Senescence Induces Proteolytically-active PAPP-A Secretion and Association with Extracellular Vesicles in Human Pre-adipocytes

Abstract

Senescence is a cellular response to various stressors characterized by irreversible cell cycle arrest, resistance to apoptosis and expression of a senescence-associated secretory phenotype (SASP). Interestingly, studies where senescent cells were deleted in mice produced beneficial effects similar to those where the zinc metalloproteinase, PAPP-A, was deleted in mice. In this study, we investigated the effect of senescence on PAPP-A secretion and activity in primary cultures of adult human pre-adipocytes. Cultured pre-adipocytes were isolated from subcutaneous (Sub) and omental (Om) fat. Senescence was induced with low dose etoposide. PAPP-A protein was measured by an ultrasensitive PAPP-A ELISA. PAPP-A proteolytic activity was measured by a specific substrate cleavage assay. Senescence significantly increased PAPP-A levels in both Sub and Om conditioned medium (CM) 8- to 15-fold over non-senescent CM. Proteolytic activity reflected PAPP-A protein with 12- to 18-fold greater activity in senescent CM versus non-senescent CM. Furthermore, PAPP-A was found at high levels on the surface of extracellular vesicles secreted by senescent pre-adipocytes and was proteolytically active. In conclusion, we identified enzymatically active PAPP-A as a component of human pre-adipocyte SASP. This recognition warrants further investigation of PAPP-A as a new biomarker for senescence and a potential therapeutic target to control the of spread of senescence in adipose tissue.

Graphical Abstract

INTRODUCTION

Senescence is a complex response of a cell to various stressors that is characterized by irreversible cell cycle arrest, resistance to apoptosis, and expression of a senescence-associated secretory phenotype (SASP) comprised of cytokines, chemokines, proteases, inhibitors, and growth factors^1–3. Identification and understanding of the different factors present in various SASPs and their function is incomplete. SASP can vary depending on the inducer of senescence, duration in senescence, and cell type. Importantly, the effects of SASP can spread to nearby non-senescent cells⁴. This is a paracrine mechanism by which senescent cells can impact tissue and organ function despite initial small numbers. Interestingly, the secretion of extracellular vesicles/exosomes by senescent cells has been reported recently and may be involved in far-reaching communication^{5, 6}. These senescent-associated extracellular vesicles (EVs) appear to be novel SASP factors⁷.

Cellular senescence can be beneficial in preventing proliferation of pre-neoplastic cells, regulating tissue remodeling during embryogenesis, and promoting optimal wound healing. However, accumulation of senescent cells with age appears to have adverse consequences on tissue and organ function³. Indeed, SASP has been postulated to make a significant contribution to age-related inflammation and associated pathologies^{8, 9}. Genetic and pharmacological approaches have been used to selectively target and clear senescent cells in various mouse models of aging and age-related disease¹⁰. Removal of senescent cells resulted in decreased atherosclerotic plaque progression¹¹, prevention of pulmonary fibrosis¹², alleviation of obesity-related metabolic dysfunction^{13, 14}, reduced liver lipid^{13, 14}, delay in thymic involution^{15, 16}, decreased indices of frailty¹⁷, and enhanced insulin sensitivity¹³. Baker et al.¹⁸ demonstrated that late-life clearance of senescent cells in wild-type mice could attenuate progress of several established age-related disorders and extend median lifespan. Several papers have identified senescent cells associated with aging and age-related pathologies in human tissues, as well¹⁹. Although senolytic drugs that target cell death-resistant pathways show great promise^20–22, a better understanding of cell senescence, and especially SASP, will be critical for guiding successful clinical trials.

Increased insulin-like growth factor (IGF) signaling has been associated with aging and many age-related diseases. Indeed, reduction of IGF-I signaling has been shown to prolong lifespan in diverse species^{23, 24}. However, very little is known about the IGF system in senescence. It has been shown that acute treatment of cells with IGF-I stimulates proliferation, while prolonged treatment of cells with IGF-I promotes senescence in a variety of cell types^25–27.

