Paracrine Proliferative Signaling by Senescent Cells in WHO Group 3 Pulmonary Hypertension? Age Corrupting Youth

“For never-resting time leads summer on

To hideous winter and confounds him there”

- William Shakespeare, Sonnet 5

The term “Pulmonary hypertension” (PH) is a simple and arguably simplistic acknowledgement that the mean resting pulmonary artery pressure (PAP) is greater than 25 mmHg. It does not define pathophysiology, direct therapy or predict prognosis. More clinically useful is the World Health Organization (WHO) classification system, which acknowledges 5 PH Groups, each somewhat homogenous in terms of pathophysiology, lung histology and prognosis¹. Group 1 PH is a collection of syndromes characterized by marked pulmonary arterial obstruction, including patients with idiopathic and familial pulmonary arterial hypertension, connective tissue diseases, congenital heart diseases or hemoglobinopathies. Group 2 PH is associated with left heart disease (valvular and ventricular). Group 3 PH (cor pulmonale), relevant to this editorial, is associated with lung diseases, such as COPD, and with chronic hypoxia. Group 4 and 5 PH are associated with chronic thromboembolic disease or miscellaneous systemic conditions, such as sarcoidosis, respectively. PH-specific therapies (L-type calcium channel blockers, prostanoids, phosphodiesterase 5 inhibitors, and endothelin antagonists) are approved only for use in Group 1 PH. Since Group 2 and 3 PH are much more prevalent than Group 1 PH there is value in understanding their pathophysiology in hopes of developing rational therapies. Group 3 PH portends poor prognosis in COPD; however, it is uncertain whether agents that reverse this PH (which is usually modest in severity) improve prognosis^{2, 3}.

From animal models (rodents exposed to chronic hypoxia) and humans we have learned that Group 3 PH is due to a combination of pulmonary vasoconstriction and obstructive vascular remodeling, due to a shift in the balance of proliferation to apoptosis in smooth muscle cells (PASMC) and fibroblasts to favor proliferation. Histologically, medial thickness increases in resistance PAs due to distal extension of muscle and local transdifferentiation/migration of myofibroblasts or pericytes. Many abnormalities contribute to the proliferative phenotype in Group 3 PH, including increased levels of serotonin and the serotonin transporter, suppressed expression of bone morphogenetic protein receptor (BMPR2) and mitochondrial metabolic changes that inhibit pyruvate dehydrogenase and favor glycolysis (see review⁴). Metabolic and proliferative changes in the lung in Group 3 PH are significantly tied to activation of transcription factors, notably HIF-1α and NFAT (see review⁴). Mice that are haploinsufficient for HIF-1α are significantly protected from chronic hypoxic PH⁵. HIF-1α is a central feature of Group 3 PH and its activation results both from physiologic, hypoxia-induced changes in redox signaling, as well as pathologic mechanisms that may be shared with Group 1 PH, including epigenetic silencing of superoxide dismutase 2⁶ and activation of the endoplasmic reticulum stress protein, NOGO⁷. Reinforcing the proliferative diathesis in Group 3 PH, is an impairment of apoptosis that is manifest in downregulation of the oxygen-sensitive potassium channel, Kv1.5 and de novo expression of the antiapoptotic protein, survivin (see review⁴).

On this background emerges the report of Noureddine et al⁸. They offer an additional theory for how PASMC proliferation might be stimulated. Noureddine et al⁸ discuss findings from 124 patients with COPD who underwent right heart catheterization (RHC) and telomere length measurement (in circulating white blood cells). The majority of these patients have been previously reported by this group in publications that suggest that interleukin 6 (IL-6) levels are elevated and contribute to Group 3 PH and that telomeres are shortened in leukocytes from COPD patients, consistent with increased cellular aging in COPD^{9, 10}. They have previously concluded that leukocyte telomere shortening is correlated with the patient’s age as well as PaO₂ and PaCO₂¹⁰. What is new in the current report is the demonstration of senescence and a related paracrine proliferative diathesis in PASMC explanted from a new cohort of 14 COPD patients during lung resection for localized tumors. These COPD patients were compared to 13 patients with history of smoking but without COPD who underwent surgery. Some of the COPD patients had Group 3 PH, although the severity of the PH was mild. The pulmonary vasculature was assessed by quantitative histology. Immunohistochemical markers of proliferation (Ki67) and senescence (cyclin dependent kinase inhibitors p16 and p21) were used to localize senescent and proliferating cells in lung specimens. Cells cultures from the vasculature were examined to measure their predisposition for senescence and their ability to produce proliferative, paracrine factors, including interleukins (IL-6, IL-8) and tumor necrosis factor (TNF-α). Intriguingly they discovered an inverse relationship between telomere length and mean PAP and pulmonary vascular resistance (PVR) in COPD patients. Pulmonary vascular remodeling and the number of both senescent and proliferating cells were all greater in COPD patients versus the control cohort of smokers. Interestingly, and paradoxically, the senescent cells were almost exclusively confined to the media and were adjacent to areas of marked cell proliferation.

