Lymphocyte radiosensitivity in BRCA1 and BRCA2 mutation carriers and implications for breast cancer susceptibility

Authors


Abstract

There is conflicting evidence as to whether individuals who are heterozygous for germ-line BRCA1 or BRCA2 mutations have an altered phenotypic cellular response to irradiation. To investigate this, chromosome breakage and apoptotic response were measured after irradiation in peripheral blood lymphocytes from 26 BRCA1 and 18 BRCA2 mutation carriers without diagnosed breast cancer, and 38 unaffected age, ethnically and sex-matched controls. To assess the role of BRCA1 and BRCA2 in homologous recombination, an S phase enrichment chromosome breakage assay was used. BrdUrd incorporation studies allowed verification of the correct experimental settings. We found that BRCA1 mutation carriers without cancer had increased chromosome breaks as well as breaks and gaps per cell post irradiation using the classical G2 assay (p = 0.01 and 0.004, respectively) and the S phase enrichment assay (p = 0.01 and 0.01, respectively) compared to age-matched unaffected controls. BRCA2 mutation carriers without cancer had increased breaks as well as breaks and gaps per cell post irradiation using the S phase enrichment assay (p = 0.045 and 0.012, respectively). No difference was detected using the G2 assay (p = 0.88 and 0.40 respectively). BRCA1 and BRCA2 mutation carriers had normal cell cycle kinetics and apoptotic response to irradiation compared to age-matched controls. Our results show a demonstrable impairment in irradiation induced DNA repair in women with heterozygous germline BRCA1 and BRCA2 mutations prior to being diagnosed with breast cancer. © 2007 Wiley-Liss, Inc.

Exposure of cellular DNA to ionising radiation (IR) inflicts various types of damage.1 It is established that the creation of double strand breaks (DSB) represents the principal lesion that can be lethal if unrepaired, or mutagenic if misrepaired. They can be removed from the genome by 2 types of genetically largely independent repair mechanisms: Homologous recombination (HR) or nonhomologous end-joining (NHEJ).

It has been suggested that breast cancer patients have an abnormality in the repair of chromosome breaks and gaps post irradiation in peripheral blood lymphocytes (PBL) compared to control groups.2 It has also been shown that heterozygous mutation carriers of chromosome breakage disorders such as Nijmegen breakage syndrome are potentially at an increased risk of developing breast cancer and demonstrate in vitro chromosome radiosensitivity.3

Increased radiosensitivity has been described using both the classical G2 and MNT assays in the same breast cancer populations compared to control groups, although there was no correlation between G2 and MNT radiosensitivity results between these groups of individuals.4

A recent study found that previous radiotherapy in women with breast cancer strongly increased both spontaneous and in vitro radiation-induced micronuclei (MN) levels, and to a lesser extent, the radiation-induced DNA damage assessed by the Comet assay.5

The BRCA1 and BRCA2 proteins are involved in the repair of DSB by HR.6 HR involves the use of the homologous chromatid as a template for error free repair.7 The homologous chromatid is synthesized during the S phase of the cell cycle and the BRCA genes are active in the S and G2 phases of the cell cycle.8 In contrast, NHEJ, which is traditionally considered error-prone, reseals DSBs efficiently throughout all cell cycle phases with limited or no homology requirements.6 Therefore, the classical G2 assays may not be the only (or the most appropriate) method of testing the role of the BRCA genes in DNA repair. We have therefore used and verified a chromosome breakage assay to assess DNA repair following damage in S phase. The assay aims to score the cells (metaphases) that were originally in S phase at the point of irradiation but have passed through the cell cycle to metaphase at the point of harvesting.

There is no evidence that BRCA1 and BRCA2 mutation carriers have more severe side-effects from radiotherapy or chemotherapy than age-matched controls.9, 10 There is also no evidence to suggest a higher risk of second breast tumors in BRCA1 and BRCA2 mutation carriers who receive radiotherapy compared to other forms of treatment. However, carriers of the CHEK2 mutation c.1100delC have been shown to be at a higher risk of second breast tumors after radiotherapy and, like BRCA1 and BRCA2, the CHEK2 gene is involved in DNA repair pathways.11 It has also been shown that women exposed to high levels of irradiation at a young age such as after Chernobyl had a higher rate of breast cancer when compared to a control group over the same period of time.12 Therefore a significant abnormality in cellular response to irradiation detectable in heterozygotes for BRCA1 or BRCA2 mutations might have implications for the choice of breast screening modality and the choice of treatment with radiotherapy for breast cancer in these individuals.

