Lung fibroblasts from patients with emphysema show a reduced proliferation rate [1, 2], altered growth factor response [3] and lower number of population doublings in long-term culture [1]. Together with clinical observations, these findings lend support to the hypothesis that premature aging of structural cells is involved in the pathogenesis of emphysema. Senescent cells not only loose their ability to divide and respond to mitogenic stimuli but also display alterations in morphology and metabolic profile [4]. This phenotype can be induced by oxidative stress [5], in association with epigenetic changes in gene expression [6, 7]. As fibroblasts provide part of the lung's structural support and matrix that is essential for its integrity [8], a senescent phenotype could affect tissue microbalance and structural maintenance of the lung. We thus focused on lung fibroblasts as important players, keeping in mind that it is unlikely that alterations found in these cells are strictly limited to this type of structural cell.

One well-known marker of cellular senescence is senescence-associated β-galactosidase (SA-β-Gal) [9, 10]. Its expression depends on confluence [11] and aged cells are positive for SA-β-Gal most likely due to an increased lysosomal content [10].

Among the mechanisms implicated in cellular aging, the telomere hypothesis [12] is based on the fact that telomere length is reduced in each cell division. A length below a critical value induces cell cycle exit and thereby limits the cell's replicative capacity. Indeed, telomeres shorten during aging of cultured fibroblasts [13] and their initial length correlates with replicative capacity [14]. However, an unaltered telomere length would not disprove the hypothesis of aging, as replicative senescence can also be mediated by telomere-independent mechanisms [4].

To elucidate further potential mechanisms, targets selected from an exploratory 12 k cDNA array analysis were reevaluated by quantitative PCR (qPCR), with emphasis on genes related to proliferation and aging. We focused on insulin-like growth factor-binding proteins (IGFBP), as they might mediate between systemic and local alterations in COPD. IGFBP-3 [15] and IGFBP-related protein (rP)-1 (IGFBP-7) [16, 17] are associated with senescence, and IGFBP-5 is involved in regulating lung matrix composition [18] and development [19]. It was found to be downregulated with increasing age [20] but upregulated in whole lung samples from severe emphysema [21]. IGFBP-rP2 (CTGF, connective tissue growth factor) and IGFBP-rP4 (Cyr61, cysteine-rich angiogenic inducer 61) are also of interest in this respect [22]. To cover a broad mechanistic spectrum of further candidates that are known to be implicated in cell cycle regulation or senescence, we selected FOSL1 (fos-like antigen 1, Fra-1), a family member of Fos transcription factors [23], LOXL2 (lysyl oxidase-like 2), a member of the lysyl oxidase (LOX) family [24], OAZ1 (ornithine decarboxylase antizyme 1), an inhibitor of the ornithine decarboxylase [25], and CDK4 (cyclin-dependent kinase 4).

Thus the aim of the present study was to further characterize the phenotype of primary parenchymal lung fibroblasts in emphysema and to obtain further clues regarding the hypothesis that premature cellular aging plays a role in this disease. For this purpose we compared SA-β-Gal activity, telomere length, and the expression of a selected panel of genes between lung fibroblasts from patients with emphysema and control patients.

As a result we found that a higher proportion of fibroblasts from patients with emphysema exhibited SA-β-Gal activity and that these cells showed an increased expression of senecence-associated IGFBP-rP1 and IGFBP-3 genes and of IGFBP-3 protein, whereas no difference in telomere length could be detected compared to fibroblasts from controls.

In the present study we found an increased staining for SA-β-Gal and a qPCR-confirmed upregulation of senescence-associated IGFBP-3 and IGFBP-rP1 in cultured primary parenchymal lung fibroblasts from patients with emphysema; this was supplemented by detection of higher protein levels of IGFBP-3. A comprehensive exploratory microarray analysis suggested that more genes were down- than upregulated in emphysema, though a number of differences could not be confirmed in qPCR. Taken together with the already known reduction in proliferation rate and capacity, these findings provide further evidence for a senescent phenotype of lung fibroblasts in emphysema, in line with the hypothesis, that premature aging of these cells is one of the relevant pathogenetic factors. As mean telomere length was unaltered, the senescent phenotype is more likely to be mediated by telomere-independent mechanisms.

