1. Quantification of T-cell dynamics from telomeres to DNA labelling José Borghans Dpt. Immunology University Medical Center Utrecht

2. How long does a T cell live? Estimates vary widely How is T-cell turnover disturbed in HIV infection, leukemia, or after stem cell transplantation?

3. What is the difficulty? How to follow a lymphocyte from its birth to its death? Extrapolation from mice to men v.v.

4. 3 years of age 23 years of age Mackall et al. 1995 Experiments of nature (1): T-cell reconstitution after chemotherapy Underestimate?: Cells may also die during reconstitution Overestimate?: Cells undergo little competition

5. Michie et al. Nature 1992 CD45RA (“naive”) CD45RO (“memory”) Immunological memory resides in a population with rapid turnover: Memory T cells have a shorter lifespan (~250 d) than naive T cells (~1000 d) Experiments of nature (2): Loss of T cells with chromosome damage Caution: Cells have DNA damage, and cell numbers are low

6. Static versus dynamic markers of T-lymphocyte turnover Static Ki67-expression (protein expressed in G 1,S,G 2,M phase) Annexin V staining (stains phosphatidylserine translocation) Dynamic Natural markers: T-cell telomere lengths T-cell receptor excision circles Labelling: (CFSE labelling) BrdU labelling Stable isotope labelling

7. Weng et al. PNAS 1995 Changes in telomere lengths are no direct measure of T-cell division

8. Changes in TRECs do not directly reflect thymus output Douek et al. Nature 1998

9. Static versus dynamic markers of T-lymphocyte turnover Static Ki67-expression (protein expressed in G 1,S,G 2,M phase) Annexin V staining (stains phosphatidylserine translocation) Dynamic Natural markers: T-cell telomere lengths T-cell receptor excision circles Labelling: (CFSE labelling) BrdU labelling Stable isotope labelling

10. Dynamic markers (1) BrdU Labelling BrdU = 5-bromo-2’-deoxyuridine Nucleoside analog incorporated instead of thymidine

11. Determine percentage BrdU+ cells by FACS analysis Kovacs et al. 2001 How to quantify leukocyte turnover?

12. Model for BrdU labelling up-labelling: determined by p+d down-labelling determined by p-d

13. Expected changes in the percentage of BrdU+ cells p-d p+d See today’s exercise But… possible toxicity, almost only done in mice, only short-term labelling Kovacs et al. 2001

14. Deuterium ( 2 H) replaces hydrogen in DNA of proliferating cells Dynamic markers (2) Stable isotope labelling Measure deuterium-enrichment using GC/MS Main advantage: no interference with cell dynamics

15. 2 H 2 -glucose 2 H 2 O Intravenously Hellerstein et al. 1999, 2003 McCune et al. 2000 Mohri et al. 2001 Ribeiro et al. 2002 Macallan et al. 2003, 2004 Wallace et al. 2004 Long-term administration Hellerstein et al. 2003 Vrisekoop et al. 2008 Stable isotope labelling

16. Enrichment in DNA of cells Fraction labeled DNA (!) L changes by: Label U(t) p -  L - dL 2 H 2 O labeling does not distinguish between production in thymus and periphery! Yields two parameters: p and d p from upslope, d from up- and downslope

17. Asquith et al. 2002 d p Typically: p < d Paradox: d is typically larger than p Conclusion: d is no good measure of average turnover rate It represents the turnover of labelled cells

18. Short labelling periods give higher estimate of d Asquith et al. 2002

19. Average turnover (p) needs data during up-labeling period D-glucoseD2O Advantages of heavy water: long-term labeling possible many data points during labeling

20. BrdU Potentially toxic Measures labeled cells Up: p+d Down: p-d Deuterium labeling Non-toxic (non-radioactive) Measures labeled DNA strands Up: p (and d ) Down: -d

21. NaiveT-cell dynamics of mice and men

22. From mice to men... (How) can we extrapolate? 2 years 80 years mice men 2 years 80 years mice men How to scale? Young/old mouse = young/old human? Effect of thymic involution during aging?

23. Of mice... Naive T-cell dynamics using 2 H 2 O H-O-H Label acquisition: - proliferation - thymic output Loss of label: - cell death - differentiation/migration C57Bl/6 2H2O2H2O 4w % labeled time

24. Short lifespan of naive T cells in mice Expected lifespan: Naive CD4: 6 weeks Naive CD8: 11 weeks

25. And men... T-cell dynamics using 2 H 2 O H-O-H % labeled time ~ 400 ml D2O on day 0 ~ 50 ml D2O daily during 9 weeks

26. T-cell turnover in healthy individuals Vrisekoop et al. PNAS 2008 Lifespan 6.0 9.4 0.6 1.0 year

27. T-cell turnover in young healthy men Vrisekoop et al. PNAS 2008 200 Percentage labeled 20 40 60 0 50 100150 Time (days) 20 40 60 0 0 Naive CD4 Naive CD8 Young adult mice (12 week old) Lifespan 6.0 9.4 0.6 1.0 year

28. Labeled naive T cells present 3 years after stop of label For all fits of naive T cells: p = d All naive T-cells are very long-lived in humans

29. Scaling from mice to men... 2 years 80 years mice men Mice live 80 weeks naive CD4 T cells 6 weeks, CD8 11 weeks Humans live 80 years naive CD4 T cells 6 years, CD8 10 years

30. From mice to men... (How) can we extrapolate? 2 years 80 years mice men How to scale? Young/old mouse = young/old human? Effect of thymus involution during aging?

