within 1 week, more than 90 % of the injected dose is usually recovered in the urine, either as parent gemcitabine (1 %) or dFdU (99 %) . In addition, the formation of dFdCTP and dFdCDP from dFdCMP is reduced through deamination of dFdCMP to 2′,2′-difluorodeoxyuridine monophosphate (dFdUMP) by dCMP deaminase. Notably, an elevated concentration of dFdCTP inhibits dCMP deaminase, determining a “self-potentiation” of the drug activity , which is also caused by the increase in dFdCTP accumulation induced by dFdU in a time-dependent manner . dFdCTP also inhibits CTP synthetase, affecting RNA synthesis by depletion of CTP, while the latter decreases dCTP synthesis [22, 23]. Finally, a recent study demonstrated that gemcitabine can inhibit the enzyme thymidylate synthase presumably through the phosphorylated metabolite dFdUMP. Inhibition of this enzyme enhances the mis-incorporation of 2′-deoxyuridine into DNA, causing indirect damage .
A considerable inter-patient variability has been described in gemcitabine accumulation, and the pharmacokinetics of gemcitabine and its main metabolite dFdU in plasma have been evaluated in multiple studies. Gemcitabine plasma concentrations generally reach a plateau after 15–30 min during the standard 30 min infusion protocol. Linear pharmacokinetics have been described over the range 40–3650 mg/m2, and nonlinear pharmacokinetics at higher doses [19, 25, 26]. Mean gemcitabine peak plasma concentrations ranged from 24 μM at 800 mg/m2  to 32 μM at 1000 mg/m2 , around 53–70 μM at a dose of 1250 mg/m2 [29, 30], 68–79 µM at 2350 mg/m2 and between 320 and 512 µM at the MTD of 5700 mg/m2 . Up to at least at gemcitabine 1250 mg/m2, deamination was linear with mean plasma dFdU concentrations being 1.25 times higher as compared to dFdU levels using gemcitabine 1000 mg/m2. Linearity was lost at doses higher than 3650 mg/m2 . The clearance of gemcitabine in the plasma is also rapid (i.e., T1/2 of 5–20 min). More than 75 % of gemcitabine is metabolised to dFdU and excreted in the urine in the first 24 h . This clearance is independent of dose over the linear range (i.e., up to 3650 mg/m2), but proportional to creatinine clearance. At the highest doses, the clearance was lower; moreover, the clearance was 1.5-fold higher in men (8.6 l/m2) compared to women (5.7 l/m2) . The pharmacokinetic elimination half-life for dFdU varies between 2 and 24 h, and it is still present systemically in concentrations greater than 1 μM up to 1 week after dosing . Of note, since dFdU is not protein bound, its plasma concentration, up to 460 μM , depending on the dose administered, is freely available. These concentrations are cytotoxic [20, 32] and could have significant implications in the clinical use of gemcitabine alone or in combination with other therapies, such as radiation , since dFdU has a radiosensitizing effect by itself.
Since gemcitabine is often given in combination with other cytotoxic and targeted drugs, the effect of combination therapy on the pharmacokinetics has been investigated in several clinical studies, since theoretically co-medication can affect both drug metabolism and elimination. However, in the most widely used combination with cisplatin or paclitaxel, no evidence was found for an effect of these drugs (as well as oxaliplatin and carboplatin) on both gemcitabine and dFdU pharmacokinetics, investigated within the same patients and between patients [27, 28, 30, 34, 35]. Similarly, the proteasome inhibitor bortezomib and the farnesyltransferase inhibitor SCH66336 did not affect the pharmacokinetics of gemcitabine or dFdU, alone or in the combination with cisplatin [36, 37]. Moreover, no effect of the VEGFR inhibitor SU5416 was observed , while the EGFR inhibitor gefitinib tended to increase the exposure to gemcitabine . However, the other EGFR inhibitor erlotinib did not affect pharmacokinetics of the gemcitabine prodrug LY2334737 itself or of gemcitabine and dFdU . Hence, from the point of view of pharmacokinetics, it can be concluded that in general gemcitabine can safely be combined with other drugs, both other cytotoxics and novel targeted drugs. Naturally, this does not exclude that gemcitabine affects the mechanism of action of other drugs or that these other drugs affect intracellular metabolism of gemcitabine. Two examples include the potentiation by gemcitabine of cisplatin adduct formation and the selective effect of bortezomib on intracellular gemcitabine activation [27, 40].
Less data are available on the pharmacokinetics of dFdCTP, which should be measured with more sensitive LCMS assays . However, several studies demonstrated that cells exposed to gemcitabine have saturable accumulation of the dFdCTP, and the optimal plasma concentration of gemcitabine that maximized the rate of formation of dFdCTP was approximately 20 μmol/l [42, 43]. This is accompanied by a change in the pattern of elimination with monophasic elimination at low concentrations and biphasic elimination described after the threshold has been reached.
Since the optimal intracellular accumulation of dFdCTP was achieved with dose rates of 10 mg/m2/min [25, 43], a number of phase I trials have explored the possibility to prolong the duration of infusion time, while other trials escalated both the dose and infusion duration [25, 44, 45]. The rationale for prolonged dosing received a major boost when a randomised phase II trial in a clinically relevant scenario (pancreatic cancer) demonstrated that prolonged infusion at a rate of 10 mg/m2/min, compared to the standard dosing regimen with 30 min infusion, was associated with increased accumulation of dFdCTP, as well as with a significant increase in response rate and a trend for increased survival . Similar trials in different tumor types confirmed the pharmacokinetic finding, but were underpowered to demonstrate survival differences . Unfortunately, a large phase III study in pancreatic cancer showed that the pharmacological advantage failed to translate into a significant survival advantage .
