¹⁴C BOMB-PULSE DATING AND STABLE ISOTOPE ANALYSIS FOR GROWTH RATE AND DIETARY INFORMATION IN BREAST CANCER?

Abstract

The purpose of this study was to perform an initial investigation of the possibility to determine breast cancer growth rate with ¹⁴C bomb-pulse dating. Tissues from 11 breast cancers, diagnosed in 1983, were retrieved from a regional biobank. The estimated average age of the majority of the samples overlapped the year of collection (1983) within 3σ. Thus, this first study of tumour tissue has not yet demonstrated that ¹⁴C bomb-pulse dating can obtain information on the growth of breast cancer. However, with further refinement, involving extraction of cell types and components, there is a possibility that fundamental knowledge of tumour biology might still be gained by the bomb-pulse technique. Additionally, δ ¹³C and δ ¹⁵N analyses were performed to obtain dietary and metabolic information, and to serve as a base for improvement of the age determination.

INTRODUCTION

Large amounts of the radioactive carbon isotope ¹⁴C were produced during atmospheric testing of nuclear weapons in the late 1950s and early 1960s. As a consequence, the concentration of ¹⁴C in air was almost doubled by 1963. Since the ¹⁴C produced was incorporated in atmospheric CO₂ and introduced into the global carbon cycle, all organisms living during the bomb-pulse era, including humans, have been labelled with bomb-¹⁴C^(1–3). When the Limited Test Ban Treaty from 1963 was implemented, the atmospheric ¹⁴C concentration commenced decreasing mainly due to uptake of ¹⁴CO₂ into the oceans and also into the biosphere (see Figure 1).

Figure 1.

¹⁴C specific activity in atmospheric CO₂ and oceans representative for the northern hemisphere (expressed in units of Fraction Modern, F¹⁴C⁽⁴⁾; F¹⁴C is ~1.0 before the nuclear weapons tests, corresponding to the natural ¹⁴C/¹²C ratio of about 10⁻¹⁰%^(5–7)).

Already in the early 1970s, Harkness and Walton⁽⁸⁾ realised the potential of using bomb-¹⁴C as a kinetic tracer in humans. Since then, the well-known decreasing atmospheric ¹⁴C concentration has provided useful information in several fields in the medical sciences, see e.g. Falso and Buchholz⁽⁹⁾ and Spalding et al. ⁽¹⁰⁾. For instance, this so called ¹⁴C bomb pulse has served in studies of turnover rates in human cells and tissues (e.g. in eye lenses⁽¹¹⁾, fat cells⁽¹²⁾ and Achilles tendon⁽¹³⁾). The development of various diseases has also been investigated using the ¹⁴C bomb-pulse technique, with ¹⁴C measurements on e.g. gallstones⁽¹⁴⁾ and Alzheimer plaques⁽¹⁵⁾. A common denominator for these studies is slow turnover (years). Another example is Gonçalves et al. ⁽¹⁶⁾, who demonstrated that human atherosclerotic plaques, known to cause e.g. heart attacks and strokes, develop slowly (mean biological ages 5–15 years). Quantification of the age of various structures in the plaque has gained a better understanding of the development of the plaque and opened for improved treatment methods.

The bomb-pulse technique has to the present authors’ knowledge not yet been used to study cancer. Understanding of the time cause for the development of human cancer tissue is of utmost importance for improved prevention and treatment strategies. So far, certain assumptions have had to be made in order to estimate how long a cancer has grown. There are theoretical models to describe tumour growth rates, such as an exponential growth model or the Gompertz Model, where the latter takes into consideration a slower growth rate as the tumour size increases. These models have been applied to cancer in general⁽¹⁷⁾ and to breast cancer in particular^(18–20). These models have been shown to fit experimental and clinical data, such as observations of the doubling of tumour volume on serial mammograms. From these studies, the average tumour volume doubling time for human breast cancer has been estimated to be between 105 and 327 days^(21–29). Theoretically, this means that a breast cancer that is 10-mm large at detection (30 volume doublings) has been growing for 9–27 years. However, these studies are limited due to the lack of direct observations of tumour age.

The main purpose of this study was to perform an initial investigation of the possibility to determine the growth rate of breast cancer with ¹⁴C bomb-pulse dating. An additional aim was to highlight the limitations of the ¹⁴C bomb-pulse technique of human tissues due to dietary variations. Furthermore, the paper pays attention to how dietary information can be obtained from stable isotope ratio analysis of carbon and nitrogen, also providing means to increase the accuracy of bomb-pulse dating. Stable isotope ratio analysis may also give important information relating cancer development to diet.