It seemed more than coincidental that studies where senescent cells were deleted in mice produced a beneficial phenotype similar to those where PAPP-A was deleted in mice (Table 1). PAPP-A is a novel zinc metalloprotease that can increase pericellular IGF bioavailability through cleavage of inhibitory IGFBPs, in particular IGFBP-4²⁸. PAPP-A is a secreted protein that tethers to the surface of cells through heparan sulfate-like proteoglycan moieties in an autocrine/paracrine fashion. IGF bound to IGFBP-4 is unable to activate receptors. However, upon cleavage of IGFBP-4 by PAPP-A, IGF is liberated from the complex in the pericellular environment and IGF signaling is initiated. PAPP-A-induced enhancement of local IGF action through proteolysis of IGFBP-4 has been demonstrated in several in vitro systems and appears to serve a similar function in vivo²⁸. Conversely, inhibition of PAPP-A expression or its proteolytic activity represents an innovative approach to decrease IGF availability with resultant restraint of IGF-I receptor signaling. Thus, inhibition of PAPP-A through gene deletion in mice had many beneficial effects, including a remarkable extension of lifespan by ~30%²⁹. A significant increase in lifespan was also seen with PAPP-A knock-out (KO) mice on a high fat diet²⁹, and when PAPP-A gene expression was knocked-down in adult mice³⁰. PAPP-A KO mice, like senescent cell-targeted mice, showed resistance to atherosclerotic plaque progression^{31, 322}, resistance to visceral obesity with high fat diet^{33, 34}, reduced liver lipid^{33, 34}, delay of thymic involution with persistence of a youthful T-cell phenotype in aged mice³⁵, prevention of sarcopenia with age³⁶, and enhanced insulin sensitivity³⁷. We can also inhibit the ability of PAPP-A to cleave IGFBP-4 with a novel neutralizing monoclonal antibody generated against a unique exosite in PAPP-A (mAb-PA1/41)^{38, 39}. Weekly treatment of apolipoprotein E KO mice (wild-type alleles for PAPP-A) with this monoclonal antibody reduced atherosclerotic plaque burden by 70–80%³¹. Visceral obesity in mice on high fat diet was also prevented with genetic or pharmacological inhibition of PAPP-A³⁴.

Table 1.

PAPP-A and Cell Senescence

Effects on aging and age-related disorders	Delete Senescent Cells	Delete PAPP-A
Increase lifespan	Yes10,18	Yes28–30
Impede atherosclerosis	Yes11	Yes31,32
Prevent visceral obesity	Yes13,14	Yes33,34
Reduced liver lipid	Yes13,14	Yes33,34
Delay Thymic involution	Yes15,16	Yes35
Prevent sarcopenia	Yes17	Yes36
Enhance insulin sensitivity	Yes13	Yes37

For this study, we posed specific questions about PAPP-A and senescence using primary cultures of adult human pre-adipocytes. Pre-adipocytes are also referred to as adipocyte-derived stem cells or fat progenitor cells. Adipose tissue is the largest source of progenitor cells^{2, 10}, and human adipose tissue exhibits a very high accumulation of senescent cells with age¹⁹.

MATERIALS and METHODS

Materials.

Cell culture components were from Gibco Life Technologies. Bovine serum albumin (BSA) was from MilliporeSigma. Recombinant human IGF-II was from R&D Systems and IGFBP-4 was a gift from our colleague, Professor Claus Oxvig (Aarhus University, Denmark). Anti-IGFBP-4 antibodies were purchased from Abcam: N-terminal ab92625 and C-terminal ab77350. Secondary antibodies were from LiCor, Inc: Donkey anti-goat (926–68074) and donkey anti-rabbit (926–32213). Etoposide was obtained from Enzo Life Sciences. picoPAPP-A ELISA kits and monoclonal antibody generated against a unique exosite in PAPP-A (mAb-PA1/41)^{38, 39} were obtained from Ansh Labs (Webster, TX).

Cells.

Primary human pre-adipocytes isolated from subcutaneous (Sub) and omental (Om) fat from the same subject were plated and cultured, as previously described⁴⁰. These are essentially pure populations of preadipocytes⁴¹. All experiments used isolations from three different donors; age and sex are presented in Supplemental Table 1. Cells were grown in αMEM supplemented with penicillin (100 IU/ml), streptomycin (100 μg/ml), and 10% heat-inactivated fetal bovine serum. After washing in phosphate-buffered saline (PBS), experiments were performed in serum-free DMEM containing 0.1% BSA. Cells were used at passages 4–9.

Low dose etoposide was used to induce DNA damage as a genotoxic stress leading to cell senescence, as previously described⁴². For induction of senescence, human pre-adipocytes at 60% confluency were treated for 48 h with 25 μM etoposide in growth media. For non-senescent cells, vehicle (0.025% DMSO) was added to the growth media. Etoposide and DMSO were removed and replaced with fresh media every three days. Senescent and non-senescent cells were changed to serum-free media with experimental additions for 72 h and conditioned medium (CM) was collected, aliquoted and stored frozen at −80°C.