Collectively, these clinical studies indicate that the anti-tumor effect of gemcitabine is schedule dependent and that lower doses can be efficacious. Therefore, it could be advantageous to deliver gemcitabine in a manner where it can achieve prolonged systemic exposure, good efficacy with lower toxicity along with added flexibility of administration and greater patient convenience, such as using an oral formulation . However, administering gemcitabine orally to patients has been limited by low oral bioavailability, high first-pass clearance, variable systemic exposures during dose escalation studies and observation of gastrointestinal toxicity including nausea, vomiting and diarrhea.
Dysregulation at the germinal level: pharmacogenetics of gemcitabine
Factors extracted from either clinical or pathological data such as age, performance status, comorbidity and disease stage or grade provide a crude discrimination of prognosis, but are often not predictive and not helpful for the choice of the best chemotherapeutic regimen for a given patient. Novel approaches to stratify patient’s prognosis or toxicity may be offered by pharmacogenetic analyses of selected candidate polymorphisms that could influence the expression of genes involved in drug metabolic pathways.
Historically, pharmacogenetics is indeed defined as the study of germline mutations (e.g., single-nucleotide polymorphisms affecting genes coding for enzymes responsible for drug pharmacokinetics), whereas pharmacogenomics refers to the role of both acquired and inherited genetic differences in relation to drug behavior through a systematic examination of genes, gene products and inter- and intra-individual variation in gene expression and function using new genomic technologies . However, in oncology, pharmacogenetics is often considered as concerning the individual patient’s features and pharmacogenomics as those of the tumor.
CDA deregulations and clinical outcome
Gemcitabine is primarily detoxified in the liver by CDA into dFdU, with a Km of approximatively 96 µM [50, 51]. Usually, 90 % of gemcitabine is detoxified by CDA, and variations in enzymatic activity impact greatly on drug pharmacokinetics and pharmacodynamics. Mice with impaired CDA displayed sharp overexposure to the drug with subsequent unrecoverable hematological toxicities , thus highlighting the correlation between CDA deficiency, overexposure to gemcitabine and increased risk of severe toxicities. Indeed, a variety of studies and case reports have found a correlation between CDA deficiency syndrome and an increase in severe hematological toxicities in patients undergoing gemcitabine-based therapy [52, 53]. Of note, the first-ever reported case of toxic death related to CDA deficiency in an ovarian cancer patient treated with the gemcitabine—carboplatin was published in 2007 . Profound functional deficiency was retrospectively evidenced, with heterozygous CDA*2 genotype. Subsequent genetic investigations revealed a new intronic mutation (i.e., 154 + 37G>A) on the CDA gene, likely to have caused the lethal toxicities . Of note, other studies have shown that patients with lower CDA activity also tend to display higher response rates and better survival [55, 56]. On the contrary, it has been observed that about 15 % of the Caucasian adult population display CDA activities significantly higher than the median values of adult populations (i.e., over 6 U/mg), making them prone to therapeutic failure because most of standard dosing of gemcitabine will be metabolized in the liver before it even reaches the tumor tissues . A pilot study involving 40 patients treated by gemcitabine-based regimens for pancreatic cancer confirmed that patients displaying CDA ultrametabolizer phenotype were fivefold more at risk to have a progressive disease than patients with normal CDA status . As expected, these patients had milder toxicities than patients with normal or lower CDA activity, an observation completely in line with previous reports about CDA and gemcitabine-related toxicities. Overall, all these studies, conducted by independent groups and involving patients treated with gemcitabine used alone or in a combination for a variety of settings, demonstrate how CDA status greatly affects clinical outcome in patients undergoing gemcitabine-based treatments.
CDA genetic polymorphisms
CDA polymorphisms and their differential functional and clinical effects
Different activity (increased or reduced) among haplotypes 
Gemcitabine induced neutropenia 
79 A>C (Lys27Gln)
208 G>A (Ala70Thr)
Reduced activity 
435 C>T (Thr145Thr)
Reported in Asians, Africans and Caucasians 
Unaltered activity 
CDA 435 C/C associated with better response and progression-free survival while C/T with a significantly increased risk of non-hematological toxicity 
Indeed, the large inter-individual variability reported with CDA activity is only partly explained by the genetic background. In addition, because more than 1000 genetic variations have been evidenced, SNP-candidate studies are probably underpowered strategies, yielding conflictual data [51, 55, 64, 68, 71, 72]. For instance, the 79A>C polymorphism (i.e., CDA*2) leads to lysine to glutamine permutation in position 27, with no impact on the catalytic site eventually, but other factors might play a role. Of note, ethnicity plays a crucial role in the allelic frequencies of this variant because minor allele frequency (MAF) ranges from 10 % in African population, 15 % in Asian population, but up to 35 % in Caucasians [73, 74]. The phenotypic impact of this allelic variant and its consequence in the clinical outcome in patients treated with nucleoside analogs remain controversial: a decrease in CDA activity has been measured for the Lys27Lys variant [75, 76], whereas other studies suggest no variation [52, 77] or lower activity for the Gln27Gln variant [78, 79]. These differences may be partly explained by variations in study design such as patient selection, ethnicity and treatment regimens [80, 81, 82]. A pivotal study has recently been published, collecting data about CDA catalytic activity according to substrates, both natural and synthetic . This biochemical study, inspired from a previous work in 2012 by Baker and collaborators , highlights that catalytic efficiency of CDA enzyme is dependent on the genetic sequence encoding for the protein, but also on the drugs used as substrates. Surprisingly, an increase in CDA catalytic efficiency was observed for the CDA27Gln protein with natural cytidine analogs and cytarabine, but surprisingly a decrease was found for other substrates such as 5-azacytidine, 6-azacytidine and fazarabine. This is in agreement with the results of the study by Giovannetti et al. , in which CDA27Gln activity was investigated with gemcitabine, with a decrease in deaminase activity being observed with this polymorphism.
