Relative telomere lengths (RTL) measured with flow cytometry-based fluorescence in situ hybridization (flow-FISH) were validated by comparing RTL to telomere lengths previously determined by mean telomere restriction fragment (TRF) Southern blots in a well-characterized normal HMEC strain and an immortal cell line derivative [21
]. Normal pre-stasis finite lifespan (primary) 184D at passage 4 (~11.5
kb mean TRF) and passage 12 (~9.2
kb mean TRF) and immortal non-malignant 184A1 at passage 49 (~4.1
kb mean TRF) exhibited strong correlation with RTL values measured by flow-FISH that were normalized to 184D passage (p)4 (Figure A). Thus, flow-FISH RTL measurements were comparable to the TRF measurements obtained using Southern blotting, however RTL does not provide absolute length of the telomeres in question.
Figure 1 No age-related telomere length reduction was observed in HMEC strains. (A) A linear regression comparing Flow-FISH measured RTL values with mean TRF values in a well characterized pre-stasis strain and an immortal derivative with critically short telomeres, (more ...)
To determine whether telomere lengths differed as a function of age in HMEC, RTL were measured in 41 strains of pre-stasis finite lifespan HMEC at 4th
passage, whose donors ranged in age from 16 to 91 years. Linear regression analysis of RTL measurements showed no significant age-dependent changes in telomere lengths (Figure B), nor did RTL values show significant correlation with total population doublings that the HMEC strains underwent prior to arresting at stasis (Figure C). Previously, we reported age-dependent changes in the distributions of CD227+/CD10- LEP, CD227-/CD10+ MEP, and cKit + HMEC progenitors [1
]. RTL values of the 41 strains at p4 also showed no correlation with the relative distribution of LEP, MEP, and cKit HMEC lineages (Figure D). Thus shortening of normal pre-stasis HMEC telomeres was not observed as functions of age or of lineage distributions.
To determine whether FACS-based RTL measurements would reveal that telomere length varied as a function of lineage in HMEC, as was previously reported from quantitative (Q)-FISH measurements in breast tissue sections [19
], RTL were measured in LEP, MEP (Figure A), and cKit + lineages (Figure B) that were FACS-enriched from p4 HMEC strains derived from a 19y woman (strain 240L) and a 91y woman (strain 805P). Strain 805P had a larger proportion of CD227+
LEP and cKit+
progenitor HMEC compared to the younger 240L strain, as we predicted based on their relative ages. Enrichment of the three lineages was verified by automated image analysis of immunofluorescent staining for intermediate filament protein keratin (K)14 and K19 expression in sorted HMEC. Consistent with our previously reported observations of age-dependent lineage phenotypes, analysis confirmed that in HMEC from the 19y woman the MEP were enriched for the K14+
phenotype, LEP for the K14-
phenotype, and cKit progenitors for the K14+
phenotype (Figure C), whereas in the 91y woman all lineages were unusually enriched for K14 expression consistent with previous observations [1
] (Figure D). Within both strains, the RTL significantly differed by lineage, where both LEP and cKit + cells exhibited shorter telomeres than the MEP (Figure E). Although significant, the differences in RTL between the lineages of the two strains were less than 1.5-fold, and the measurements of mixed populations, as in Figure B, essentially represent average RTL. If anything the average RTL of older strains would be biased in the direction of shortened telomere lengths, thus making our conclusion that HMEC telomeres are unlikely to shorten significantly with age all the more strongly.
Figure 2 Relative telomere length and telomerase activity varies by lineage within cultured strains. (A) Representative FACS analyses of CD227 and CD10 expression in 4th passage HMEC strains isolated from a 19 year old woman (240L) and a 91 year old woman (805P). (more ...)
To determine whether telomerase activity changed with age, TRAP activity was measured in protein extracts from 34 HMEC strains at p4. Among the strains, 12 of 34 exhibited obvious ladders indicating TRAP activity (Figure A). The average age of HMEC strains that exhibited TRAP activity was 29.7y± 3.9y, whereas the average age of TRAP-negative strains was 49.0y± 4.9y (Figure B). This was confirmed by image quantification of the TRAP blots such that the ladder positive compared to the ladder negative group showed activity of 1.2
0.1 versus 0.8
0.1 relative arbitrary TRAP units (P
0.03), respectively. TRAP activity did not correlate with RTL (Figure C) or with the number of population doublings the strains underwent before entering stasis (Figure D). Higher TRAP activity did correlate with decreased proportions of LEP (Figure E) and cKit + HMEC (Figure G) but did not correlate with proportions of MEP (Figure F). Telomere lengths and number of population doublings were independent of the level of detectable TRAP activity. However, more TRAP activity was detected in younger HMEC strains, and therefore also correlated with the age-dependent distributions of LEP and cKit+
progenitors. Consistent with suggestions that telomerase activity may be stress sensitive [23
], we hypothesize that aging-related stresses lead to a global decline in telomerase activity within HMEC.
Figure 3 Telomerase activity in HMEC decreases with age. (A) TRAP gels for 34 strains of 4th passage pre-stasis HMEC. Histograms showing distributions, with their associated means, of (B) age, (C) RTL, (D) PD at stasis, (E) percentage LEP of total, (F) percentage (more ...)