Previous in vitro studies have used different assays for DSB repair capacity in BRCA1 and BRCA2 mutation carriers, to assess radiosensitivity, and whether it was related to increased breast cancer risk. However, the findings of these studies are inconsistent, probably due to small numbers of patients13 and the type of assay performed.14 In addition, most studies did not distinguish results obtained from patients who had received treatment for breast cancer and those who had not.15

By assessing radiation-induced MN formation, 10/11 BRCA1 mutation carriers have been shown to have elevated radiosensivitiy compared to 2/17 unaffected controls.16 However, no difference was seen between 4 BRCA1 mutation carriers and 4 unaffected controls post irradiation using the comet assay, which is a sensitive test for measuring the induction and repair of DNA damage. This suggests that germline BRCA1 mutations may directly or indirectly reduce the fidelity of DNA repair (detected using the MN assay) rather than the induction and repair of DNA damage (detected using the Comet assay).17

Whether developing breast cancer itself or chemotherapy/radiotherapy alters chromosome radiosensitivity has not been formally assessed in prospective studies. As part of our case-control study of newly diagnosed breast cancer and unaffected age-matched controls, we measured chromosome radiosensitivity in 5 individuals at the point of diagnosis and then again one year post treatment. No statistical difference between the pre and post treatment groups was seen.18 This has also been found by other groups.12 However, in 12 newly diagnosed breast cancer patients, we have shown that apoptotic response to IR was lower post treatment in every individual compared to levels assessed prior to therapy at the point of diagnosis, by an average of 15%. As we have demonstrated that apoptotic response to IR usually reduces with age by about 0.5% per year, this suggests a biological ageing effect of treatment of ∼30 years over an 18-month period using this assay. These results demonstrate the need for a study where the effect of breast cancer and its treatment are not potential confounders when interpreting chromosome radiosensitivity data in BRCA1 and BRCA2 mutation carriers and age-matched controls.

In this article, we describe the results of radiosensitivity studies in PBL in 26 BRCA1 and 18 BRCA2 mutation carriers who had not been diagnosed with breast cancer (hereby referred to as BRCA1 and BRCA2 mutation carriers) and 38 age, sex and ethnically matched unaffected controls never diagnosed with a malignancy.

Material and methods

Recruitment

The 26 BRCA1 and 18 BRCA2 mutation carriers were identified from Clinical and Cancer Genetics Departments from 3 London Hospitals, and blood samples taken from them. Thirty-eight controls were women; their sex, ethnicity and age were matched ±5 years and blood samples were taken as soon as possible after each mutation carrier was recruited. Thirty-five of the controls had had a normal mammogram in the previous 6 months. Five controls, members of BRCA1 and BRCA2 mutation families who had tested negative for the family mutation, were also recruited. Other controls were recruited from the breast outpatient clinics and the mammographic screening unit at St George's Hospital and had been discharged without a diagnosis of breast disease or a significant family history of breast cancer.

To assess any effect of treatment on chromosome radiosensitivity, 5 newly diagnosed breast cancer patients, unselected for family history, were recruited and tested at the point of diagnosis and again at follow-up one year after treatment. S phase enrichment chromosome radiosensitivity was also measured in 5 BRCA1 mutation carriers and 2 BRCA2 mutation carriers after treatment for breast cancer.

Ethical approval was obtained for our study from the local ethical committees. Permission to test the radiosensitivity of PBL of the BRCA mutation carriers was obtained separately.

G2 radiosensitivity assay

Investigations of spontaneous and gamma-ray induced chromosome breakage studies were performed on PBL.17, 18 Ten milliliter blood samples were taken in Lithium Heparin (BD Vacutainer, Becton Dickenson, Oxford, UK) and brought to the laboratory as soon as possible, at room temperature until set up for culture, always within 8 hr. For each patient and control 4 whole blood cultures were established from heparinised blood samples, using complete medium 199 containing 20% newborn calf serum, 2% phytohaemaglutinin (PHA), 80 IU/ml penicillin and 80 μg/ml streptomycin (Invitrogen Corporation, Paisley, Scotland, UK).