In addition to the CDA*2 allelic variant, another polymorphism has been studied extensively: CDA*3, resulting from the substitution of alanine to threonine in position 70 because of the 208G>A SNP in the coding sequence. With the CDA*3 variant, impact on CDA phenotype is more univocal because researchers all agree that a decrease in CDA activity is found for the protein encoded by 208 A/A variant [68, 74, 83]. Indeed, deaminase activity was found to be 100-fold lower than normal CDA with respect to all tested drugs with this allelic variant . Of note and unlike CDA*2, which is found in every population, but in different proportions, CDA*3 has never been detected in Caucasian populations, but is only found in Africans and Asians. To date, the clinical impact of CDA*3 genotype in patients treated with gemcitabine has been repeatedly reported in Japanese patients only [74, 83, 84]. In addition, two studies have aimed at establishing the respective MAF of 79A>C, 208G>A and 435C>T in both Asian and Caucasian populations [67, 85]. Few differences were observed in MAF of 435C>T allelic variant when comparing these two ethnicities, whereas discrepancies were evidenced for the 79A>C and 208G>A variants. For the CDA*2 allelic variant, twice as many individuals carry a wild-type genotype with a lower incidence of C/C genotype in Asians, as compared with white people. The discrepancy is more marked for the rs60369023 variant because no individual, whether in African-Americans, Chinese-Americans, or Caucasian-Americans, was carrying the minor allele A. Only 11 patients were heterozygous among over 400 Korean and Japanese patients, and none of them was found to be homozygous for the CDA*3 variant . These data confirm that screening for 208G>A single-nucleotide polymorphism has a clinical, yet limited, meaning in Asian or more significantly in Japanese populations only. As mentioned above, numerous other genetic variations have been identified, but no study has established a clear link between a given genotype and the resulting phenotype yet, apart for the −31delC variant (rs532545), a CDA promoter deletion possibly resulting in an amplification of the CDA gene with functional (i.e., ultrametabolizer phenotype) impact eventually . These data call for more sophisticated multigenic or haplotype-based studies to establish a clear genotype-to-phenotype relationship with CDA and gemcitabine.
Dysregulation at the tumor level: pharmacogenomics of gemcitabine
Several determinants for efficacy have been identified with gemcitabine at the tumor level. Because gemcitabine requires facilitated transport for cellular uptake , several studies evaluated the expression levels of the plasma membrane human equilibrative nucleoside transporter-1 (hENT1) and human concentrative nucleoside transporter-3 (hCNT3), showing their prognostic and predictive roles of drug activity in patients undergoing gemcitabine-based regimen [86, 87, 88, 89]. Higher uptake in cancer cells with high levels of both transporters could explain the marked increase in disease-free survival and overall survival observed in pancreatic cancer patients administered with gemcitabine. Of note, several polymorphisms affecting the genes coding for hENT1 and hCNT3 have been described. These polymorphisms might impact protein expression, but the functional and clinical significance of these polymorphisms have yet to be defined .
Other determinants for response at the tumor level include the expression of dCK, the rate-limiting enzyme activating the prodrug gemcitabine to active nucleotides , deoxycytidylate deaminase that metabolizes active phosphorylated nucleotides into inactive metabolites  and RR that is one of the gemcitabine targets . These genes have many polymorphisms, which could impact on drug efficacy, but the relevance of all these markers has not been fully confirmed at the bedside. Large prospective control studies are necessary to confirm the role of these determinants in response to gemcitabine.
However, the candidate-gene approach used in most of these studies cannot establish if a positive association is due to linkage with untyped functional variant alleles or due to intragene interaction. Drug efficacy and toxicity may also be influenced by other genes and pathways, which will be undetected by single-polymorphism analysis. Therefore, alternative approaches through the broader application of new genomic technologies might be necessary to identify novels biomarkers of gemcitabine efficacy and toxicity, and to bring us closer to tailor-made therapy for individual patients . Li et al.  used such a method to identify novel genes involved in gemcitabine metabolism. After analyzing data with 26,653 probe sets, the researchers identified 15 genes where mRNA expression correlated with cytidine analog sensitivity and, from there, selected FKBP5 for further functional variation. FKBP5 is a gene involved in steroid receptor mutation and is a binding partner for rapamycin, suggesting that it may have a role in the apoptosis pathway. Overexpression of FKBP5 correlated with increased gemcitabine chemosensitivity. Using FKBP5 siRNA-treated cells, the researchers showed a decrease in downstream enzyme caspase-3/7 activity, confirming that the activation of the apoptotic pathway was affected by the downregulation of FKBP5 expression.
Recently, it has also been shown that tumor blood perfusion and vessels’ density could be associated with response to drug-based therapy, probably in relation with a drug delivery issue. In this respect, the Hedgehog signaling pathway has been suggested as being a new determinant of response with gemcitabine, since it is associated with production of a desmoplastic tumor stroma and a reduced tumoral perfusion eventually, probably through the expression of Gli family transcriptor factors. As a result, Hedgehog signaling could prevent at least partly gemcitabine to be delivered to tumors [94, 95]. This rising concern in both experimental and clinical oncology about the role Hedgehog protein plays, highlights the fact that beyond pharmacological molecular determinants, the issue of drug delivery becomes more and more critical [96, 97]. This issue is currently addressed by recent efforts to develop “gemcitabine 2.0.” forms likely to display improved distribution and cellular uptake profiles.