This leads to the question. “How can cells that don’t proliferate be associated with increased proliferation?” Noureddine et al demonstrate that culture media of PASMC from COPD patients (enriched in senescent cells, evident from reduced population doubling level, PDL) stimulate more proliferation than does media from PASMC derived from control smokers without COPD. The COPD-senescent cell cultures appeared to be causing proliferation by their exaggerated secretion of IL-6, IL-8, and TNF-α (a paracrine mechanisms) and by local contact mechanisms (that enhance PASMC migration). Thus Noureddine et al suggest that senescent PASMC from COPD patients cause otherwise healthy cells to proliferate excessively, which they speculate may explain the vascular remodeling and PH in this cohort (Figure). This is an intriguing idea, new to the PH field, although already being considered in the world of oncology¹¹. The authors’ findings are potentially noteworthy because they represent a first attempt to associate telomere shortening with Group 3 PH. However, some of the novelty is reduced by their two recent reports in similar patients and the fact that it is unclear how many of this cohort actually had PH (as defined by a mean PAP >25 mmHg and a PVR > 3 Wood Units).

Cell senescence was originally described in cells that had ceased proliferating and was felt to be a part of the natural aging process¹². In human cells, the arrest is triggered by telomere shortening, although the exact relationship between telomere shortening and senescence signaling is ill-defined¹³. In addition to replicative telomere shortening, cellular senescence can also be induced via reactive oxygen species or other stressors, which in turn upregulate cell cycle inhibitors such as p16 (p16INK4A) and p21(p21CIP1/WAF) ¹⁴. Senescent cells do not divide even after exposure to mitogenic factors and develop a phenotypic flat morphology with increased β-galactosidase activity. These senescent cells are not innocent by-standers, but rather have paracrine effects secreting factors with mitogenic, angiogenic, and antiapoptotic effects¹⁵. And it is in this way that Noureddine et al ⁸ in this issue of Circulation Research postulate that senescent cells fulfill the Shakespearean role of “never-resting time” leading the surrounding cells to an unhealthy, irreversible “hideous winter” of vascular remodeling and pulmonary hypertension (Figure 1). While this is a very novel approach to understanding the hyperproliferative state in PH, this interesting study has several caveats and limitations.

Figure 1.

Noureddine et al⁸ suggest that senescent PASMC from COPD patients cause otherwise healthy cells to proliferate excessively, which they speculate may explain the vascular remodeling and PH in this cohort.

Question 1: Is telomere shortening related to PH or to age and oxidant stress?

The patients in this study, like those previously presented in patients with pulmonary fibrosis¹⁶, are at risk of having Category 3 PH. However, the majority of these COPD patients had normal PA pressures or at worst mild PH. The mean PAP in the cohort that underwent RHC was technically normal (24.6 mmHg) and the average PVR was only slightly greater than normal (3.1 Wood Units). This raises the question of whether the telomeric shortening relates to PH or to the many other abnormalities in COPD. The authors acknowledge that telomere shortening in COPD and pulmonary fibrosis is not specific to the vasculature and in their prior study correlated with patient age and PO₂¹⁰. Moreover, in an elegant study of pulmonary fibrosis patients (which showed most had significant telomere shortening in leukocytes) there was no evidence of telomere shortening in the group 1 PH patients they used as controls¹⁶. Since the PH is much more severe in Group 1 vs Group 3 PH. this argues against the PH itself being directly related to telomere shortening. Perhaps the age or worsened oxygenation in COPD patients has the more direct relationship to telomere length with the mildly elevated PA pressures being a covariable. Unfortunately there was not a multivariate analysis showing that the relationship between telomere length and either PVR or proliferation was independent of age, PO₂, PCO₂ or quantified cigarette exposure. This is important because one of the strongest correlates of decreased telomerase even in the current study was age. The uncertainty about the directness of the link between proliferation and PH is compounded by the fact that a large part of the data (Table 1 from reference 8) derives from leukocytes (suggesting senescence is a systemic rather than a lung specific process).