Cultures were incubated at 37°C for 65 hr prior to irradiation treatment using a gamma cell 1000 Elite irradiator (Nordion International, Kanata, Ontario, Canada) (Caesium-137 source); the dose rate was calculated at yearly intervals and exposure times adjusted accordingly to ensure a consistent dose was used throughout the study (dose rates were: year 1, 198.43 ± 0.59% Gy/hr; year 2, 193.90 ± 0.59% Gy/hr; year 3, 189.48 ± 0.59% Gy/hr): For the G2 assay, one culture was exposed to 1 Gy and one was untreated. Cultures were reincubated for a further 30 min, and then 0.2 μg/ml colcemid (Invitrogen Corporation, Paisley, Scotland, UK) was added for 1 hr prior to harvesting using standard protocols (5 min 0.075 M KCL hypotonic treatment followed by 3 changes of 3:1 methanol: acetic acid fixative). For the S phase enrichment assay, another culture was exposed to 3 Gy and one mock treated, were reincubated for 10-hr and processed in the same way. Slides made from the cell suspensions in a Thermotron environment control cabinet (Thermotron Industries, Sittingbourne, UK) were stained with Leishman's stain, and then coded prior to scoring for chromosome breakage.

Fifty metaphases were scored for chromosome breaks (aligned and mis-aligned discontinuities greater than a chromatid width) and gaps (discontinuities smaller than a chromatid width) under the light microscope for each radiation treatment and from untreated cultures (Fig. 1). The number of each type of chromosomal aberration was recorded separately, enabling statistical analysis to be performed on breaks alone and of breaks and chromatid gaps together. The G2 and S phase enrichment assay results are presented separately.

Figure 1.

Peripheral blood lymphocyte metaphase spread. Example of metaphase spread of peripheral blood lymphocyte from a breast cancer patient post irradiation demonstrating chromosome break ** and gap*.

BrdUrd pulse-chase assays to verify S phase enrichment chromosome radiosensitivity assay and cell cycle kinetics

The S phase enrichment assay was verified in cultured PBL from 4 individuals between the ages of 23 and 43, without a diagnosed malignancy, to determine the average length of time required for S phase cells to pass through to metaphase from S phase after 3Gy irradiation using flow cytometry.

The PBL were cultured for 3 days in RPMI culture medium (with 2% PHA added) as for the G2 assay. Two micromolar BrdUrd was added for 1 hr and then washed 3 times in prewarmed media to remove the BrdUrd and the cells were then exposed to 3 Gy or mock treated. The cells were then reincubated for 9, 11 or 13 hr prior to fixation in 70% ethanol at 4°C as stated in the apoptotic assay19. BrdUrd labeling was carried out after removal of the ethanol, DNA denaturation in HCL and incubation with 1:5 dilution of monoclonal mouse-anti-BrdUrd (Becton Dickinson) for 1 hr followed by incubation with rabbit anti-mouse FITC (DAKO) for a further hour. After washes, the cells were stained for DNA content by addition of propidium iodide (PI) in a volume of 1 ml with Rnase I. (The method of flow cytometric measurement and analysis of BrdUrd labeling in DNA content in lymphoid cells is described in detail in Ref.20).

The same BrdUrd pulse-chase experimental technique was used to measure cell cycle kinetics 0, 8 and 12 hr after mock treatment or 3 Gy in 4 BRCA1, 4 BRCA2 mutation carriers and 4 age-matched unaffected controls.

