Aim, design and setting of the study
The main aim of the present study is therefore to assess the clinical and genomic safety of the treatment with a common formulation of
Ginkgo biloba L. leaf extract (IDN 5933 also known as Ginkgoselect®Plus, Indena S.p.A., Milan, Italy). The study design is a randomized clinical trial, placebo controlled, double blinded, performed in elderly institutionalized subjects. Given the obvious difficulty of assessing the risk associated to treatment using cancer incidence, we identified a battery of tests that may allow the detection of early events in the multistep process of carcinogenesis, especially as far as liver carcinogenesis is concerned. The first indexes to be evaluated were the presence of DNA damage and genomic instability, events both validated as precursors of cancer development. The comet assay and the micronucleus test in peripheral blood lymphocytes were evaluated for this purpose, respectively [
32‐
36]
. To provide a more specific assessment concerning events associated to hepato-carcinogenesis, the expression of some critical genes, i.e., c-
myb, p53, and
ctnnb1 (β-catenin), reported to be modulated in early stages of carcinogenesis [
37‐
39], was comparatively evaluated in treated subjects versus placebo. Finally, the presence of adverse clinical events, or the presence of liver injury associated to the experimental treatment was evaluated in all participants to the clinical trial.
The GiBiEx multicenter, randomized. Double blind, placebo-controlled clinical trial compared subjects receiving twice-daily doses of either 120-mg of Ginkgo biloba L. leaf extract (IDN 5933) or placebo for a 6 months period. A population of 140 elderly residents (≥ 65 years) was enrolled among the residents of three nursing homes of the San Raffaele network, i.e., the San Raffaele Montecompatri, the San Raffaele Rocca di Papa, and the San Raffaele Sabaudia, between June and November 2015. Patients with previous report of increased bleeding tendency receiving treatment with anticoagulant and antiplatelet drugs, with, cognitive impairment, refusing to sign the informed consent, and with a life expectancy of less than 1 year were not considered eligible for enrollment. Overall, 74 subjects didn’t meet the eligibility criteria and were excluded from the study. Eventually, 66 subjects (age > 65) signed the informed consent and accepted to participate to the study: 26 males 839.4%) and 40 females (60.6%). All participants were randomly allocated to 120 mg IDN 5933, twice per day in tablet form, or to placebo with a 1:1 ratio via computer-generated randomized allocation by an independent data management contractor with blocking by site.
The
Ginkgo biloba leaf extract (IDN 5933 also reported as Ginkgoselect®Plus) utilized in this study was produced by INDENA S.p.A. Milan, Italy and contains 24.3% flavone glycosides and 6.1% of terpene lactones (2.9% bilobalide, 1.38% ginkgolide A, 0.66% ginkgolide B, 1.12% ginkgolide C) as determined by HPLC. Ginkgoselect®Plus is obtained by extracting the dried leaves of
Ginkgo biloba using ethanol: water (70:30
v/v). The crude extract is then purified by subsequent concentration to water for the removal of biflavone compounds. Then two consecutive resin columns are used for the elimination of sugars and gingkolic acids respectively. The obtained product is then dried for residual ethanol and water elimination. Ginkgolic acids are considered unwanted substance of the extract and must be not more than 5 ppm (NMT 5 ppm). Only water and ethanol are used as solvents for the preparation of IDN 5933 (see Additional file
1).
At the start of the study (T0) all participants were administered a questionnaire concerning demographics characteristics, life-style, smoking habit, and relevant clinical history including concomitant treatments (see Additional file
2). A sample of 20 ml of blood was collected from each participant at T0 and at the end of the study (T1). The duration of the study was 6 months: 34 subjects received twice-daily doses of 120-mg of experimental treatment,
Ginkgo biloba L. leaf extract IDN 5933; 32 subjects received placebo.
A total of 19 subjects (28.8%), withdrew from the study for the following reasons: death (1 subject), for acute pancreatitis in a chronic renal failure patient, discharge (10 subjects), admission to another hospital for worsening state of health (4 subjects), and discontinued treatment (4 subjects). The study was completed by 47 subjects (71.2%, 18 males and 29 females): 20 subjects in the placebo group (9 males and 11 females) and 27 subjects in the treatment group (9 males and 18 females).
