Increased cellular senescence is inevitable in aging organisms and is often ascribed to telomere erosion that prompts replicative senescence. However, senescence can also be induced by various external stimuli, leading to stress-induced premature senescence, a phenomenon observable both in young and aged organisms. The process can be traced by analyzing the level or activity of established senescence markers, although none of them is specific for a given type of senescence. Thus, in this study we propose CLDN1 as a cell-specific marker of stress-induced premature senescence in VSMC.
The basis for the presented research on CLDN1 in senescence has been provided by transcriptomic analysis of prematurely (doxorubicin- and curcumin-induced) and replicatively senescent VSMCs. The microarray analysis indicated that the mRNA level was significantly upregulated in prematurely senescent cells while no changes were observed in replicative senescence when compared to young cells. However, when the mRNA level was analyzed with qPCR, upregulated expression was observed in both types of senescence. A similar observation was described in the original paper discussing SEMP1, where mRNA expression was elevated in replicatively senescent human mammary epithelial cells (HMECs) [14]. Unfortunately, no information regarding premature senescence has been provided. In our experiment, despite the observed higher transcript level in both SIPS and RS VSMCs, increased protein expression was noted solely in DOX and CUR-treated cells, suggesting additional post-transcriptional regulation of the mRNA level, for example, by miRNA [12]. Studies analyzing skin grafts from photoprotected (buttocks), and unprotected (forearm) tissue areas collected from young and elderly donors, provided interesting observations, aligning with our findings to some extent [13]. In those studies, the term “intrinsically senescent cells” can be interpreted as replicatively senescent cells and photoaged skin cells as prematurely senescent cells, as this tissue has been exposed to UV-radiation, which is a commonly known inducer of senescence. In intrinsically senescent cells (RS) claudin-1 expression was significantly downregulated with no changes in mRNA level compared to cells from young donor. This is similar to the phenomenon observed in our experiment. However, when photoaged (SIPS) and photoprotected skin from elderly patients were compared, the decrease in protein and mRNA level were even more pronounced in the sun exposed tissue. Interestingly, no differences were noted between forearm skin and the photoprotected skin from young individuals. It should be noted that these are different types of cells than those used in our experiments. Such conflicting results may be cell type-dependent, as it has been proven that the expression level of claudin-1 in normal or aging cells can be tissue-specific [31]. Our results suggest that in VSMCs, increased claudin-1 expression is a characteristic feature of SIPS.
Our research also aimed to elucidate the reasons behind the differences in CLDN1 expression between SIPS and RS and to understand the underlying epigenetic regulatory mechanisms. Results of the ChIP-seq analysis suggest that expression of the CLDN1 gene may be regulated by the enrichment in H3K4me3 at the transcription start site (TSS). It has been demonstrated that the overall levels of H3K4me3 decrease in senescent cells (unpublished data, manuscript in preparation) while on the other hand, a broad H3K4me3 peak at the TSS of CLDN1 is detected in both types of senescence. This may facilitate transcription, potentially by allowing more efficient binding of transcription factors to the TSS.
As cells senesce, their surface area increases, which may influence the formation of contact points between cells and the potential development of tight junctions (where claudin-1 is located). We observed that in SIPS, cells may form small clusters, and wondered whether the enhanced contact might be the reason for higher protein expression. Similar although less frequent observations were made for RS cells, but this depended on the donor. This led to the expansion of the experimental setup, in which we analyzed both young and RS cells in their nearly 100% confluence state. Additionally, we have investigated whether the higher protein level is associated solely with irreversible cell cycle arrest. To test this, cells were cultured in a serum-free medium for 7 days, resulting in their reaching a quiescent state. Such cells are temporarily arrested in the cell cycle and resume division upon reintroduction of serum into the medium. The Western blotting analysis showed no increase in CLDN1 in either case. The result suggests that the elevated CLDN1 level is not associated with increased intercellular contact or cell cycle exit.
We tested if the increased CLDN1 expression in SIPS cells might result from the prolonged culture, as these cells were incubated for 7 days without passage. Recent studies performed on human corneal epithelial cells (HCECs) showed that prolonged 14-day cultivation led to a significant increase in claudin-1 expression, which was noted from the 6th day of incubation [32]. The 14-day culture was supposed, based solely on cell division halting (cell count), to mimic cellular senescence. However, it is plausible that these cells reached 100% confluence, which potentially led to contact inhibition, and this could not be assigned to cellular senescence per se. Furthermore, the conditions of prolonged and confluent cell culture might have induced cellular stress, contributing to altered protein expression. As mentioned above, our confluent cells did not exhibit an increase in CLDN1 expression. To further investigate the effects of prolonged cell culture, we conducted claudin-1 expression kinetics in VSMCs over four consecutive days. The results, however, were inconclusive and remain under investigation.
Immunocytochemical staining demonstrated that in PS cells CLDN1 localizes predominantly within the cell nucleus, which was confirmed on the protein level by analyzing the cytoplasmic and nuclear cell fractions. This nuclear localization is uncommon, and the consequences of altered localization are not well understood. The existing data are contradictory and mostly focused on cancer cells. For example, studies have shown that in thyroid cancer cells, the translocation of CLDN1 to the nucleus and increased protein levels led to enhanced cell proliferation and migration [33]. In contrast, the increase in cytoplasmic fraction in melanoma cells drove metastasis and worsened survival outcomes, with nuclear localization having little effect on cell migration [34]. In colon cancer cells, however, increased levels of CLDN1 in both the cytoplasm and nucleus were associated with enhanced cell migration [35].
Following the preliminary characterization of CLDN1 expression in VSMCs, we employed it as a marker of SIPS in cells isolated from atherosclerotic plaques of six donors. To assess the proliferative state of these cells, we used ANLN. Our analysis revealed that all six samples predominantly consisted of non-dividing cells, as indicated by the absence of the ANLN band in Westen blotting.
Focusing on CLDN1 expression, only samples 3 and 4 displayed strong bands in Westen blot indicating high CLDN1 level, suggesting that these cells might have undergone premature senescence. In contrast, samples 1 and 2 exhibited SIPS and replicative senescence (RS) characteristics, though the proportion might be unbalanced given the faint CLDN1 bands observed. Samples 5 and 6, on the other hand, appeared to have a predominant population of RS cells. CLDN1 expression differences in isolated smooth muscle cells reflect the heterogeneity of atherosclerotic plaques, some of which are stable, some unstable and may cause vascular incidents. Better characterization of atherosclerotic plaques may allow for the selection of clinically significant changes, especially in coronary and cerebral arteries.
We also conducted parallel analyses on fibroblasts using the same experimental approach (results not shown). While the pattern of increased CLDN1 expression in SIPS fibroblasts mirrored that observed in VSMCs, the results related to RS were inconclusive. This area remains under investigation.
Claudin-1 is a widely studied protein; however, its role is mostly described in cancer cells and other malignancies or ailments (tissue fibrosis, stroke, wound healing etc.) [8, 36–38]. Its dysregulation can be used as a prognostic marker of breast cancer [39] or sleep apnea [40], and has recently been used as a therapeutic target of colon cancer [41], hepatocellular carcinoma [42] or tissue fibrosis [37]. So far, no research has been conducted on the role of CLDN1 in atherosclerosis and we propose that CLDN1 might serve as a stress-induced senescence marker in VSMCs and perhaps a future therapeutic target in this condition.