Human Lung Epithelial Cells Divide >200 Population Doublings without Engaging a Telomere Maintenance Mechanism

The “Hayflick limit” is a “mitotic clock” and primary cells have a finite lifespan that correlates with telomere length. However, introduction of the telomerase catalytic protein component (TERT) is insufficient to immortalize most, but not all, human cell types under typical cell culture conditions. Originally, telomerase activity was only detected in cancer cells but is now recognized as being detectable in transit amplifying cells in tissues undergoing regeneration or in extreme conditions of wound repair. Here we report that in vitro low stress culture conditions allow normal human lung basal epithelial cells to grow for over 200 population doublings without engaging any telomere maintenance mechanism. This suggests that most reported instances of telomere-based replicative senescence are due to cell culture stress-induced premature senescence. One Sentence Summary Human lung cells growing in reduced stress conditions can divide well beyond the Hayflick limit.

as originally described for fetal human lung fibroblasts is due to critically shortened telomeres or cell culture shock-induced premature senescence (9).
Telomerase is a conserved ribonucleoprotein enzyme complex (10) that uses an RNA template to reverse transcribe and add TTAGGG n DNA sequences at mammalian chromosome ends during DNA replication (11,12). Initially, telomerase activity was only associated with advanced human tumors and cancer cell lines while most somatic tissues tested were telomerase negative (13). Further investigations detected telomerase activity in a subset of fast proliferating normal human cells, including hematopoietic tissues (14)(15)(16)(17), skin (18,19), hair follicles (20), and intestinal mucosa (21). Epithelial cell turnover in the lung occurs at a slower pace in the absence of damage. For example, ciliated tracheal cells have a half-life of six months and ciliated bronchial cells have a half-life of seventeen months in mouse lung (22).
While telomerase activity has not been previously reported in adult human lung tissue, it may be transiently expressed in basal progenitor cells during injury repair. The important role of telomerase in regeneration of adult tissue is underscored by the manifestation of genetic diseases in individuals with telomere spectrum disorders, ranging from bone marrow failure to idiopathic pulmonary fibrosis (23,24). Lung basal progenitor cells have been shown to have long-term replicative capacity in vivo for lung repair and regeneration (25,26), but senesce in standard in vitro culture conditions after several passages. Recently, we demonstrated that less stressful conditions for long-term expansion of primary human bronchial epithelial basal cells (HBECs) in vitro include co-culturing with an irradiated fibroblast feeder layer, ROCK inhibitor, and 2% oxygen (ROCKi conditions) (27). While differentiated lung epithelial cells are exposed to 21% atmospheric oxygen in vivo, basal lung stem cells residing near the basement membrane are exposed to much lower oxygen levels. Therefore, low oxygen culture conditions reduce the oxidative stress and DNA damage that occur in standard cell culture conditions that is not representative of in vivo conditions (28). For these reasons, the improved ROCKi conditions were modified from conditional reprogramming of cells (CRC) as originally described (29) to include 2% oxygen in addition to a change from standard epithelial cell proliferation F-media to a more defined Bronchial Epithelial Growth Media (BEGM) (27). In CRC conditions and 21% oxygen, primary HBECs only grow for about 50 population doublings (29). Replacement of the fibroblast feeder layer with pharmacological inhibition of PAK1-ROCK-Myosin II and TGF-β signaling also extends primary HBEC proliferation in vitro to 50 population doublings, but accumulation of large cells preceded senescence and correlated with shortened telomeres (30).
We hypothesized that HBECs in ROCKi conditions would exhibit an extended lifespan compared to HBECs in standard culture conditions, but would senesce when telomeres reached a critically short telomere length or engage a telomere maintenance mechanism. Here we report the in vitro culture of primary HBECs well beyond the Hayflick limit without engaging a telomere length maintenance mechanism for over 200 population doublings ( Figure 1A, blue line).