Clinical conditions of all patients were monitored at the beginning, during the six months of treatment, and at the end of the study by the medical personnel of each nursing home, after receiving special training. Biological samples were collected at time T0 and at time T1 to evaluate liver injury [gamma-glutamyl transferase (γGT), alanine aminotransferase (ALT), aspartate aminotransferase, (AST)]. DNA damage and genomic instability [Comet Assay and Micronucleus assay (MN)]. A subgroup of 17 individuals (8 treated versus 9 placebos) was monitored to compare the expression patterns of genes putatively associated to early events of hepatic carcinogenesis, i.e., p53, c-myb and ctnnb1.
Micronucleus assay
Blood samples from 47 subjects were obtained by venipuncture, collected in heparinised tubes, and then transported in cold (4 °C) to the laboratory where they were processed. All samples were coded and analyzed under blind conditions
. The micronucleus assay (MN) was performed following the protocol previously described by Fenech [
40,
41]. The cultures were established in duplicate by adding 0.5 ml of whole blood to 4.5 ml of culture medium: RPMI 1640 Medium Gibco™ (Thermo Fisher Scientific, MA USA) containing 15% heat inactivated fetal bovine serum Gibco™ (Thermo Fisher Scientific, MA USA), 1% phytohaemagglutinin Gibco™ (Thermo Fisher Scientific, MA USA), 1% L-glutamine (200 mM) (Biowest Europe, Nuaillé – France) and 1% penicillin (5000 U/ml)/streptomycin (5000 μg/ml) Gibco™ (Thermo Fisher Scientific, MA USA). Cells were incubated at 37 °C, 5% CO
2. Cytochalasin-B (6 μg/ml) (Sigma-Aldrich S.r.l. Milan, Italy) was added at 44 h to prevent cytokinesis. After 72 h of incubation, cells were centrifugated, collected, and treated with a mild hypotonic solution (0.075 M KCl at 4 °C), and then centrifuged immediately and fixed in Carnoy’s solution [3:1 methanol–acetic glacial acid). Air dried slides were prepared and stained with 5 μg/ml 4, 6-diamidino-2-phenylindole (DAPI) (Sigma-Aldrich S.r.l. Milan, Italy).
Microscope analyses were performed by employing a fluorescence microscope Axio Imager Z2 microscope (Carl Zeiss Microscopy GmbH, Jena, Germany) equipped with a Automated SlideFeeder ×80, connected to motorized and controlling the microscope components for automated focusing, light source adjustment (for bright field imaging) and fluorescence filter changes. A minimum of 1000 binucleated cells were automatically scored to determine the number of MN and binucleated cells with MN (BN) with Metafer software (Carl Zeiss AxioImager.Z2, Jena, Germany). After the automated scan, the image gallery was visually reviewed by an experienced scorer, following criteria described by Fenech for MN and BN [32, 33, and 41].
Alkaline comet assay
Lymphocytes were separated by gradient centrifugation of blood on LeucoSep. After washing pellet was resuspended in cryopreservation medium (Synth-a-Freeze® Thermo Fisher Scientific, Waltham, MA USA) and stored at −80 °C. Cryopreserved lymphocyte aliquots were thawed, immediately diluted in 5 ml PBS and centrifuged 5 min at 180 g to remove the cryopreservation medium. The pellet was resuspended in PBS and cell suspension was processed for the alkaline comet assay. The assay was performed essentially as previously described [
42]. Two slides were prepared for each experimental point. After overnight lysis at 4 °C in lysing solution [2.5 M NaCl, 100 mM Na
2EDTA, 10 mM Tris, [pH 10] containing 10% DMSO and 1% Triton X-100], slides were placed in electrophoresis buffer [300 mM NaOH, 1 mM Na
2EDTA, pH >13], and left in the solution for 25 min at 4 °C. Electrophoresis was carried out at 4 °C for 25 min, 27 V (0.8 V/ cm) and 300 mA. Slides were then neutralized, fixed and air-dried. Before scoring, slides were stained with 12 μg/ml ethidium bromide (Sigma-Aldrich) and examined at 200× magnification with an Olympus fluorescence microscope. Slides were analyzed blindly with a computerized image analysis system (Delta Sistemi, Italy). For each sample, 150 cells were analyzed from 2 different slides; the percentage of DNA in the tail of the comet (% TI) was used as the parameter for evaluation of DNA damage. Heavily damaged cells, i.e. hedgehogs (as determined by visual scoring or cells having more than 80% DNA in the tail), were not included in the measurement but they were counted and their percentage was calculated per sample.