In order to overcome various forms of drug resistance and/or to improve drug delivery, recent studies evaluated several gemcitabine prodrugs. Since many clinical studies showed that a low expression of hENT1 was associated with a poor survival of pancreatic cancer patients receiving gemcitabine [97, 98, 99], Clavis Pharma developed an elaidic acid prodrug modified at the 5′-sugar position, CP-4126, which was able to bypass hENT1. CP-4126 was similarly effective as gemcitabine in various model systems in vivo and showed an oral efficacy, possibly because CP-4126 could also inhibit CDA, preventing or reducing its first-pass effect . Because of its efficacy in phase II studies, CP-4126 (as CO-101) was tested in a phase II randomized, multicenter trial in comparison with gemcitabine as first-line therapy in metastatic pancreatic cancer patients . Mandatory tumor biopsy specimens were evaluated for their hENT1 expression in both treatment groups. The similar effect of CO-101 and gemcitabine demonstrated an effective conversion of the prodrug to gemcitabine. However, the study did not reach its anticipated endpoint, an increased efficacy of CO-101 compared to gemcitabine in the low hENT1 group of patients. Of note, no difference in survival was found also between the low and high hENT group treated with gemcitabine, suggesting that the role of hENT1 is less important in metastatic disease than after surgery, as shown in the patients in the earlier adjuvant studies. However, it is also possible that the lack of difference between the patients with low and high hENT1 expression was due to a lower specificity of the antibody, which was different from that used in earlier studies.
Another gemcitabine prodrug currently in clinical development is LY2334737, which has valproic acid attached to the N4 of the base. Valproic acid is cleaved off by carboxylesterase 2 (CES2), which is high in the liver and gastrointestinal tract, resulting in an early cleavage of the molecule, as well as in a better antitumor activity in tumors with a high CES2 activity. A phase I dose-finding study with oral administration (daily for 14 days, with 1 week rest) showed linear pharmacokinetics, until the maximal tolerated dose of 40 mg in Caucasian and 30 mg in Japanese patients, with a lower toxicity profile compared to oral gemcitabine [39, 100]. However, using other schedules (every other day for 21 days followed by 1 week rest, or daily for 7 days every other week), the maximal tolerated dose was 90 mg/m2, which was recommended for phase II studies for the every other day schedule .
Nucana developed another type of prodrug, NUC-1031, which is a gemcitabine analog to which a phosphoramidate ProTide moiety has been added. This novel nucleotide evades all three main cellular resistance mechanisms associated with gemcitabine (i.e., nucleoside transport, dCK-mediated activation and CDA-mediated degradation). NUC-1031 showed activity in cell culture and in in vivo models, including xenografts resistant to systemic gemcitabine treatment, while dFdCTP reaching tenfold higher levels in white blood cells than was found for gemcitabine at similar doses [102, 103]. More recently, NUC-1031 showed clear signs of clinical activity in patients with gynecological cancers. This agent was well tolerated, and a phase Ib study of NUC-1031 in combination with carboplatin is ongoing, while phase III studies are planned in both platinum sensitive and refractory gynecological cancers .
Another approach to increase delivery consists in the use nanoparticles that can be designed to allow controlled/sustained drug release . These systems are more stable than liposomes, but retain their low immunogenicity. Gemcitabine-loaded gold nanoparticles targeted to the epidermal growth factor receptor with cetuximab had an increased targeting and activity of gemcitabine in pancreatic tumors in vitro and in mouse tumor models . Similarly, gemcitabine covalently coupled with the natural lipid 1,1′,2-tris-nor-squalenic acid (squalene) at its 4-amino moiety, resulting in 4-(N)-tris-nor-squalenoyl-gemcitabine, which spontaneously assemble into a hexagonal structure with an aqueous core, was active in both human and murine leukemia resistant cell lines and tumors .
Gemcitabine can also be covalently coupled via polyethylene glycol (PEG) to another molecule that can target the complex to tumor cells, such as for PEG-gemcitabine conjugates to folic acid, binding specifically to the folate receptors, which are highly overexpressed on the surface of many cancers . Finally, gemcitabine can also be loaded in PEGylated liposomes .
Conclusions and perspectives
There are few drugs in oncology that are as old, but still so widely used as chemotherapeutic targets such as gemcitabine. Gemcitabine is indeed approved and commonly used, alone or in combination, for the treatment of several tumor types, such as NSCLC, pancreatic, bladder, ovarian and breast cancer. In order to improve its antitumor activity while reducing toxic effects, many studies investigated pharmacokinetics  and/or the impact of genetic polymorphisms and CDA activity [51, 55], as well as tumor-specific expression of hENT1 mRNA and protein [87, 88], on gemcitabine toxicity and efficacy. These factors appear to be the most promising predictive indicators of outcome in patients receiving gemcitabine chemotherapy [9, 10].
However, most pharmacogenetic studies were retrospective and monocentric, without multiple correction and validation in broader populations. Most phenotypic studies used different methods and specimens; for example, a number of assays have been used to determine CDA activity in various blood compartments . Moreover, most clinical trials on gemcitabine combinations were performed without previous preclinical studies evaluating molecular mechanisms and markers of drug synergistic interaction, while pharmacogenomics studies on tumor specimens did not evaluate tumor heterogeneity and possible evolution of cancer cells after tumor relapse, which should be faithfully documented within multiple samples of the single tumor as well as repeated biopsies.
Therefore, future efforts should be redirected at identifying, both in preclinical models and in the clinical setting, either sensitive or non-responding genotypes or phenotypes. These profiles should be identified with validated methods, which should be used for the appropriate patient enrollment into subsequent prospective studies.
Hopefully, in the near future, the availability of validated genetic/phenotypic platforms will lead to the selection of key factors responsible of the chemosensitivity and toxicity to gemcitabine-based treatments and guide in the choice of more effective rationally based tailor-made treatments for each patient.
This work was supported by grants from Cancer Center Amsterdam (CCA) Foundation (to GJP and EG), Istituto Toscano Tumori (ITT) and Associazione Italiana per la Ricerca sul Cancro (AIRC) (to EG).
Compliance with ethical standards
Conflict of interest
Joseph Ciccolini owns a patent on CDA testing (US 20130011392 A1).