Bomb-pulse dating and stable isotope analysis

The ¹⁴C bomb-pulse technique uses ¹⁴C data from atmospheric clean-air CO₂ to translate the ¹⁴C specific activity of the sample into a calendar date (so called CaliBomb dates, see ‘Material and Methods’ section). For humans, carbon enters the body mainly through the diet. Thus, ¹⁴C data from clean-air CO₂ may not be fully representative for humans.

A previous study by Georgiadou et al. ⁽³⁰⁾ used human blood serum samples from a biobank to estimate the accuracy of bomb-pulse dating on human material. Blood serum samples collected from residents of Malmö (Sweden) in 1978 exhibited CaliBomb dates between 3.0 ± 0.4 years before the collection date and 0.2 ± 0.5 years after the collection date (the average deviation from collection date was −1.5 ± 0.7 years). Two major effects associated with the age deviation, competing in opposite directions, were identified: (1) delay time between production and consumption of foodstuffs, which can explain CaliBomb dates obtained before the collection date and (2) excessive consumption of marine food products, which has the potential of producing CaliBomb dates after the collection date.

As discussed in the study by Georgiadou et al. ⁽³⁰⁾, there are also other factors that have the possibility to influence the obtained CaliBomb date, however, for the majority of the population probably to a more limited extent than delay time and marine food consumption. Anthropogenic ¹⁴C released from nuclear power plants or from research laboratories, industry or hospitals using ¹⁴C as a tracer has the potential of producing too old CaliBomb dates (at the declining bomb-pulse curve)^{(31, 32)}. On the other hand, fossil fuel-based products in food industry (e.g. in CO₂ used in cultivation in greenhouses) as well as food grown in heavily industrialised areas may lead to foodstuffs having lower ¹⁴C specific activity than clean air, thus contributing to producing too young CaliBomb dates (at the declining bomb-pulse curve)⁽³³⁾.

¹⁴C bomb-pulse dates reported in the literature usually do not take into account that there is a delay time between production and consumption of the food^{(33, 34)}. Neither is it commonly considered that marine foodstuffs can influence the accuracy of the calibration (see the difference in marine and atmospheric calibration curve in Figure 1). However, since a large consumption of marine foodstuffs has the potential to affect the calibration curve (see Figure 1) on an individual basis, information about the diet may be important in bomb-pulse dating of human tissue samples. Additionally, the individual diet is of interest for studies of the development of various diseases, including breast cancer⁽³⁵⁾.

One technique to assess the diet of an organism is analysis of the ratios of stable isotopes of carbon and nitrogen (expressed as δ ¹³C and δ ¹⁵N in ‰, where δX = [(R_sample/R_standard) − 1] × 1000, and R is the ratio of heavy to light isotope, i.e.¹³C/¹²C or ¹⁵N/¹⁴N) (see e.g. Schoeller⁽³⁶⁾). δ ¹³C and δ ¹⁵N provide dietary information since the stable isotope ratios in tissues and organs of the consumer reflect those of the diet with a small shift. This discrimination of one of the stable isotopes occurs in each step of the food chain. The difference in δ-value between an organism and its diet, referred to as the discrimination factor, is generally about 1‰ for δ ¹³C and 3‰ for δ ¹⁵N⁽³⁷⁾. The discrimination factor, however, also depends on various factors such as species, age, metabolic processes and environmental conditions (see e.g. Caut et al. ⁽³⁷⁾ and O’Connell et al. ⁽³⁸⁾). Thus, different trophic levels (positions in the food chain) are characterised by different ranges of δ values (see Figure 2).

Figure 2.

Generalised isotopic trophic level diagram for terrestrial and marine food webs (values from Georgiadou et al. ⁽³⁰⁾). The δ ¹³C values of C3 plants (e.g. wheat and potatoes) differ from those of C4 plants (e.g. maize and sugar cane) due to photosynthetic pathways. For the Swedish diet, of relevance for this study, C3 plants dominate over C4 plants.

Despite the complexity and variation of discrimination factors, several studies have shown that δ ¹³C and δ ¹⁵N are valuable biomarkers of dietary intake (see e.g. Georgiadou et al. ⁽³⁰⁾ and Patel et al. ⁽³⁹⁾ and references therein). The approach can additionally be used as a tool to identify samples where extreme dietary conditions (high intake of marine products) might influence the ¹⁴C bomb-pulse date (decreasing the applicability of the atmospheric calibration curve).

