Significant human beta-cell turnover is limited to the first three decades of life as determined by in vivo thymidine analog incorporation and radiocarbon dating

Abstract

Aims: Diabetes mellitus results from an absolute or relative deficiency of insulin-producing pancreatic β-cells. The turnover rate of adult human β-cells remains unknown. We employed two techniques to examine adult human islet β-cell turnover and longevity in vivo.

Methods: Subjects enrolled in National Institutes of Health clinical trials received thymidine analogs [iododeoxyuridine (IdU) or bromodeoxyuridine (BrdU)] 8 d to 4 yr prior to death. Archival autopsy samples from 10 patients (aged 17-74 yr) were employed to assess β-cell turnover by scoring nuclear analog labeling within insulin-staining cells. Human adult β-cell longevity was determined by estimating the cells' genomic DNA integration of atmospheric (14)C. DNA was purified from pancreatic islets isolated from cadaveric donors; whole islet prep DNA was obtained from a 15-yr-old donor, and purified β-cell DNA was obtained from two donors (ages 48 and 80 yr). (14)C levels were then determined using accelerator mass spectrometry. Cellular "birth date" was determined by comparing the subject's DNA (14)C content relative to a well-established (14)C atmospheric prevalence curve.

Results: In the two subjects less than 20 yr of age, 1-2% of the β-cell nuclei costained for BrdU/IdU. No β-cell nuclei costained in the eight patients more than 30 yr old. Consistent with the BrdU/IdU turnover data, β-cell DNA (14)C content indicated that the "birth date" of cells occurred within the subject's first 30 yr of life.

Conclusions: Under typical circumstances, human β-cells and their cellular precursors are established by young adulthood.

Trial registration: ClinicalTrials.gov NCT00006505.

Figures

Figure 1

In vivo thymidine analog labeling of human pancreas. Costaining of thymidine analog and pancreatic hormones. These images were obtained using an Axioskop 2 microscope with a 10× objective and a 0.63× converter and captured with a Hamamatsu Orca ER digital camera using Open Lab 5.2 software. Left, Pancreatic islet from a 20-yr-old subject, stained with IdU (green), insulin (red), and DAPI for nucleus (blue). The two white arrows show two nuclei positive for IdU, completely surrounded with insulin, providing evidence for β-cell turnover (20× magnification). Middle, Pancreatic islets from a 45-yr-old subject. There were no IdU-positive β-cells. The two white arrows represent cells that have divided outside of pancreatic islets (10× magnification). Right, Pancreatic islet from the same 20-yr-old subject as in left panel. Many proliferating cells within the islets are not mature endocrine cells: IdU (green); somatostatin, glucagon, or pancreatic polypeptide (blue); and insulin (red).

Figure 2

Pronsulin-positive β-cells by flow cytometry. Left, A representative pancreatic islet cell preparation, analyzed by cell size (forward scatter, x-axis) and cellular complexity (side scatter, y-axis). Right, A distinct and sortable proinsulin-positive islet cell population. Middle, Unstained control preparation. SSC-A, Side scatter area; FSC-A, forward scatter area; PE, phycoerythrin.

Figure 3

14C dating of human β-cells. Explanation of 14C curve. The black curve in all graphs shows the atmospheric concentrations of 14C over the decades since 1920. The sharp rise in the atmospheric 14CO2 concentration during 1955–1963 was due to atmospheric testing of nuclear weapons. The decline since 1963 is due to the elevated 14CO2 moving to the oceans and biosphere. As the 14CO2 entered the biosphere, it labeled all living things with the elevated 14C as the carbon moved up the food chain. Because DNA gets new carbon at cell division, the date of DNA synthesis is reflected in its 14C/C concentration. A, The measured 14C concentration plotted on the y-axis (1) is related to the atmospheric 14C concentration on the established atmospheric 14C bomb curve (2). The average birth date of the cell population can be inferred by where the data point intersects on the x-axis (3). F14C notation represents the 14C/C ratio associated with the 14C bomb pulse where F14C = 1 occurred in the late 19th century before burning fossil fuels started to depress the 14C/C ratio of atmospheric CO2 (14). B, 14C concentration in the DNA of β-cell and non-β-cell fractions. Vertical lines mark the birth year, square symbols with the dashed lines represent non-β-cell fractions, and the circles with dotted lines represent the β-cell fraction. The first subject was born in 1926 (blue lines) and was 80 yr old at the time of death from head trauma with a glycosylated hemoglobin of 5.7%. His β-cell 14C level corresponded to 1956 on the black curve. The second subject was born in 1957 (green lines) and was 48 yr old at the time of death from a subarachnoid hemorrhage with a glycosylated hemoglobin of 5.5%. That subject’s β-cell 14C levels corresponded to 1958 or 1988. The data from these two subjects suggest that birth dates of β-cells do not exceed 30 yr after the subjects’ actual birth date. C, 14C concentrations in the DNA from whole islets (red circle) and spleen cells (yellow square) of a subject born in 1990 (15 yr old, healthy, died in a car accident). D, Expanded view of panel C from 1980 to 2010. The 14C concentrations in DNA from whole islets and spleen cells of the same subject born in 1990. The 14C levels in the spleen (yellow square) follow the bomb curve, suggesting 100% annual turnover of DNA. The islets (red circle and line) have a slower than atmospheric decline in 14C, suggesting a 20% annual 14C turnover, which corresponds to 40% annual cell turnover, because only half of the cell’s carbon is renewed when cells divide.