Senescence-associated β-galactosidase staining and cell number.

SA-β-gal staining was performed as previously described⁴². DAPI (Thermo Scientific 62248) was used at a 1:1000 dilution for cell number. Cytation 5 Imaging Reader (Bio Tek) was used for imaging and quantification of SA-β-gal and DAPI.

RT-qPCR.

Total RNA was isolated from cultured cells by washing twice with cold phosphate buffered saline (PBS), lysing with 1 ml of Trizol (Ambion Life Technologies, Carlsbad, CA) and further processed as per manufacturer’s instruction. RNA (1 μg) was reversed transcribed with the SuperScript III Frist-Strand Synthesis System (Life Technologies) and evaluated by quantitative real-time PCR using the CFX Connect Real-Time System with iTAQ Universal SYBR Green Supermix (Bio-Rad, Hercules, CA). Amplification plots were analyzed using CFX Maestro Software version 4.1 (Bio-Rad). [p21 primers for RT-qPCR were forward: TGTCCGTCAGAACCCATGC; reverse: AAAGTCGAAGTTCCATCGCTC]. To create standard curves, designed primers were PCR’d with cDNA from known cell culture samples. Amplified PCR products were purified through QIAquick Gel Extraction Kit (Qiagen Hilden, Germany), quantified and serial diluted from 10⁸ to 10³ molecules. Relative quantification and fold changes were based on the standard curve for each gene.

PAPP-A protein and enzymatic activity.

Ultrasensitive ELISA kits that measure human PAPP-A were used according to manufacturer’s instructions to determine levels of PAPP-A in CM. For proteolysis assays, human IGFBP-4 (4 nM) pre-complexed to IGF-II (2 nM) was incubated in cell-free CM without and with inhibitory PAPP-A monoclonal antibody (mAb-PA1/41, 10⁻⁷ M) for 2 hours at 37°C, as previously described⁴³. IGF must be bound to IGFBP-4 for IGFBP-4 to be a substrate for PAPP-A proteolysis, and IGF-II is more effective than IGF-I in this cell-free assay^44–46. mAb-PA1/41 specifically inhibits PAPP-A-mediated IGFBP-4 proteolysis^{38, 39}. CM was then incubated for the indicated times at 37°C. Western blotting was performed as previously described⁴³. Briefly, samples were separated by 12% SDS-PAGE under reducing conditions, blotted onto a PVDF membrane, blocked, and probed for IGFBP-4 fragments using specific N- and C-terminal antibodies. Fluorescently-labeled secondary antibodies were used for detection of intact and cleaved IGFBP-4 by LI-COR Odyssey imager and Image Studio software.

Extracellular vesicle isolation.

The isolation procedure for EVs was taken from Barile et al. using serial ultracentrifugation⁴⁷. Freshly collected CM from T75 flasks was centrifuged 3,000 × g for 15 minutes at 4°C to remove any cells. The supernatant was moved to a new tube and centrifuged at 10,000 × g for 15 minutes at 4°C to remove cellular debris. This supernatant was collected into a 10 ml conical bottom ultracentrifuge tube (Beckman Coulter) and centrifuged 100,000 × g for 90 minutes in an L-70 ultracentrifuge (Beckman) using an SW 41 swinging bucket. The supernatant was discarded and the EV-enriched pellet reconstituted in PBS or solubilized in RIPA buffer (0.15 M NaCl, 0.5% NP-40, 0.1% SDS, 50 mM Tris). Total protein concentrations were determined by using Pierce BCA Protein Assay Kit (ThermoFisher) according to manufacturer’s instructions.

Adipokines.

Pre-adipocyte CM were analyzed using a Luminex Discovery Panel from R&D Systems by the Immunochemical Core Laboratory of Mayo Clinic.

Statistical analyses.

Comparisons of Sub and Om pre-adipocytes from each patient were analyzed by paired t-tests. Comparisons of non-senescent vs. senescent cells were analyzed by non-paired t-test. Significance was set at P < 0.05

RESULTS

Are subcutaneous- and visceral fat-derived pre-adipocytes differentially responsive to etoposide-induced senescence?

To address this question, we assessed etoposide-induced senescence by observing characteristic physical changes in the cells, i.e., enlarged and flatten cell morphology, and SA-β-gal staining of (Sub) and omental (Om) pre-adipocyte pairs. There was a marked 50–65% decrease in cell number in etoposide-treated compared to non-treated cells, with apparent cessation of proliferation (Fig.1A), at which times 90–100% of cells were positive for SA-β-gal staining. Representative SA-β-gal staining is shown in Figure 1B.