Gene expression of c-myb, p53, and ctnnb1
An aliquot of lymphocytes isolated as described above, was centrifuged and pellet was cryopreserved at −80 °C. Total RNA was extracted from blood samples by Trizol® (Invitrogen, Thermo Fisher Scientific, and Waltham, MA USA) protocol. The amount and purity of the extracted RNA was evaluated by fiber optic spectrophotometer (Nanodrop ND-1000, NanoDrop Technologies, Wilmington, DE, USA) calculating the 230/260 and 260/280 absorbance ratios. Two hundred nanograms of total RNA were retro-transcribed into total cDNA by TaqMan® Reverse Transcription Reagent (Applied Biosystems, Thermo Fisher Scientific), according to manufactures’ indications. Potential contamination by genomic DNA was verified by amplifying the human β-actin gene (Fw primer: 5′ TGA CGG GGT CAC CCA CAC TGT GCC CAT CTA 3′; rev primer: 5’ CTA GAA GCA TTT GCG GTG GAC GAT GGA GGG 3′) in a conventional thermocycler according to the following Polymerase Chain Reaction (PCR) conditions: one cycle at 95 °C (10 min), 40 cycles at 95 °C(30 s)-58 °C (30 s)-72 °C (30 s), one cycle at 72 °C (5 min). Reactions were run in 20 μl total volume by Taq Polymerase (Euroclone, Milan, Italy); PCR products (663 bp amplicon for the cDNA, 867 bp amplicon for the genomic DNA) were visualized on 1.5% agarose gel by ethidium bromide staining.
Analysis of the human (hsa)
c-myb, p53 and
ctnntb1 genes expression was carried out with 1 μL of cDNA using SYBR Green master mix (Applied Biosystems) and analyzed on a Eco™ Real-Time PCR System (Illumina,
San Diego, CA, USA). All reactions were run in quadruplicate and the relative abundance of the specific mRNA level was calculated by normalizing to the GAPDH expression using the 2-ΔΔCt method [
43]. Primers sequence is as follows:
-
hsa_c-myb Fw 5’ ACCATGACTATGATGGGCTGC 3′
-
hsa_c-myb Rev. 5’ TCCCCAAGTGACGCTTTCC 3′
-
hsa_p53 Fw 5’ TAACAGTTCCTGCATGGGCGGC 3′
-
hsa_p53 Rev. 5’ AGGACAGGCACAAACACGCACC 3′
-
hsa_ctnnb1 Fw 5’ AGCTTCCAGACACGCTATCAT 3′
-
hsa_ctnnb1 Rev. 5’ CGGTACAACGAGCTGTTTCTAC 3′
-
hsa_GAPDH Fw 5’GAG TCA ACG GAT TTG GTC GT3’
-
hsa_GAPDH Rev. 5’ GAC AAG CTT CCC GTT CTC AG 3′
Data were analyzed in each sample in terms of increase/decrease of the specific gene expression at T1 compared to T0, with a cut-off value ≥ 2-fold change.
Statistical analysis
A pilot study with this sample size will identify with 95% confidence a problem with a 5% probability to occur in a potential study participant [
44]. Descriptive statistics were measured for all patients’ characteristics. The presence of heterogeneity between groups was tested with the Chi-square test for qualitative variables and with the Students T-test for continuous variables. Logarithmic transformation of data was applied when necessary to normalize the distribution. Several variables were categorized before analysis, e.g., marital status, education, weight, height, smoking habit, comorbidities, and liver function tests. The comparison of clinical and molecular indexes of DNA damage in the two groups of treatment, were evaluated with univariate analysis, and subsequently with multiple regression analysis, using general linear models, to take into account the influence of age, sex, smoking status, and other covariates as mental and behavioral disorders.
Statistical analysis concerning the MN assay evaluated the frequency of micronucleated cells per 1000 binucleated cells. To take into account the large overdispersion, due to the random nature of variables, a mixed distribution was applied to fit regression models, i.e. the Poisson and Negative Binomial Models. A lognormal model was applied to the fit Tail intensity in the comet assay.
In the Poisson and Negative Binomial Models, a specific analysis was used (through differences in starting values and convergence criteria of the algorithms) to deal with the overdispersion, scaling the standard errors using the square root of the Pearson chi-square dispersion. The standard errors were adjusted to compensate for the overdispersion in both methods. The parameters estimation was accomplished via maximum likelihood, and the AIC - Akaike’s Information Criterion [
45] was taken into account to select the best model. For each model the Mean Ratio (MR), together with asymptotic 95% confidence intervals (95% CI) was computed. Each model included age, sex and smoking habit, and actual confounders.
The critical limit for significance was set at
p < 0.05. The statistical softwares used for the analysis were Stata [
46], and SPSS (Statistical Package for Social Science, Version 16.0) [
47].