Results and Discussion
When HBECs are switched from less stressful ROCKi conditions back to standard growth media and 2% oxygen around PD 70 ( Figure 1A, red line) or PD 200 ( Figure 1A, green line), cells only proliferate for a few passages before undergoing cell culture stress-induced premature senescence. This indicates that normal cell cycle checkpoints are intact and are activated in more stressful culture conditions. Measurement of telomere length over time ( Figure 1B) shows gradual shortening of telomeres up to PD 214, almost 4-times the number of doublings determined by Hayflick. Mouse 3T3 J2 lethally irradiated feeder fibroblasts are telomerase positive with long telomeres and were depleted by culturing HBECs for two passages in filtered media conditioned by the feeder layer for telomere length assays. To get an accurate measurement of the shortest telomeres which closely correlate with cellular senescence (31,32), the telomere shortest length assay (TeSLA) was employed. TeSLA quantitates the length of all the shortest telomeres that are below detection using other telomere measurement methods (33). TeSLA showed a gradual shortening of telomere length up to PD 214, at which point the average telomere length was ~1 kb with about 86% of telomeres below 1.6 kb ( Figure 1C). In standard growth conditions, primary HBECs only proliferate for twenty to thirty PD (27,34), at which point telomeres are still well above an average of 4 kb. Therefore, the growth arrest of primary HBECs in standard growth conditions is likely due to stress-induced premature senescence rather than critically short telomeres.
At PD 214, the growth rate of ROCKi HBECs slowed down and the population of cells appeared to be undergoing telomere-induced replicative senescence as demonstrated by increasing numbers of large and β-galactosidase positive cells ( Figure S1). Prior to terminating the experiment, HBECs were seeded at clonal density in ROCKi conditions to partially phenocopy a lung injury to determine if there were any rare cells with increased growth capacity within the remaining cell population. Interestingly, we observed that about 15% of the cells seeded were able to grow into robust colonies that could be serially passaged. Telomere length of ROCKi HBEC colonies measured by TRF (Figure 2A) and TeSLA (Figure 2B), showed gradual increase of telomere length between PD 240 and PD 335 that then plateaued and was maintained at an average of 1.7 kb for an additional 200 PDs. The change in telomere length over time measured by TRF and TeSLA was supported by telo-FISH on metaphase spreads of ROCKi HBECs from low, mid, and high PD ( Figure 2C). Prior to subculturing HBECs at clonal density after PD 214 there was no evidence for subsets of cells with longer telomeres as determined by TeSLA analyses ( Figure 1C) and telomere in situ hybridization (data not shown).
To investigate the telomere maintenance mechanism employed by HBECs in ROCKi conditions after clonal expansion, we first tested for the telomerase-independent mechanism of alternative lengthening of telomeres (ALT). The C-rich circular DNA characteristic of cells using . CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint . http://dx.doi.org/10.1101/474270 doi: bioRxiv preprint first posted online Nov. 19, 2018; ALT was not detected in low, mid, and high PD ROCKi HBECs ( Figure 3A). Also, ALT associated PML bodies, a hallmark of most ALT cells (35), did not co-localize with telomeres ( Figure 3B). Finally, the absence of ALT was further supported by the lack of long and heterogeneous telomeres (Figures 2A, B) (35). To rule out that the observed replicative lifespan extension was due to chromosomal end-fusions, we performed Bal 31 digestion, which is a highly specific single/double-stranded exonuclease that digests telomere ends, which would be protected from digestion if end-to-end fusions occurred ( Figure 3C). Bal31 digestion resulted in complete degradation of telomere DNA, indicating there were no chromosome end-fusions "protecting" telomeres. Additionally, telomere fusions were not observed in late passage HBECs using pancentromere-telomere FISH (data not shown) and chromosome counts in ROCKi HBECs ( Figure 3D).
We next tested if telomerase reactivation could be detected after PD 214. Droplet digital PCR quantitation of telomerase activity using the telomere repeat amplification protocol (ddTRAP) for HBECs at PD 314 indicated low, but above background levels, of telomerase activity compared to cells at PD 32 and 235 ( Figure 4A), suggesting telomerase may be transiently expressed in a subset of cells. To investigate whether the low levels of telomerase activity were functional, we transfected both PD 14 (long telomeres) and PD 324 (critically short telomeres) with PX458 CRISPR/Cas9 plasmid expressing guide RNA to knockout the TERT gene. CRISPR transfection resulted in about 80% reduced growth of colonies in PD 324 HBECs but not PD 14 HBECs ( Figure 4B). If late passage cells did not have telomerase we would have expected similar colony formation to early passage HBECs. To further test the role of telomerase in late PD ROCKi HBECs, we treated HBECs with the telomerase mediated telomere targeting agent 6-thio-2'deoxyguanosine (6-thio-dG), which has been shown to induce rapid telomere dysfunction and cell death in telomerase positive cells, but not telomerase negative cells (36). Treatment of PD 14 HBECs with 6-thio-dG for 3 clonal passages (~30 PD to . CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint . http://dx.doi.org/10.1101/474270 doi: bioRxiv preprint first posted online Nov. 19, 2018; PD 44) did not result in reduced colony size and number but PD 233 colonies were growing slowly and in the presence of 6-thio-dG colony formation was essentially eliminated after 3 clonal passages in 6-thio-dG to PD 263 ( Figure 4C). The results of both the CRISPR and 6thio-dG experiments demonstrate low density cloning of ROCKi HBECs after PD 214 may reactivate telomerase activity analogous to injury repair in the lung but only at higher passage number when telomeres are critically short. We previously reported the human TERT gene is very close to a telomere end on chromosome 5p and that TERT can be transcriptionally activated when telomeres are short using a mechanism termed telomere looping or telomere position effects over long distances that may explain the activation of telomerase in these cells at late PD (37). We suggest that senescence of cultured cells results from both extrinsic and intrinsic factors. One set of senescence triggers is the extrinsic culture environment that can be initiated by a multitude of disparate stress factors, while the second set of senescence triggers is intrinsic and depends on the machinery that monitors telomere length (9). In addition, as recently reviewed (38), the senescence research field would profit from clearly defining more specific descriptors of the term cellular senescence. In summary, we have developed methods for maintaining normal human lung epithelial cells in culture long-term that should expand the use of these cells for basic research, genetic engineering that requires clonal expansion, regenerative medicine and drug screens.