Apoptotic assay

Ten milliliter of whole blood was collected in heparinised tubes and taken rapidly (maximum 5 hr) to the laboratory for separation of mononuclear cells. After removal of plasma, PBL were separated by centrifugation on Histopaque (Sigma), washed and resuspended in 10 ml RPMI (Roswell Park Memorial Institute) 1640 medium (Cancer Research UK, Lincoln's Inn, London) containing 10% serum plus antibiotics. Cell concentrations were adjusted to 5 × 105 PBL/ml. Ten milliliter of the suspension was added to a series of T25 tissue culture flasks (Nunc, Roskilde, Denmark) and cells were cultured for 70 hr without PHA stimulation (as PHA stimulation greatly reduces apopototic response) and then irradiated (4 Gy using a Gammacell 1000 Elite supplied with source by MDS Nordion, High Wycombe, Bucks containing a Caesium 137 source with a dose rate of 412 Gy/hr) or mock-treated and cultured for a further 24 hr, at which time they were split into 2 aliquots and fixed in 70% ethanol. The apoptotic assay was more sensitive at 70 than at 24 hr as constitutive levels of apoptosis were still low but induced levels after irradiation were increased. During these pilot experiments we also demonstrated that apoptotic percentage reaches a maximum plateau between 18 and 30 hr following irradiation and thus 24 hr was chosen as the assay point in all further studies.20 After removal of ethanol, 2 × 106 cell aliquots were subjected to DNA denaturation by exposure to 0.1 M HCL at 37°C for 12 min. After washes, cells were stained for DNA content by addition of PI final concentration 50 μg/ml (Sigma, Poole, Dorset) in a volume of 1 ml for a minimum of 30 min prior to measurement of red fluorescence (PI), forward and 90° light scatter on a Becton Dickinson, FACSCalibur. At least 10,000 cells per sample were scanned and data stored in list mode prior to analysis using CellQuest software. Doublet discrimination using pulse area/width analysis on the PI signal was used to remove cell clumps from the analysis.

The extent of apoptosis was assessed in both aliquots by measuring the percentage of cells in a sub-G1 peak on DNA profiles. The apoptotic response to radiation was defined as the average increase in apoptosis seen when comparing the irradiated with the unirradiated sample (% apoptotic cells in the sub-G1 peak after 4 Gy minus % apoptotic cells in the sub-G1 peak after 0 Gy).21

Results

G2 and S-phase enrichment assay

BRCA1 mutation carriers.

The BRCA1 mutation carriers were aged between 24 and 65, average age 41, and their controls had an average age of 41 (range 21–62 year). The BRCA mutation carrier results were compared to the age-matched controls using the paired t test. Twenty-six BRCA1 mutation carriers had a significantly higher number of breaks as well as breaks and gaps per cell compared to their matched controls using the G2 phase (p = 0.01 and 0.004 respectively: Fig. 2). Twenty-one BRCA1 mutation carriers also had a significantly higher number of breaks as well as breaks and gaps per cell compared to their matched controls using the S phase enrichment assays (p = 0.01 and 0.01, respectively: Fig. 3).

Figure 2.

G2 phase chromosome radiosensitivity in BRCA1 and BRCA2 mutation carriers versus unaffected controls. The mean number of chromosome breaks, including and excluding gaps, per cell scored in BRCA1 and BRCA2 mutation carriers and age-matched controls without a diagnosed malignancy at thirty minutes after 1 Gy or mock treatment (G2 assay). For each individual, 50 metaphases were scored to produce a mean number of chromosome breaks per cell for that condition. These means were used to create standard error bars for each patient group. Chromosome breaks include discontinuities (greater than the width of a chromatid) or with malalignment and the smallest number of breaks required to induce a rearrangement, such as a triradial. (See text for intra-individual variation).

Figure 3.

S phase enrichment chromosome radiosensitivity in BRCA1 and BRCA2 mutation carriers versus unaffected controls. The mean number of chromosome breaks, including and excluding gaps, per cell scored in BRCA1 and BRCA2 mutation carriers and age-matched controls without a diagnosed malignancy at 10 hr after 3 Gy or mock treatment (S phase enrichment assay). For each individual, 50 metaphases were scored to produce a mean number of chromosome breaks per cell for that condition. These means were used to create standard error bars for each patient group.

BRCA2 mutation carriers.

The BRCA2 mutation carriers had an average age of 42 (range 29–36 year) and their controls had an average age of 43 (28–61 year). No difference was detected in breaks or breaks and gaps per cell using the G2 assay (p = 0.88 and 0.40, respectively: Fig. 2). Fifteen BRCA2 mutation carriers without cancer had increased breaks as well as breaks and gaps per cell, post irradiation using the S phase enrichment assay (p = 0.045 and 0.012, respectively.

The BRCA2 mutation carriers had a significantly higher number of breaks as well as breaks and gaps per cell compared to their matched controls using the S phase enrichment assay (p = 0.045 and 0.012, respectively: Fig. 3).