Previous studies already demonstrated that lung fibroblasts from patients with emphysema exhibited a reduced proliferation rate and capacity in vitro [1, 2]. An increase over time in the proportion of senescent, cell cycle-arrested cells could well be a contributor to tissue destruction. It seems conceivable that such deficiencies favour the onset of emphysematous lesions, and indeed such alterations have been found in senescence-accelerated mice [28]. To check this hypothesis, we first assessed the proportion of cells staining positive for SA-β-Gal, which is considered as a marker of cellular senescence [9]. For this assay we compared the staining between groups after comparable culture times in vitro, as a rise in the percentage of SA-β-Gal positive cells can also be observed during aging of cells in culture.

Telomere length, an important marker of cellular aging, which represents a mitotic clock counting down in aging cells, was similar in emphysema and controls. The assay employed is an established procedure and has been successfully used to reveal, for example, shorter telomere lengths in lymphocytes of smokers [29]. The validity of our data was indicated by the similar pattern observed in the duplicate determinations, as well as by the fact that telomere length was close to previously reported values [13]. It might be argued that fibroblasts in emphysema underwent more replications in vivo due to the need for repair of tissue damage and therefore should have shorter telomeres. The characteristics of cell proliferation curves [1] suggest that fibroblasts from emphysema display replicative senescence about 6 population doublings earlier than controls. Assuming a shortening by about 50 bp in each fibroblast replication [13], this difference would correspond to telomeres being about 300 bp shorter in emphysema compared to controls. In opposite to this, mean telomere length as measured in the present study was 400 bp greater. This implies a difference in length of up to 700 bp contra hypothesis which renders it unlikely that shortening of telomeres explained the difference in fibroblast phenotypes. This is true even though the scatter was large and the number of patients investigated was limited. In fact, a statistical analysis showed a less than 5 % probability of obtaining the observed result if the hypothesis of shortened telomeres in emphysema was true. In addition we would like to note that the experiments were performed in early passages. Thus it seems unlikely that the higher in vitro proliferation rate of controls diminished a potential difference to an extent, that it was even reversed into the opposite.

This suggests the presence of telomere-independent replicative senescence which is a well-known phenomenon potentially involving a variety of pathways, including p16 [4, 30]. On the basis of this, it does not seem likely that telomere length was the major determinant of the observed alterations in emphysema. It certainly would not explain the differences in proliferation rate, SA-β-Gal staining and gene or protein expression that occurred at comparable telomere lengths.

Two cDNA arrays were used to find hints on differentially expressed genes under baseline culture conditions. mRNA of fibroblasts from patients typical of their group was pooled and analyzed. Based on the results and a comprehensive literature study, the expression of selected genes was then reevaluated in independent cultures from individual patients. As the available cDNA was limited, we focused on a small number of genes associated with senescence and cell cycle, which appeared interesting or novel with regard to the pathogenesis of emphysema. Special attention was paid to using only fibroblasts from cultures with a reproducible proliferation rate to ensure comparability with previous results.

Among the genes that were most upregulated on the array was IGFBP-rP1, whose expression is known to increase during senescence [17]. This family of compounds appeared of particular interest, as it might also provide a bridge between local and systemic effects in COPD via insulin-related pathways, similar to IGFBP-3 and -5. For IGFBP-3 and IGFBP-rP1 the upregulation in emphysema was confirmed by qPCR. Furthermore, increased concentrations of IGFBP-3 were detected in cell culture supernatants of fibroblasts from patients with emphysema. In the qPCR analysis there was also a trend (p = 0.07) towards upregulation of IGFBP-rP2, which had been previously described as overexpressed in lung fibroblasts from emphysema, together with IGFBP-rP4 [22]. We believe that the facts that these authors studied patients with more severe emphysema, as well as differences in methodology are responsible for the differences between the findings.