- 1.Burris HA 3rd, Moore MJ, Andersen J, Green MR, Rothenberg ML, Modiano MR, Cripps MC, Portenoy RK, Storniolo AM, Tarassoff P, Nelson R, Dorr FA, Stephens CD, Von Hoff DD (1997) Improvements in survival and clinical benefit with gemcitabine as first-line therapy for patients with advanced pancreas cancer: a randomized trial. J Clin Oncol 15:2403–2413PubMedGoogle Scholar
- 4.Kroep JR, Giaccone G, Tolis C, Voorn DA, Loves WJ, Groeningen CJ, Pinedo HM, Peters GJ (2000) Sequence dependent effect of paclitaxel on gemcitabine metabolism in relation to cell cycle and cytotoxicity in non-small-cell lung cancer cell lines. Br J Cancer 83:1069–1076PubMedPubMedCentralCrossRefGoogle Scholar
- 8.van Riel JM, Peters GJ, Mammatas LH, Honeywell RJ, Laan AC, Ruyter R, van den Berg FG, Giaccone G, van Groeningen CJ (2009) A phase I and pharmacokinetic study of gemcitabine given by 24-h hepatic arterial infusion. Eur J Cancer 45:2519–2527. doi: 10.1016/j.ejca.2009.05.025 PubMedCrossRefGoogle Scholar
- 24.Honeywell RJ, Ruiz van Haperen VW, Veerman G, Smid K, Peters GJ (2015) Inhibition of thymidylate synthase by 2′,2′-difluoro-2′-deoxycytidine (Gemcitabine) and its metabolite 2′,2′-difluoro-2′-deoxyuridine. Int J Biochem Cell Biol 60:73–81. doi: 10.1016/j.biocel.2014.12.010 PubMedCrossRefGoogle Scholar
- 27.van Moorsel CJ, Kroep JR, Pinedo HM, Veerman G, Voorn DA, Postmus PE, Vermorken JB, van Groeningen CJ, van der Vijgh WJ, Peters GJ (1999) Pharmacokinetic schedule finding study of the combination of gemcitabine and cisplatin in patients with solid tumors. Ann Oncol 10:441–448PubMedCrossRefGoogle Scholar
- 28.Kroep JR, Giaccone G, Voorn DA, Smit EF, Beijnen JH, Rosing H, van Moorsel CJ, van Groeningen CJ, Postmus PE, Pinedo HM, Peters GJ (1999) Gemcitabine and paclitaxel: pharmacokinetic and pharmacodynamic inter- actions in patients with non-small-cell lung cancer. J Clin Oncol 17:2190–2197PubMedGoogle Scholar
- 29.Kuenen BC, Rosen L, Smit EF, Parson MR, Levi M, Ruijter R, Huisman H, Kedde MA, Noordhuis P, van der Vijgh WJ, Peters GJ, Cropp GF, Scigalla P, Hoekman K, Pinedo HM, Giaccone G (2002) Dose-finding and pharmacokinetic study of cisplatin, gemcitabine, and SU5416 in patients with solid tumors. J Clin Oncol 20:1657–1667PubMedCrossRefGoogle Scholar
- 30.Kroep JR, Smit EF, Giaccone G, Van der Born K, Beijnen JH, Van Groeningen CJ, Van der Vijgh WJ, Postmus PE, Pinedo HM, Peters GJ (2006) Pharmacology of the paclitaxel-cisplatin, gemcitabine-cisplatin, and paclitaxel-gemcitabine combinations in patients with advanced non-small cell lung cancer. Cancer Chemother Pharmacol 58:509–516PubMedCrossRefGoogle Scholar
- 32.Ruiz van Haperen VW, Veerman G, Boven E, Noordhuis P, Vermorken JB, Peters GJ (1994) Schedule dependence of sensitivity to 2′,2′-difluorodeoxycytidine (Gemcitabine) in relation to accumulation and retention of its triphosphate in solid tumour cell lines and solid tumours. Biochem Pharmacol 48:1327–1339PubMedCrossRefGoogle Scholar
- 34.Poplin E, Feng Y, Berlin J, Rothenberg ML, Hochster H, Mitchell E, Alberts S, O’Dwyer P, Haller D, Catalano P, Cella D, Benson AB 3rd (2009) Phase III, randomized study of gemcitabine and oxaliplatin versus gemcitabine (fixed-dose rate infusion) compared with gemcitabine (30-minute infusion) in patients with pancreatic carcinoma E6201: a trial of the Eastern Cooperative Oncology Group. J Clin Oncol 27:3778–3785. doi: 10.1200/JCO.2008.20.9007 PubMedPubMedCentralCrossRefGoogle Scholar
- 35.Faivre S, Le Chevalier T, Monnerat C, Lokiec F, Novello S, Taieb J, Pautier P, Lhommé C, Ruffié P, Kayitalire L, Armand JP, Raymond E (2002) Phase I-II and pharmacokinetic study of gemcitabine combined with oxaliplatin in patients with advanced non-small-cell lung cancer and ovarian carcinoma. Ann Oncol 13:1479–1489PubMedCrossRefGoogle Scholar
- 36.Voortman J, Smit EF, Honeywell R, Kuenen BC, Peters GJ, Van de Velde H, Giaccone G (2007) A parallel dose-escalation study of weekly and twice-weekly bortezomib in combination with gemcitabine and cisplatin in the first-line treatment of patients with advanced solid tumors. Clin Cancer Res 13:3642–3651PubMedCrossRefGoogle Scholar
- 37.Theodore C, Geoffrois L, Vermorken JB, Caponigro F, Fiedler W, Chollet P, Ravaud A, Peters GJ, Balincourt C, Fumoleau P (2005) Multicentre EORTC study 16997: feasibility and phase II trial of farnesyl transferase inhibitor and gemcitabine combination in salvage treatment of advanced urothelial tract cancers. Eur J Cancer 41:1150–1157PubMedCrossRefGoogle Scholar