Question 2: Is senescence specific to PASMC in the lung?

The authors find that telomere shortening occurs in circulating leukocytes of COPD patients as well as cultured PASMCs. Furthermore, they also report that there is increased staining for the stress-induced senescence marker p16 in PASMC and lung endothelium, again highlighting that the process is not unique to PASMCs. However, it is not clear how the senescence found in leukocytes and other cells impacts PASMC proliferation to promote PH. The authors present evidence that it is the senescent PASMC in the vascular wall that drive the remodeling process (Figure).

Question 3: Is there precedent for PASMC diversity in the vasculature?

Noureddine et al offer evidence that collections of senescent PASMC cause normal cells in the PA wall to proliferate by both paracrine and nonparacrine signaling⁸. This idea has 2 central tenants: First, that there is diversity in the types of PASMC that populate the vascular wall and second, that paracrine signaling can drive hyperplasia. There is precedent for the existence of both radial and longitudinal diversity of PASMC within the arterial wall. Not only are their diverse populations of PASMC in the vasculature, but this diversity relates to parameters relevant to Group 3 PH. There is a radial diversity of PASMC with populations of variable proliferative capacity as one proceeds from lumen to adventitia. Most of the proliferative response in PH occurs in meta-vinculin negative PASMC¹⁷. In addition, there is longitudinal diversity (in resistance versus conduit PAs) such that PASMC that mediate hypoxic pulmonary vasoconstriction are predominantly found in resistance PAs. This diversity reflects preferential expression of O₂-sensitive potassium channels and O₂-responsive mitochondria in PASMC from resistance versus conduit PAs^{18, 19}.

On the second count, there is precedent for senescent cells stimulating proliferation in bystander cells. In prostate cancer, senescent cells may stimulate hyperplasia of nearby normal epithelium¹¹. In contrast to this finding, enhancing cellular senescence has been exploited to regress several cancers, including lymphoma, osteosarcoma and hepatocellular carcinoma²⁰. Indeed there is a growing effort to counteract proliferation of cancer cells by enhancing oncogene-induced senescence, an intrinsic tumor suppressive mechanism²¹. Alimonti et al recently used a senescence response that is distinct from oncogene-induced senescence (PTEN-loss-induced cellular senescence), to regress prostate cancer in a human xenograft model²¹.

The proliferative effects of senescent cells SMC proliferation reported by Noureddine et al is also at odds with findings in the systemic vascular bed by Bennett et al. In a study of SMC in human atherosclerotic plaques they found earlier senescence and slower rates of cell proliferation in SMC from atherosclerotic plaque²². Thus, the relationship between telomerase, senescence and proliferation is relevant to important diseases (COPD, cancer and atherosclerosis) but the relationship is complex. More research is required before we will have confidence whether one should enhance or repress senescence in treating PH.

Question 4: How do the many previously described mechanisms of excessive PASMC proliferation relate to this new senescence mechanism?

In brief, the answer is, “we don’t know”. The field of PH research will be well served when there is more cross-testing of extant mechanisms because each is likely to be a part of the explanation, rather than the whole.

At the end of Sonnet 5, Shakespeare reassures the reader that age has not completely corrupted youth and there is some solace to be found in the beauty of one’s off-spring which triumphs over the efforts of Time such that “substance still lives sweet”. Noureddine et al⁸ describe increased presence of senescence cells in the pulmonary vasculature of patients with COPD with some pessimism, implicating them in the PA-SMC hypertrophy and elevated PA pressure. However, if we are to follow the cancer and atherosclerosis literature perhaps all is not lost and there may be benefit from enhancing cell senescence.

The Noureddine paper should stimulate further study of senescent cells in PH now that their presence in the vascular wall has been revealed. Ultimately we need to know whether we need to eliminate old cells before they corrupt the youth or whether one should exploit the fatigue of the senescent cell to therapeutic benefit.