Figure 1.

Effect of etoposide-induced cell senescence (Sen) on (A) cell number and (B) senescence-associated B-galactosidase staining in Subcutaneous (Sub) and Omental (Om) pre-adipocytes

The p21 gene is a marker of cell cycle arrest and its upregulation has been shown to maintain viability of DNA damage-induced senescent cells⁴⁸. There was a significant elevation (~2-fold) of p21 expression in both Sub and Om pre-adipocytes in response to etoposide-induced senescence (Table 2).

Table 2.

p21 mRNA Expression (x10⁴)

	Non-Senescent	Senescent	P-value
Sub	3.5 ± 0.14	5.7 ± 0.21	0.001
Om	4.1 ± 0.21	5.6 ± 0.23	0.008

Expression of p21, a marker of cell cycle inhibition, was measured by RT-qPCR.

Results are mean ± SEM, n = 3 subjects.

Levels of another gene marker of cell cycle arrest, p16, did not change in these cells with induction of senescence (data not shown). p21 and p16 mediate different pathways and at different phases of senescence and are not always up-regulated in parallel^{49, 50}. Thus, morphological changes, increased SA-β-gal staining, decreased cell proliferation, up-regulation of p21, and induction of common SASP factors, IL-6 and MCP-1 (see Table 6), are consistent with a senescent phenotype in both Sub and Om pre-adipocytes.

Table 6.

Cytokine/SASP factor expression in CM from Non-Senescent and Senescent pre-adipocytes Seventy-two hour CM from pre-adipocyte isolates from subcutaneous and omental fat depots. Results are mean ± SEM of three subjects and expressed as pg/10⁴ cells. All samples (in duplicate) were run together in the same assay. IL (Interleukin)-6, IL-8, MCP (Monocyte Chemoattractant Protein)-1, HGF (Hepatocyte Growth Factor), PAI-1 (Plasminogen Activator Inhibitor)-1 IL-1β and TNF (Tumor Necrosis Factor)-α were below the limits of quantitation in the Luminex Discovery Panel.

pg/104 Cells
	Non-Senescent	Senescent	P-value
IL-6
Sub	235 ± 97	1,695 ± 136	0.0009
Om	589 ± 304	1,972 ± 152	0.0152
IL-8
Sub	115 ± 103	117 ± 19	0.986
Om	109 ± 93	176 ± 48	0.557
MCP-1
Sub	75 ± 56	5,647 ± 229	<0.0001
Om	143 ± 83	2,507 ± 581	0.0158
HGF
Sub	49 ± 44	60 ± 35	0.854
Om	15 ± 6	104 ± 73	0.291
PAI-1
Sub	17,430 ± 7154	163,486 ± 111,421	0.261
Om	33,200 ± 10,813	46,020 ± 23,718	0.649

Does senescence affect PAPP-A expression and proteolytic activity?

Senescent and non-senescent Sub and Om cells were washed and changed to serum-free medium for 72 hr. Results (expressed as ng PAPP-A protein/10⁴ cells) are presented in Table 3.

Table 3.

PAPP-A Expression in CM from Non-Senescent and Senescent pre-adipocytes.

PAPP-A (ng/104 Cells)
	Sub	Om
Non-Senescent
Subject A	0.7	2.0
Subject B	1.1	6.2
Subject C	0.7	2.1
Senescent
Subject A	8.8	19.4
Subject B	16.2	55.1
Subject C	5.9	23.8

Human PAPP-A protein levels in CM were measured by picoPAPP-A ELISA and expressed as ng/10⁴ cells. All samples (in duplicate) were run together in the same assay.

In non-senescent cells, PAPP-A levels in Om CM were 3- to-6 fold higher than in Sub CM. This is similar to what we have previously reported⁴⁰, but did not reach statistical significance in this study. Senescence increased PAPP-A levels 8- to 15-fold over non-senescent cells in both Sub and Om CM, with absolute levels of PAPP-A being 2- to 4-fold higher in Om CM than in Sub CM. The effect of senescence was significant for Sub CM (P = 0.004) and Om CM (P = 0.011), Figure 2.

Figure 2.

Effect of senescence on PAPP-A protein secretion in pre-adipocytes isolated from subcutaneous (Sub) and omental (Om) fat depots.