Materials and Methods
3T3 J2 cell culture.
The 3T3 (J2 strain) Swiss mouse fibroblast cell line was purchased from Tissue Culture Shared Resource (TCSR), Lombardi Comprehensive Cancer Center, Georgetown University and tested . CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint . http://dx.doi.org/10.1101/474270 doi: bioRxiv preprint first posted online Nov. 19, 2018; negative for mycoplasma. The cell line does not produce murine viruses and was irradiated at 30 Gray (Gy) with gamma radiation to provide an irradiated fibroblast feeder layer.

Telomere Shortest Length Assay (TeSLA)
The TeSLA was performed as described previously (33) to measure the average and the shortest telomere lengths.

Collection of chromosome spreads
HBEC cells were treated with 5µM colchicine (Sigma-Aldrich) and incubated for 12 hours at 37°C to enrich the fraction of mitotic cells. Successively, cells were collected and re-suspended in 75mM KCl hypotonic solution (Sigma-Aldrich, St. Louis, MO) for 20 minutes at 37°C, followed . CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
Finally, cells were dropped onto slides and air dried.

Ploidy Analysis
Ploidy was evaluated performing chromosome count. Cells treated with colchicine and fixed as reported in the previous section, were stained and mounted 4′,6-diamidino-2-phenylindole . CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
After 48 hours, slides were rinsed in PBS 1X for 5 minutes and then fixed in 4% formaldehyde in PBS at 4°C for 20 minutes. Slides were permeabilized with 0.1% Triton X-100 (Sigma) in PBS at RT for 10 minutes and blocked with BSA 10% in PBS at 37°C for 30minutes. Cells were then incubated with the mouse primary anti-PML antibody (PG-M3; Santa Cruz Biotechnology, Santa Cruz, CA) diluted 1:100 in PBS for 3 hours at RT. After washing with 0.05% Triton X-100 in PBS for 5 minutes, cells were incubated with a secondary goat anti-mouse IgG conjugated with Alexa 488 (Invitrogen, Carlsbad, CA) diluted 1:300 in PBS for 1 hour at RT. After immunostaining, telomeric FISH was performed as described in "Telomeric Qualitative FISH" section. Pictures were taken at 100X magnification with multiple Z-stack (5-10) at 0.4µm intervals using an Axiovert 200M Fluorescent Microscope (Carl Zeiss). Images were analyzed with ImageJ v1.52a.

C-circle assay.
The C-circle assay was performed as described previously (39). Briefly, Hinf1/Rsa1/Alu1digested genomic DNAs were incubated with phi29 polymerase (New England Bio) and reaction buffer (0.75 U of phi29 polymerase, 0.1 mg/ml bovine serum albumin (BSA), 0.2 mM deoxynucleoside triphosphate mix, and 1× phi29 buffer) at 30°C for 12 h and then at 65°C for 20 min. Samples were loaded into slot blots and then hybridized with 32 P-labeled C-probe (CCCTAA) 3 under native conditions to measure the amplified C-circle level.

Droplet Digital TRAP Assay (Telomerase Activity)
Quantitation of telomerase enzyme activity was performed as described previously (40).
. CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.

6-thio-deoxyguanosine Treatment.
For treatment at high density, HBECs were seeded in triplicate at 30,000 cells per well of a 6well plate with 8x10 4 irradiated 3T3s, in BEGM with 5% FBS and 10 µM ROCK inhibitor and incubated in a 5% CO 2 incubator at 37⁰C for 7 days. Cells were treated with 3 µM 6-thiodeoxyguanosine or DMSO/water (1:1) control 24 hours after seeding and then every 3 days. On the final day, cells were counted and reseeded for further treatment. For treatment at clonal density, HBECs were seeded in triplicate at 100 cells per well of a 6-well plate with 8x10 4 irradiated 3T3s, in BEGM with 5% FBS and 10 µM ROCK inhibitor and incubated in a 5% CO 2 incubator at 37⁰C for 10 to 14 days. Cells were treated with 3 µM 6-thio-deoxyguanosine (Metkinen Oy, Kuopio, Finland) or DMSO/water (1:1) control 24 hours after seeding and then every 3 days. On the final day, media was aspirated and cells fixed/stained with 6% glutaraldehyde/0.5% crystal violet (Sigma) for 1 hour at room temperature with rocking. Dishes were rinsed with warm water until water was clear and the resulting clones in each condition were imaged and counted.
. CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.

β-galactosidase Staining
A senescence β-galactosidase staining kit (Cell Signaling, Danvers, MA) was used to perform βgalactosidase Staining according to the manufacturer's instructions.

Material Availability
All data are available in the manuscript or the supplementary materials.
. CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.      HBECs transfected with CRISPR/Cas9 hTERT plasmid did not show a large change in colony growth (11.2%) compared to untransfected cells with 13.5% colony growth. C) Treatment of ROCKi HBECs with 3 µM 6-thio-dG for 3 passages at clonal density inhibits colony formation compared to control as demonstrated by reduced colony size and number observed in PD 263 colonies after crystal violet staining.
. CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.