The serum study⁽³⁰⁾ showed that the age deviation (difference between CaliBomb date and actual sampling date) correlated strongly with δ ¹³C, which was interpreted as influence from marine dietary components. δ ¹⁵N, however, did not show any correlation with age deviation in that study. δ ¹³C data from the serum study have also been used to demonstrate the possibility to develop general methods increasing accuracy as well as precision of the age determination using the ¹⁴C bomb-pulse technique⁽⁴⁰⁾. However, more data are needed to produce a general correction model.

MATERIAL AND METHODS

Frozen breast cancer tissues from 11 women collected in 1983 were retrieved from the South Swedish Breast Cancer Group Biobank. In three cases, the samples were large and sub-samples from each tumour could be taken (maximum of three sub-samples per cancer).

Prior to ¹⁴C analysis, parts of each tissue sample were graphitised according to the procedures described by Andersson Georgiadou et al. ⁽⁴¹⁾ and Genberg et al. ⁽⁴²⁾. Total carbon in the graphite samples ranged between 32 and 124 µg. The ¹⁴C/C ratio, expressed as F¹⁴C⁽⁴⁾, was analysed with accelerator mass spectrometry (AMS) at the Lund Single Stage AMS facility at Lund University^{(43, 44)}. The analytical precision of the F¹⁴C measurements for human tissue samples is usually 0.5–1% for this instrument⁽⁴⁰⁾. The obtained F¹⁴C values were translated into calendar years using the CaliBomb software (http://calib.qub.ac.uk/CALIBomb/), choosing the calibration dataset ‘Levin’ (representative for European clean-air CO₂) and ‘smoothing’ (averaging the dataset) of 0.5 years. A given analytical uncertainty in F¹⁴C results in various uncertainties in the CaliBomb date depending e.g. on the steepness of the declining bomb-pulse curve and on seasonal variations (see Figure 1).

δ ¹³C and δ ¹⁵N were analysed using isotope ratio mass spectrometry (IRMS) at the Department of Biology, Lund University, using between 0.28 and 0.40 mg of dried tumour material for each sample. The analytical precision obtained for standards at this instrument is usually <0.4‰ for carbon and <0.12‰ for nitrogen (1σ). Further information about the IRMS analysis can be found in Andersson Georgiadou⁽⁴⁰⁾.

RESULTS AND DISCUSSION

The results of the ¹⁴C measurements and the IRMS analysis are shown in Table 1. A few samples lack IRMS data due to insufficient amount of sample material. The analytical uncertainties of the ¹⁴C measurements were higher than normally because of the difficulty of accurately weighing the samples and due to smaller and more varying sample sizes than usually.

Table 1.

Results from the ¹⁴C and IRMS analyses of the breast tumour samples.

Sample	Year of birth	Lab code	14C analysis			IRMS analysis
			m C (µg)	F14C ± 1σ	CaliBomb date ± 1σ	m (mg)	δ 13C (‰)	δ 15N (‰)
T1	1924	GEO_1511	45	1.211 ± 0.014	1984.8 ± 1.1	0.30	−22.3	11.0
T2	1901	GEO_1512	56	1.193 ± 0.013	1986.2 ± 1.9	0.30	−22.0	11.7
T3A	1928	GEO_1513	130	1.309 ± 0.015	1979.3 ± 0.6	0.35	−22.5	13.3
T3B		GEO_1514	63	1.241 ± 0.014	1982.8 ± 0.9	0.28	−22.7	12.8
T3C		GEO_1515	88	1.240 ± 0.014	1982.9 ± 0.9
T4	1919	GEO_1516	54	1.239 ± 0.014	1982.9 ± 0.9
T5	1922	GEO_1517	53	1.222 ± 0.014	1984.1 ± 0.8	0.39	−21.2	13.2
T6	1901	GEO_1518	124	1.244 ± 0.014	1982.7 ± 0.9	0.37	−21.0	13.0
T6:2		GEO_1519	103	1.232 ± 0.014	1983.1 ± 0.8
T7	1898	GEO_1520	83	1.274 ± 0.014	1980.9 ± 1.0	0.32	−27.3	9.3
T8	1932	GEO_1521	101	1.247 ± 0.014	1982.3 ± 1.3	0.40	−23.3	11.0
T9A	1902	GEO_1522	52	1.184 ± 0.013	1987.5 ± 1.7	0.35	−22.8	9.4
T9B		GEO_1523	37	1.190 ± 0.013	1986.7 ± 1.7	0.28	−24.7	7.8
T9C		GEO_1524	32	1.157 ± 0.013	1990.2 ± 1.7
T10	1915	GEO_1525	98	1.137 ± 0.013	1992.8 ± 3.3	0.37	−20.6	13.7
T11	1911	GEO_1526	79	1.223 ± 0.014	1984.1 ± 0.7	0.32	−21.9	10.3