31. Thymus involution in mice and men MiceMen Steinman et al. 1985

32. 2 H 2 O labeling does not distinguish between production in thymus and periphery Proliferation Thymic output Differentiation Naïve Death Label acquisition: - thymic output - proliferation Loss of label: - cell death - differentiation/migration Fraction labeled DNA Time in days

33. Distinguishing between T-cell proliferation and thymic output ATx or ShamTx at 7 weeks of age Isolation of spleen, PLNs and thymus FACS analysis Mathematical modelling C57BL/6

34. Naive T-cell dynamics after thymectomy N’ = ε f(t) + r n N – d n N 2 Naive T cells: Naive CD4 (spleen) Naive CD8 (spleen) Tx Control Tx Control ε = 2.8 10 -2 r n = – 3.0 10 -3 d n = 1.7 10 -9 Cells (x10 6 ) Mouse age (weeks) ε = 2.8 10 -2 r n = – 2.0 10 -3 d n = 6.8 10 -10 0.9% of thymocytes per day

35. Naive T-cell production in mice almost completely due to thymic output! Loss Thymic output Heavy water Peripheral proliferation Total Naive T-cell production Naive CD4  p = 0.28 10 6 / day Naive CD8  p = 0.13 10 6 /day Thymectomy Thymic output Loss Peripheral proliferation Thymic output: CD4 thymic emigrants  0.35 10 6 / day CD8 thymic emigrants  0.12 10 6 / day Contribution of thymic output and T-cell proliferation in mice

36. If everything depends on thymic output, what happens when thymic output declines? Young mice:12 weeks 43 days 77 days 42 days 91 days Old mice: 85 weeks p = 0.023p = 0.024 p = 0.013p = 0.011 Time (days)‏ ?

37. Naive T-cell dynamics in mice hardly change with age The naive pool of a mouse is a big homogeneous pool of thymus emigrants Young mice:12 weeks 43 days 77 days 42 days 91 days Old mice: 85 weeks p = 0.023p = 0.024 p = 0.013p = 0.011 Time (days)‏ Naive T-cell proliferation hardly occurs at any age in mice!

38. This is completely different in humans! Evidence for naive T-cell proliferation in men…

39. TREC decline with age originally interpreted to reflect thymus decline Child Adult Douek et al. Nature 1998Rodewald Nature 1998

40. If naive T-cells proliferate homeostatically, TREC contents do decline Naive TREC decline in humans suggests that naive T-cell proliferation contributes to the naive T-cell pool in humans Hazenberg et al. 2000, Dutilh et al. 2003

41. CD31 proposed as marker for non- divided naive CD4 T cells NaiveMemory CD31+CD31- thymus CD31+ naive T cells have consistently higher TREC content than CD31- Suggestion: CD31+ naive T cells represent (non-divided) thymic emigrants CD31- naive T cells have divided Kimmig et al J. Exp Med 2002

42. However, even CD31+ naive T cells divide in humans! See also Kilpatrick et al. 2008 Age (years) 01020304050607080 Sj TREC content/ cell 0.0001 0.001 0.01 0.1 1 CD31+ CD31- CD4 SP Age (in years) TREC content per cell

43. Fewer thymus emigrants (TE) than CD31+ naive T cells NaiveMemory CD31+CD31- thymus NaiveMemory CD31- thymus CD31+ TEnon-TE

44. T cells: TRECs: TREC content: Which part of T-cell production comes from the thymus? Adapted from Hazenberg et al. 2000  t  pN +  (t) A/c =

45. Use TREC content of cord blood or SP thymocytes to measure c See also Kilpatrick et al. 2008 Age (years) 01020304050607080 Sj TREC content/ cell 0.0001 0.001 0.01 0.1 1 CD31+ CD31- CD4 SP Age (in years) TREC content per cell  t  pN +  (t) A/c =

46. In human adults <10% of naive T cells is produced by the thymus Fraction naive cells produced in thymus Age (in years) Based on CD31 Based on TREC contents

47. Extrapolating from mice to men RTE Proliferation RTE Proliferation Men 2 years 80 years mice men Prediction: without proliferation there should be no TREC dilution in mice…

48. And indeed... no TREC decline in naive T cells from mice

49. NaiveT-cell dynamics of mice and men “The best-laid schemes o' mice an' men, Gang aft agley” The most carefully prepared plans may go wrong Quantification of T-cell dynamics from telomeres to DNA labelling