S phase assay verification and cell cycle kinetics

Irradiation of PBLs immediately after pulse-chase caused a G2/M block for several hours. At 9 hr post irradiation, 23% of S phase cells have returned to G1, which is ∼2.5% per hour post reincubation, assuming a consistent movement of cells through the cell cycle. However, between 11 and 13 hr reincubation, a further 12% of labeled cells returned to G1 at a rate of 6% per hour. This suggests that there is an ∼3-fold enrichment of S phase labeled cells passing through the cell cycle back to G1 between 11 and 13 hr (Fig. 4). Ten hours post reincubation plus a further hour in colcemid was therefore chosen in the S phase enrichment assay to allow up to 11 hr for the cells to reach metaphase. This time point was used, rather than 12 or 13 hr, as the BrdUrd incorporation assay measures the proportion of cells that have passed through metaphase and returned to G1. However, to score chromosome breaks in metaphases prior to the cell dividing, one hr less was used to allow for the cell to pass through metaphase and return to G1.

Figure 4.

S phase enrichment chromosome radiosensitivity assay. The S phase enrichment chromosome breakage assay, demonstrating cells measured for red fluorescence (PI), forward and 90° light scatter on a Becton Dickinson FACSCalibur. During the S phase, BrdUrd is incorporated into DNA at the expense of thymidine. DNA content is measured along the x-axis (as it binds to propidium iodide) and BrdUrd content along the y-axis. In S phase, the DNA content is doubled and BrdUrd content is increased. This is then halved as the cell divides and returns to G1. An example trace is shown in (a) and annotated in (b).

No significant difference in cell cycle kinetics was found between the BRCA1 and BRCA2 mutation carriers and unaffected age-matched controls (Fig. 5).

Figure 5.

Cell cycle kinetics in BRCA1 and BRCA2 mutation carriers and age-matched unaffected controls. Percentage of S phase (BrdUrd labelled) cells returned to the G1 phase of the cell cycle 8 and 12 hr after 3 Gy irradiation and post mock treatment in PHA stimulated PBL from 4 BRCA1, 4 BRCA2 and 4 age-matched unaffected controls.

Apoptotic response to irradiation results

The BRCA1 and BRCA2 mutation carriers were compared to age-matched controls using the 2 tailed paired t test. No significant difference in apoptotic response spontaneously or after irradiation with 4 Gy in the BRCA1 and BRCA2 mutation carriers compared to the controls was identified (Fig. 6).

Figure 6.

Apoptotic response to irradiation in BRCA1 and BRCA2 mutation carriers and controls. Percentage of apoptotic cells in response to irradiation in BRCA1 and BRCA2 mutation carriers and age matched unaffected controls 24 hr after mock treatment, 4 Gy and induced apoptotic response (4–0 Gy) measured using flow cytometry.

The effect of breast cancer treatment on chromsome radiosensitivity

No consistent effect of treatment on chromosome breakage as measured by the G2 assay was seen in 5 newly diagnosed breast cancer patients, unselected for family history (Fig. 7). Using the binomial proportion test and 40 individual time points, the chromosome aberrations increased post treatment only in 21/40 occasions (p = 0.87).

Figure 7.

Chromosome radiosensitiivty in unselected breast cancer patients at the point of diagnosis and one-year post treatment. The number of chromosome breaks with and without gaps in 5 breast cancer patients using the G2 assay at the point of diagnosis and repeated one-year post treatment.

We measured S phase chromosome radiosensitivity in 5 BRCA1 mutation carriers (mean breaks per cell 0.62 ± 0.05 using the S phase assay) and 2 BRCA2 mutation carriers (mean breaks per cell 0.48 ± 0.26 using the S phase assay) after treatment for breast cancer. These were higher levels than in unaffected BRCA1 carriers (Fig. 3) but in BRCA2 mutation carriers levels were not increased. These numbers were too small to allow any conclusions to be drawn.

Discussion

We present chromosome breakage data on the largest series of germline BRCA1 and BRCA2 mutation carriers without a diagnosed malignancy, and age, sex and ethnically matched unaffected controls. Given concerns about the possible risks of breast screening with mammography in unaffected mutation carrier women and the potentially damaging effects of high dose radiotherapy in normal breast tissue, we suggest that it is important to assess whether there is evidence for increased radiosensitivity in such mutation carriers.

The radiosensitivity assays we have used measure radiation induced DSB in metaphase spreads that have been enriched for cells that were in G2 and/or S phase of the cell cycle at the point of irradiation. The timing of the S phase enrichment assay was defined by measuring with flow cytometry the movement of BrdUrd incorporated cells through the cell cycle after irradiation.