All samples were collected in 1983. Only results from the time after the bomb-pulse peak in 1963 are shown. A few samples lack IRMS data due to insufficient amount of sample material (m means mass of carbon (m C (ug)) and sample (m (mg)), respectively).

Table 1 only includes CaliBomb dates after the peak in 1963 (all calibrations also return a result between the mid-1950s and beginning of the 1960s). It is however not likely that the results from the rising part of the bomb-pulse curve are relevant in this study. A result from the rise of the bomb pulse would imply that the carbon in the analysed tissue was about 20 years old at the time of collection for all the subjects (not likely, the subjects were between about 30 and 60 years of age in the late 1950s). Another alternative would be that the turnover of cells is very slow resulting in an average F¹⁴C value corresponding to the rise of the bomb pulse. Neither this explanation appears likely. The most plausible explanation seems to be that the carbon in the breast tumour samples mainly originates from recently consumed food.

Figure 3 shows the results from the ¹⁴C measurements (±1σ, results only for the declining part of the bomb-pulse curve) for each subject (the year of collection was 1983). The CaliBomb date of all samples except T3A and T9C overlaps with the year of collection 1983 within 3σ.

Figure 3.

Results from the ¹⁴C measurements (±1σ) for each subject (the year of collection, 1983, is indicated in grey).

Figure 4 shows the δ ¹³C and δ ¹⁵N values obtained for the tumour samples. Sample T7 seems to lie apart from rest of the samples. In contrast to the other samples, T7 contained mainly fat tissue, alternatively necrosis, which might explain the separate location of the samples in the diagram. In stable isotope analysis of human tissues, it is also known that different tissues do not have identical δ ¹³C and δ ¹⁵N values (see e.g. Caut et al. ⁽³⁷⁾ and Tieszen et al. ⁽⁴⁵⁾). Fatty acids, e.g., are known to be depleted in δ ¹³C (having a negative discrimination factor) compared to other tissues as well as diet. Schoeller et al. ⁽⁴⁶⁾ found that δ ¹³C(plasma lipid) < δ ¹³C(plasma protein) in 10 American subjects. Tieszen et al. ⁽⁴⁵⁾ report that δ ¹³C values in gerbil were ranked as fat < liver < muscle < brain < hair (the discrimination factor between diet and fat being about −3‰). The present results for sample T7 follow the same trend as these observations and indicate that δ ¹³C(fat cells/fat tissue) < δ ¹³C(cancer cells).

Figure 4.

δ ¹³C and δ ¹⁵N obtained for the tumour samples (compare with Figure 2). A few tumour samples lack IRMS data due to insufficient amount of sample material.

Figure 5 shows the δ ¹³C and δ ¹⁵N values obtained for the tumour samples including previous measurements of serum and atherosclerotic plaques from Swedish patients (see Gonçalves et al.⁽⁴⁷⁾). It should be noted that the subjects are not the same, neither is the year of collection (serum from 1978 and plaque from 2007–2009). However, in this diagram sample T7 still lies apart from the rest of all sample types. Figure 5 also indicates linear correlations between δ ¹³C and δ ¹⁵N for the various sample types. This can be expected considering the tropic level diagram in Figure 2.

Figure 5.

δ ¹³C and δ ¹⁵N for the tumour samples and previous measurements on Swedish plaque⁽⁴¹⁾ and serum⁽³⁰⁾. Note that serum for the omnivores, in general, has higher δ ¹³C and δ ¹⁵N than vegans and lacto-ovo vegetarians (l-o-veg), in concordance with the generalised isotopic trophic diagram in Figure 2.