- 38.Giaccone G, González-Larriba JL, Van Oosterom AT, Alfonso R, Smit EF, Martens M, Peters GJ, Van der Vijgh WJF, Smith R, Averbuch S, Fandi A (2004) Combination therapy with gefitinib, an epidermal growth factor receptor tyrosine kinase inhibitor, gemcitabine and cisplatin in patients with advanced solid tumors. Ann Oncol 15:831–838PubMedCrossRefGoogle Scholar
- 39.Koolen SL, Witteveen PO, Jansen RS, Langenberg MH, Kronemeijer RH, Nol A, Garcia-Ribas I, Callies S, Benhadji KA, Slapak CA, Beijnen JH, Voest EE, Schellens JH (2011) Phase I study of Oral gemcitabine prodrug (LY2334737) alone and in combination with erlotinib in patients with advanced solid tumors. Clin Cancer Res 17:6071–6082PubMedCrossRefGoogle Scholar
- 40.Ceresa C, Giovannetti E, Voortman J, Laan AC, Honeywell R, Giaccone G, Peters GJ (2009) Bortezomib induces schedule-dependent modulation of gemcitabine pharmacokinetics and pharmacodynamics in non-small cell lung cancer and blood mononuclear cells. Mol Cancer Ther 8:1026–1036. doi: 10.1158/1535-7163.MCT-08-0700 PubMedCrossRefGoogle Scholar
- 46.Tempero M, Plunkett W, Ruiz Van Haperen V, Hainsworth J, Hochster H, Lenzi R, Abbruzzese J (2003) Randomized phase II comparison of dose-intense gemcitabine: thirty-minute infusion and fixed dose rate infusion in patients with pancreatic adenocarcinoma. J Clin Oncol 21:3402–3408PubMedCrossRefGoogle Scholar
- 47.Bengala C, Guarneri V, Giovannetti E, Lencioni M, Fontana E, Mey V, Fontana A, Boggi U, Del Chiaro M, Danesi R, Ricci S, Mosca F, Del Tacca M, Conte PF (2005) Prolonged fixed dose rate infusion of gemcitabine with autologous haemopoietic support in advanced pancreatic adenocarcinoma. Br J Cancer 93:35–40PubMedPubMedCentralCrossRefGoogle Scholar
- 48.Wickremsinhe E, Bao J, Smith R, Burton R, Dow S, Perkins E (2013) Preclinical absorption, distribution, metabolism, and excretion of an oral amide prodrug of gemcitabine designed to deliver prolonged systemic exposure. Pharmaceutics 5:261–276. doi: 10.3390/pharmaceutics5020261 PubMedPubMedCentralCrossRefGoogle Scholar
- 51.Ciccolini J, Dahan L, André N, Evrard A, Duluc M, Blesius A, Yang C, Giacometti S, Brunet C, Raynal C, Ortiz A, Frances N, Iliadis A, Duffaud F, Seitz JF, Mercier C (2010) Cytidine deaminase residual activity in serum is a predictive marker of early severe toxicities in adults after gemcitabine-based chemotherapies. J Clin Oncol 28:160–165PubMedCrossRefGoogle Scholar
- 52.Mercier C, Raynal C, Dahan L, Ortiz A, Evrard A, Dupuis C, Blesius A, Duluc M, Franceschini F, Giacometti S, Salas S, Milano G, Favre R, Seitz JF, Ciccolini J (2007) Toxic death case in a patient undergoing gemcitabine-based chemotherapy in relation with cytidine deaminase downregulation. Pharmacogenet Genomics 17:841–844PubMedCrossRefGoogle Scholar
- 53.Raynal C, Ciccolini J, Mercier C, Boyer JC, Polge A, Lallemant B, Mouzat K, Lumbroso S, Brouillet JP, Evrard A (2010) High-resolution melting analysis of sequence variations in the cytidine deaminase gene (CDA) in patients with cancer treated with gemcitabine. Ther Drug Monit 32:53–56PubMedCrossRefGoogle Scholar
- 55.Tibaldi C, Giovannetti E, Tiseo M, Leon LG, D’Incecco A, Loosekoot N, Bartolotti M, Honeywell R, Cappuzzo F, Ardizzoni A, Peters GJ (2012) Correlation of cytidine deaminase polymorphisms and activity with clinical outcome in gemcitabine-/platinum-treated advanced non-small-cell lung cancer patients. Ann Oncol 23:670–677PubMedCrossRefGoogle Scholar
- 62.Ludovini V, Floriani I, Pistola L, Minotti V, Meacci M, Chiari R, Garavaglia D, Tofanetti FR, Flacco A, Siggillino A, Baldelli E, Tonato M, Crinò L (2011) Association of cytidine deaminase and xeroderma pigmentosum group D polymorphisms with response, toxicity, and survival in cisplatin/gemcitabine-treated advanced non-small cell lung cancer patients. J Thorac Oncol 6:2018–2026PubMedCrossRefGoogle Scholar
- 64.Mercier C, Dupuis C, Blesius A, Fanciullino R, Yang CG, Padovani L, Giacometti S, Frances N, Iliadis A, Duffaud F, Ciccolini J (2009) Early severe toxicities after capecitabine intake: possible implication of a cytidine deaminase extensive metabolizer profile. Cancer Chemother Pharmacol 63:1177–1780PubMedCrossRefGoogle Scholar
- 66.Mahlknecht U, Dransfeld CL, Bulut N, Kramer M, Thiede C, Ehninger G, Schaich M (2009) SNP analyses in cytarabine metabolizing enzymes in AML patients and their impact on treatment response and patient survival: identification of CDA SNP C-451T as an independent prognostic parameter for survival. Leukemia 23:1929–1932PubMedCrossRefGoogle Scholar