The upregulation of IGFBP-3 and -rP1 can be taken as further evidence for a senescent phenotype in emphysema. As these proteins interact with mitogenic compounds such as insulin-like growth factor I and II (IGF-I, II) or insulin, an active role for IGFBPs in senescence might well be assumed. Both IGF-I and -II are produced by interstitial mesenchymal cells, epithelial cells and macrophages within the lung, as known from studies in lung fibrosis, and can regulate cell proliferation, especially in fibroblasts [31]. Stimulation of the IGF-I receptor by IGF-I, IGF-II [32] or insulin [33] can promote cell division, possibly in synergy with EGF/EGFR and/or TGF-α [34]. The interaction of IGF-I, -II and insulin with their receptors is largely regulated by IGFBPs and their related proteins [32]. Specifically, elevated mRNA [15, 30] or protein levels of IGFBP-3 were found in late passage/senescent fibroblasts [15] and IGFBP-3 is capable of interacting with IGF-I [32]. IGFBP-rP1 can inhibit the growth of cancer cells via a senescence-like mechanism, associated with SA-β-Gal staining [16]. IGFBP-rP1 was also found upregulated in senescent human mammary epithelial cells [17]. Through binding to insulin it can prevent signal transduction towards proliferation. Though the picture regarding the insulin and IGF system is known to be very complex and data are not always consistent, these findings and our results suggest that this system is involved in lung emphysema. It is also important to note that we observed the differences in fibroblast phenotype after several weeks in culture, indicating that these were neither transient nor dependent on the inflammatory environment in situ. It does not seem far-fetched to assume the persistence of alterations being at least partially due to epigenetic factors.

In performing the qPCR we additionally covered a number of genes of diverse pathways that could be altered in emphysema or cellular senescence. LOXL2 seemed of interest as involved in cross-linking collagens and elastin [24]; it has been found upregulated in fibroblasts in replicative as well as stress-induced premature senescence [30]. Overproduction of the ornithine decarboxylase (ODC) regulatory protein ODC-antizyme OAZ1 has been shown to correlate with cell growth inhibition in a variety of cell types [25]. This gene was included just because the downregulation in emphysema as indicated by the array would argue against our hypothesis. As a key member of cell cycle-associated factors, CDK4 was included, since there is evidence for a downregulation in senescent cells [35]. In addition, FOSL1 is known to be involved in proliferation and can be upregulated by cigarette smoke [23]. None of these genes turned out to be differentially regulated between emphysema and control patients according to qPCR. This does not render them irrelevant but puts additional emphasis on the findings regarding IGFBP-rP1 and -3, which showed reproducible and meaningful differences between groups. In addition, IGFBP-3 levels were elevated in supernatants of fibroblast from emphysema. These experiments were performed in the absence of fetal calf serum to avoid contributions from the serum. Although serum starvation itself could increase the amount of IGFBP-3 [15], the fact remains that this would have affected both groups. Due to the larger proliferation rate of control fibroblasts a higher total protein concentration was present in the supernatant. To reveal the relative production of IGFBP-3 we therefore normalized to total protein levels.

Due to the limited amount of cells available, it was not possible to perform all investigations in fibroblasts from the same group of patients. We ensured, however, that the groups compared were adequate in each case, by showing that they differed not only with respect to key patients' characteristics but also in fibroblast proliferation rates, as shown previously [1]. The use of different independent cultures, especially for gene expression analysis, thus involved true replicate culture, not just replicate analysis of the same RNA sample. This might well be the cause for the differences between the findings of the exploratory microarray analysis and the qPCR. On the other hand, the fact that IGFBP-3 and -rP1 were upregulated in both analyses and independent cultures, probably gives additional weight to this result.

It has been suggested, that replicative senescence of diploid cells in culture could be due to inadequate growth conditions [5]. Taking into account this, it could be argued, that our observations were at least partially the result of differences in the ability to handle oxidative stress in vitro. To resolve this issue, it would be helpful to detect senescence markers in fibroblasts of histological samples. Such analyses are, however, severely handicapped by the lack of fibroblast-specific antibodies. In addition, functional analyses are not possible in these cells without growing them in culture, and single-cell PCR requires amplification of mRNA which is an additional source of error. Thus we infer that, even if cell culture conditions should have been involved in our study, the present data provide evidence that a different phenotype of fibroblasts exists in lung emphysema. Such a different phenotype might well be present in other cells types, too, and is likely to involve epigenetic alterations. The presence of such persistent, programmed alterations might be of considerable importance for all attempts directed towards alveolar regeneration in patients with lung emphysema.

In conclusion, our data support the view that primary parenchymal lung fibroblasts from patients with emphysema show a senescent phenotype, which does not seem to be based on a reduction of telomere length. Instead, the upregulation of the senescence-associated IGFBP-3 and IGFBP-rP1 suggests that a change in the response to mitogenic and metabolic stimuli such as IGF-I, -II and insulin is involved in the previously found reduced proliferation rate in culture.