The main purpose of the δ ¹³C and δ ¹⁵N measurements of the breast tumour samples was to evaluate if the diet of any of the subjects contained a large proportion of fish products, which could influence the ¹⁴C value (according to Figure 1). High values of both δ ¹³C and δ ¹⁵N (upper right corner of Figure 2) correspond to more fish products in the diet than low values (lower left corner of Figure 2). According to Figures 4 and 5, tumour samples with high δ ¹³C and δ ¹⁵N values (upper right corner) could indicate that these subjects consume more marine foodstuffs than the others (e.g. the diet of subjects 5, 6 and 10 contains more fish products than others). The excess marine influence could subsequently affect the bomb-pulse dating (see Figure 1), and possibly give dates after the collection date (above the grey area in Figure 3). This is not observed (see Figure 3). In view of the limited amount of samples, the magnitude of the uncertainty of the measurements, the variability of discrimination factors and probably variability in growth rate of breast tumours, it is difficult to draw any general conclusions from the findings.

It is evident that further development of the application of ¹⁴C bomb-pulse dating technique for breast cancer research is needed before it can be of use. A main drawback of the technique is that the ¹⁴C content in the tumour tissue originating from the time when the tumour was in its initial stage is small relative to the ¹⁴C content from the later created part of the tumour. Analysis of extracted tumour DNA may be a way to obtain more relevant growth rate information, since 50% of carbon in DNA originates from the latest cell division⁽⁹⁾. A better approach to estimate the onset of the tumour growth would be to isolate and analyse cancer stem cell DNA. The main challenge of this approach will be to obtain samples large enough for the ¹⁴C analysis (at least 10 µg of carbon is required for the system used). An associated difficulty is that the slope of the bomb pulse nowadays is approaching zero. This means that the resulting uncertainty in the age determination increases to inapplicability when approaching present days. Thus bomb-pulse dating of fresh material is only possible for cells or tissues with a slow turnover (several years). For materials with shorter turnover, relevant age information can only be obtained from the steeper part of the bomb-pulse curve, and biological material from biobanks must be used. The available amounts of sample material in the biobanks are however often very limited.

If the estimate of the ¹⁴C turnover rate can be more exact, the information gained can be of value, primarily in the understanding of tumour biology: which is the tumour growth rate? Also, the clinical implications of the dating of breast cancer could be the revision or validation of existing tumour growth models. Furthermore, in breast cancer screening one major drawback is the overdiagnosis of slow-growing indolent tumours. Overdiagnosis occurs when women without symptoms are diagnosed with a disease that will not cause them to experience symptoms or lead to early death⁽⁴⁸⁾. It has been estimated that the overdiagnosis in breast cancer screening is 10–20%^{(49, 50)}. It is not possible to incorporate ¹⁴C bomb-pulse dating in clinical practice since the method is both laborious and expensive and most importantly the atmospheric levels of ¹⁴C are now down at a low level. That is why the ¹⁴C bomb-pulse dating technique applied on breast cancer combined with an analysis of micro- and macro-structural changes in the tumour, in order to find a signature of a slow growing cancer, could be of interest. For example, in parallel to the efforts of obtaining relevant tumour growth rate information, samples of the same breast tumour could be investigated with a high-resolution imaging technique, i.e. synchrotron radiation-based small angle X-ray scattering to detect changes in the collagen structure of the tumour tissue⁽⁵¹⁾. If structural changes in a breast tumour can be correlated to growth rate, this could provide means to find new biomarkers or a morphological signature that can be used to further refine breast cancer diagnosis and ultimately optimise the treatment, e.g. less aggressive treatment of indolent tumours.

CONCLUSIONS

This is a first pilot study assessing the potential of carbon and nitrogen isotope analysis for obtaining growth rate (¹⁴C/¹²C) and dietary (¹³C/¹²C and ¹⁵N/¹⁴N) information for breast cancer research. Stable carbon and nitrogen isotope analysis may give information about the relation between diet and breast cancer development and progression.

The measurements in this limited, initial study could not demonstrate the possibility to obtain age information on the onset of the breast cancer. However, analysis of total tumour material may primarily reflect the current turnover of the carbon in the various components of the metabolically active tumour tissue. Analysis of extracted tumour DNA, or the isolation and analysis of cancer stem cells, may be a way to obtain more relevant growth rate and age information. Furthermore, ¹³C/¹²C and ⁵N/¹⁴N analysis may provide important dietary information of value for studies of possible correlations between diet and tumour development.

FUNDING

This work was supported by Stiftelsen Hedda och John Forssmans fond [2014/12 to K.L.]; the Nils-Magnus och Irma Ohlssons stiftelse för vetenskaplig forskning och utbildning [2015/03 to K.L.] and Stiftelsen för cancerforskning vid Onkologiska kliniken vid Universitetssjukhuset MAS [2014/11 to K.L.].