Most previous studies of chromosome radiosensitivity assays have recruited breast cancer patients with and without BRCA1 or BRCA2 mutations and age-matched unaffected controls. Such patients will almost have received radiotherapy and/or chemotherapy. It is not known whether chromosome radiosensitivity in PBLs is affected by the development of breast cancer or its treatment, and most previous studies have not controlled for these variables. We have found no evidence that breast cancer treatment consistently affects chromosome radiosensitivity in a small number of cases assessed at diagnosis and one-year post treatment. Our results demonstrate the need for a study in which the effect of breast cancer and its treatment are not potential confounders in interpreting differences between in vitro chromosome radiosensitivity in BRCA1 and BRCA2 mutation carriers and age-matched controls.

Our results bring additional evidence that BRCA1 and BRCA2 mutation carriers do have, on average, increased chromosomal aberrations per cell after irradiation in PBL compared to unaffected control groups using 2 independent tests, the G2 and the S-phase enrichment assays. The aetiology of chromosome aberrations is controversial. It has been suggested that they may be markers of incomplete DNA repair, technical artefacts or the result of incomplete HR-independent DNA repair mechanisms.22, 23, 24

Chromosome breaks were scored according to the routine methods devised to identify possible Ataxia telangiectasia and Fanconi's anaemia patients at Guy's hospital. Complex rearrangements, such as tri-radials are scored according to the minimum number of chromosome breaks that would be required to produce the abnormal karyotype. Translocations are not routinely measured, as this requires the use of spectral karyotyping to be accurate, which is not practical in assessing large number of patients. This has been measured in 5 BRCA1 mutation carriers compared to age-matched controls,25 suggesting how defects in HR may increase error-prone DNA repair.

No significant differences were found between BRCA1 and BRCA2 mutation carriers and controls in cell cycle kinetics or apoptotic response of PBL after irradiation. Despite the small number of cases, these results might suggest that the increase in radiosensitivity is not due to differences in cell cycle check points or inefficient apoptosis but is more likely to be the result of a direct effect of a defect in DNA repair.

BRCA1 and BRCA2 are thought to be involved in DNA repair by HR.6 If BRCA gene haploinsufficiency is not sufficient to causethe chromosome instability, phenotype we observed, one could postulate that loss of the second allele in a proportion of lymphocytes could cause an increase in chromosome breaks in these cells in BRCA1 and BRCA2 mutation carriers. It is unlikely that this would occur in a large proportion of cells, but if this led to a large number of breaks even in a small proportion of cells, this could increase the average number of chromosome breaks scored.

Cells homozygous for BRCA1 or BRCA2 mutations have been shown to be deficient in DNA repair by HR.26 This loss of HR could enhance repair by error-prone methods such as NHEJ. The subsequent higher levels of deletions and translocations could increase the probability of neoplastic transformation by loss of tumor suppressor genes, activation of proto-oncogenes or epigenetic mechanisms.

Unlike carriers of for c.1100delC CHEK2 mutation,11 there is no evidence that there is a clinical difference from controls in the radiosensitivity of breast tissue in BRCA1 and BRCA2 mutation carriers. However, since the in vitro assays used short time points, which are unlike the clinical assessment, and clinical assays employed different radiation doses, the assays are not necessarily relevant to each other. Also the dynamics of DNA repair in lymphocytes and fibroblasts may differ.

The association between in vitro chromosome radiosensitivity in BRCA1 and BRCA2 mutation carriers who have been treated for breast cancer is well recognised. However, as we have studied BRCA1 and BRCA2 mutation carriers undiagnosed with breast cancer, we have provided stronger evidence for radiosensitivity in these individuals, which is unrelated to the development of, or treatment for, breast cancer. We could further extrapolate that radiosensitivity could be a measurable cellular phenotype in individuals with breast cancer susceptibility. This may be significant in planning future screening and treatment for these individuals.

Acknowledgements

Authors would like to acknowledge the following for their support in the recruitment of patients to this study. Mr. Hisham Hamed, Mr. Nicholas Beechey-Newman, Prof. Ian Fentiman, Mr. Jonathan Roberts, Mr. Kefah Mokbel, Mr. Anup Sharma, Dr. Louise Wilkinson and the staff of the Carrier Clinic at The Royal Marsden Hospital NHS foundation Trust, London.

Ancillary