- 68.Sugiyama E, Kaniwa N, Kim SR, Kikura-Hanajiri R, Hasegawa R, Maekawa K, Saito Y, Ozawa S, Sawada J, Kamatani N, Furuse J, Ishii H, Yoshida T, Ueno H, Okusaka T, Saijo N (2007) Pharmacokinetics of gemcitabine in Japanese cancer patients: the impact of a cytidine deaminase polymorphism. J Clin Oncol 25:32–42PubMedCrossRefGoogle Scholar
- 69.Mercier C, Evrard A, Ciccolini J (2007) Genotype-based methods for anticipating gemcitabine-related severe toxicities may lead to false-negative results. J Clin Oncol 25:4855; author reply 4855-6Google Scholar
- 71.Caronia D, Martin M, Sastre J, de la Torre J, García-Sáenz JA, Alonso MR, Moreno LT, Pita G, Díaz-Rubio E, Benítez J, González-Neira A (2011) A polymorphism in the cytidine deaminase promoter predicts severe capecitabine-induced hand-foot syndrome. Clin Cancer Res 17:2006–2013PubMedCrossRefGoogle Scholar
- 74.Ueno H, Kaniwa N, Okusaka T, Ikeda M, Morizane C, Kondo S, Sugiyama E, Kim SR, Hasegawa R, Saito Y, Yoshida T, Saijo N, Sawada J (2009) Homozygous CDA*3 is a major cause of life-threatening toxicities in gemcitabine-treated Japanese cancer patients. Br J Cancer 100:870–873PubMedPubMedCentralCrossRefGoogle Scholar
- 76.Tibaldi C, Giovannetti E, Vasile E, Mey V, Laan AC, Nannizzi S, Di Marsico R, Antonuzzo A, Orlandini C, Ricciardi S, Del Tacca M, Peters GJ, Falcone A, Danesi R (2008) Correlation of CDA, ERCC1, and XPD polymorphisms with response and survival in gemcitabine/cisplatin-treated advanced non-small cell lung cancer patients. Clin Cancer Res 14:1797–1803PubMedCrossRefGoogle Scholar
- 77.Abraham A, Varatharajan S, Abbas S, Zhang W, Shaji RV, Ahmed R, Abraham A, George B, Srivastava A, Chandy M, Mathews V, Balasubramanian P (2012) Cytidine deaminase genetic variants influence RNA expression and cytarabine cytotoxicity in acute myeloid leukemia. Pharmacogenomics 13:269–282PubMedCrossRefGoogle Scholar
- 79.Baker JA, Wickremsinhe ER, Li CH, Oluyedun OA, Dantzig AH, Hall SD, Qian YW, Ring BJ, Wrighton SA, Guo Y (2013) Pharmacogenomics of gemcitabine metabolism: functional analysis of genetic variants in cytidine deaminase and deoxycytidine kinase. Drug Metab Dispos 41:541–545PubMedCrossRefGoogle Scholar
- 80.Mahfouz RZ, Jankowska A, Ebrahem Q, Gu X, Visconte V, Tabarroki A, Terse P, Covey J, Chan K, Ling Y, Engelke KJ, Sekeres MA, Tiu R, Maciejewski J, Radivoyevitch T, Saunthararajah Y (2013) Increased CDA expression/activity in males contributes to decreased cytidine analogue half-life and likely contributes to worse outcomes with 5-azacytidine or decitabine therapy. Clin Cancer Res 19:938–948PubMedPubMedCentralCrossRefGoogle Scholar
- 82.Grepin R, Guyot M, Jacquin M, Durivault J, Chamorey E, Sudaka A, Serdjebi C, Lacarelle B, Scoazec JY, Negrier S, Simonnet H, Pages G (2012) Acceleration of clear cell renal cell carcinoma growth in mice following bevacizumab/Avastin treatment: the role of CXCL cytokines. Oncogene 31:1683–1694PubMedCrossRefGoogle Scholar
- 83.Yonemori K, Ueno H, Okusaka T, Yamamoto N, Ikeda M, Saijo N, Yoshida T, Ishii H, Furuse J, Sugiyama E, Kim SR, Kikura-Hanajiri R, Hasegawa R, Saito Y, Ozawa S, Kaniwa N, Sawada J (2005) Severe drug toxicity associated with a single-nucleotide polymorphism of the cytidine deaminase gene in a Japanese cancer patient treated with gemcitabine plus cisplatin. Clin Cancer Res 11:2620–2624PubMedCrossRefGoogle Scholar
- 87.Giovannetti E, Del Tacca M, Mey V, Funel N, Nannizzi S, Ricci S, Orlandini C, Boggi U, Campani D, Del Chiaro M, Iannopollo M, Bevilacqua G, Mosca F, Danesi R (2006) Transcription analysis of human equilibrative nucleoside transporter-1 predicts survival in pancreas cancer patients treated with gemcitabine. Cancer Res 66:3928–3935PubMedCrossRefGoogle Scholar
- 88.Farrell JJ, Elsaleh H, Garcia M, Lai R, Ammar A, Regine WF, Abrams R, Benson AB, Macdonald J, Cass CE, Dicker AP, Mackey JR (2009) Human equilibrative nucleoside transporter 1 levels predict response to gemcitabine in patients with pancreatic cancer. Gastroenterology 136:187–195. doi: 10.1053/j.gastro.2008.09.067 PubMedCrossRefGoogle Scholar
- 89.Maréchal R, Mackey JR, Lai R, Demetter P, Peeters M, Polus M, Cass CE, Young J, Salmon I, Devière J, Van Laethem JL (2009) Human equilibrative nucleoside transporter 1 and human concentrative nucleoside transporter 3 predict survival after adjuvant gemcitabine therapy in resected pancreatic adenocarcinoma. Clin Cancer Res 15:2913–2919PubMedCrossRefGoogle Scholar
- 90.Sebastiani V, Ricci F, Rubio-Viqueira B, Kulesza P, Yeo CJ, Hidalgo M, Klein A, Laheru D, Iacobuzio-Donahue CA (2006) Immunohistochemical and genetic evaluation of deoxycytidine kinase in pancreatic cancer: relationship to molecular mechanisms of gemcitabine resistance and survival. Clin Cancer Res 12:2492–2497PubMedPubMedCentralCrossRefGoogle Scholar
- 93.Li L, Fridley B, Kalari K, Jenkins G, Batzler A, Safgren S, Hildebrandt M, Ames M, Schaid D, Wang L (2008) Gemcitabine and cytosine arabinoside cytotoxicity: association with lymphoblastoid cell expression. Cancer Res 68:7050–7058. doi: 10.1158/0008-5472.CAN-08-0405 PubMedPubMedCentralCrossRefGoogle Scholar
- 95.Olive KP, Jacobetz MA, Davidson CJ, Gopinathan A, McIntyre D, Honess D, Madhu B, Goldgraben MA, Caldwell ME, Allard D, Frese KK, Denicola G, Feig C, Combs C, Winter SP, Ireland-Zecchini H, Reichelt S, Howat WJ, Chang A, Dhara M, Wang L, Rückert F, Grützmann R, Pilarsky C, Izeradjene K, Hingorani SR, Huang P, Davies SE, Plunkett W, Egorin M, Hruban RH, Whitebread N, McGovern K, Adams J, Iacobuzio-Donahue C, Griffiths J, Tuveson DA (2009) Inhibition of Hedgehog signaling enhances delivery of chemotherapy in a mouse model of pancreatic cancer. Science 324:1457–1461PubMedPubMedCentralCrossRefGoogle Scholar
- 99.Poplin E, Wasan H, Rolfe L, Raponi M, Ikdahl T, Bondarenko I, Davidenko I, Bondar V, Garin A, Boeck S, Ormanns S, Heinemann V, Bassi C, Evans TR, Andersson R, Hahn H, Picozzi V, Dicker A, Mann E, Voong C, Kaur P, Isaacson J, Allen A (2013) Randomized, multicenter, phase II study of CO-101 versus gemcitabine in patients with metastatic pancreatic ductal adenocarcinoma: including a prospective evaluation of the role of hENT1 in gemcitabine or CO-101 sensitivity. J Clin Oncol 31:4453–4461PubMedCrossRefGoogle Scholar
- 101.Faivre SJ, Olszanski AJ, Weigang-Köhler K, Riess H, Cohen RB, Wang X, Myrand SP, Wickremsinhe ER, Horn CL, Ouyang H, Callies S, Benhadji KA, Raymond E (2015) Phase I dose escalation and pharmacokinetic evaluation of two different schedules of LY2334737, an oral gemcitabine prodrug, in patients with advanced solid tumors. Invest New Drugs 33:1206–1216PubMedCrossRefGoogle Scholar
- 102.Slusarczyk M, Lopez MH, Balzarini J, Mason M, Jiang WG, Blagden S, Thompson E, Ghazaly E, McGuigan C (2014) Application of ProTide technology to gemcitabine: a successful approach to overcome the key cancer resistance mechanisms leads to a new agent (NUC-1031) in clinical development. J Med Chem 57:1531–1542Google Scholar
- 103.Ghazaly EA, Joel S, Gribben JG, Mohammad T, Emiloju O, Stavraka C, Hopkins T, Gabra H, Wasan H, Habib NA, Leonard RCF, McGuigan C, Slusarczyk M, Blagden SP (2013) ProGem1: Phase I first-in-human study of the novel nucleotide NUC-1031 in adult patients with advanced solid tumors. J Clin Oncol 31 suppl:abstr 2576Google Scholar
- 104.Blagden SP, Rizzuto I, Stavraka C, O’Shea D, Suppiah P, Patel M, Loyse N, Sukumaran A, Bharwani N, Rockall A, Gabra H, El-Bahrawy M, Wasan HS, Leonard RCF, Habib NG, McGuigan C, Gribben JG, Ghazaly EA (2015) A first in human Phase I/II study of NUC-1031 in patients with advanced gynecological cancers. J Clin Oncol 33 suppl:abstr. 2547Google Scholar
- 106.Patra CR, Bhattacharya R, Wang E, Katarya A, Lau JS, Dutta S, Muders M, Wang S, Buhrow SA, Safgren SL, Yaszemski MJ, Reid JM, Ames MM, Mukherjee P, Mukhopadhyay D (2008) Targeted delivery of gemcitabine to pancreatic adenocarcinoma using cetuximab as a targeting agent. Cancer Res 68:1970–1978PubMedCrossRefGoogle Scholar
- 107.Reddy LH, Khoury H, Paci A, Deroussent A, Ferreira H, Dubernet C, Decleves X, Besnard M, Chacun H, Lepetre-Mouelhi S, Desmaele D, Rousseau B, Laugier C, Cintrat JC, Vassal G, Couvreur P (2008) Squalenoylation favorably modifies the in vivo pharmacokinetics and biodistribution of gemcitabine in mice. Drug Metab Dispos 36:1570–1577PubMedCrossRefGoogle Scholar
- 110.Joerger M, Burgers JA, Baas P, Doodeman VD, Smits PH, Jansen RS, Vainchtein LD, Rosing H, Huitema AD, Beijnen JH, Schellens JH (2012) Gene polymorphisms, pharmacokinetics, and hematological toxicity in advanced non-small-cell lung cancer patients receiving cisplatin/gemcitabine. Cancer Chemother Pharmacol 69:25–33. doi: 10.1007/s00280-011-1670-4 PubMedCrossRefGoogle Scholar
- 111.Peters GJ, Honeywell RJ, Maulandi M, Giovannetti E, Losekoot N, Etienne-Grimaldi MC, Milano G, Serdjebi C, Ciccolini J, EORTC-Pharmacology and Molecular Mechanism Group (2014) Selection of the best blood compartment to measure cytidine deaminase activity to stratify for optimal gemcitabine or cytarabine treatment. Nucleosides Nucleotides Nucleic Acids 33:403–412PubMedCrossRefGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.