Results are mean ± SEM, n=3

*P < 0.05 Senescence (hatched bars) vs. non-senescence (solid bars)

Proteolytic activity in CM reflected PAPP-A protein levels. An example is presented in Figure 3, but the results were similar in all three subjects (Table 4 and full Westerns in supplemental Figures 1–3).

Figure 3.

PAPP-A-mediated IGF-dependent IGFBP-4 proteolysis in CM from non-senescent and senescent subcutaneous (Sub) and omental (Om) pre-adipocytes. Arrows indicate intact IGFBP-4 and 18-kD N-terminal fragment of IGFBP-4. Molecular weight standards are on the left.

Table 4.

PAPP-A-mediated IGF-Independent IGFBP-4 proteolysis in CM from Non-Senescent and Senescent pre-adipocytes.

18 kD IGFBP-4 fragment (densitometry units)
	Sub	Om
Non-Senescent
Subject A	0.18	0.25
Subject B	0.06	0.16
Subject C	0.10	0.19
Senescent
Subject A	0.80	1.12
Subject B	1.10	0.86
Subject C	0.83	0.96

PAPP-A-mediated IGFBP-4 proteolysis in CM was measured as described in Materials and Methods.

In CM from non-senescent cells, proteolytic fragments of IGFBP-4 were only seen in the presence of exogenous IGF-II, indicating the IGF-dependent nature of PAPP-A-mediated proteolysis of IGFBP-4^44–46. This activity was inhibited by mAb-PA1/41 indicating specificity for PAPP-A^{38, 39}. As measured by the fluorescent intensity of the 18 kD IGFBP-4 fragment, proteolytic activity in Om CM was 2- to 3-fold greater than in Sub CM from non-senescent cells. CM from senescent cells had markedly increased proteolytic activity that was 12- to 28-fold over CM from non-senescent cells (Fig. 4). With the 6h incubation time, there was no apparent difference in proteolytic activity between Sub and Om CM from senescent cells. There were some IGFBP-4 fragments in the absence of added IGF-II (Control lanes) in the senescent cell CM of all three subjects. This could reflect endogenous IGF secretion by senescent cells⁵¹ or, more likely, the very high levels of PAPP-A in CM from senescent cells.

Figure 4.

Effect of senescence on PAPP-A-mediated IGFBP-4 proteolysis in CM from pre-adipocytes isolated from subcutaneous (Sub) and omental (Om) fat depots, as measured by IGFBP-4 fragments (in densitometry units).

Results are mean ± SEM, n=3

***P < 0.001 Senescence (hatched bars) vs. non-senescence (solid bars)

Is PAPP-A found on extracellular vesicles (EVs) released by senescent pre-adipocytes?

Subcutaneous and omental pre-adipocytes were treated and cultured as above. CM from senescent and control cells were collected and processed for EVs; the isolation procedure is taken from Barile et al. using sequential ultra-centrifugation⁴⁷. EVs were assayed for PAPP-A protein and proteolytic activity, as above. EV PAPP-A was expressed as ng/mg protein (Table 5). The difference in Sub vs Om in non-senescent cells was approximately 2-fold (Om > Sub), but there was considerable variability among the three subjects. Likewise, senescence increased EV-PAPP-A but with variable results (2- to 7-fold). Proteolytic activity in senescent EVs is presented in Figure 5 (full Western in supplemental Figure 4).

Table 5.

PAPP-A expression in EVs secreted by Non-Senescent and Senescent pre-adipocytes.

PAPP-A (ng/mg protein)
	Sub	Om
Non-Senescent
Subject A	23	98
Subject B	25	31
Subject C	35	61
Senescent
Subject A	59	237
Subject B	175	226
Subject C	44	477

Human PAPP-A protein levels in solubilized EVs were measured by ELISA and expressed as ng/mg protein. All samples (in duplicate) were run together in the same assay.

Figure 5.

PAPP-A-mediated IGF-dependent IGFBP-4 proteolysis in solubilized and intact extracellular vesicles (EVs) from senescent pre-adipocytes. Arrows indicate intact IGFBP-4 and 18-kD N-terminal fragment of IGFBP-4. Molecular weight standards are on the left.

We compared EVs solubilized in RIPA buffer (total) to those brought up in PBS (intact). We found marked IGF-dependent IGFBP-4 proteolytic activity (measured as decrease in intact IGFBP-4 fluorescence) in both RIPA (11% of control) and PBS (8% of control) preparations which was completely inhibited by mAb-PA1/41. The difference in the protease assay for CM and EV protein likely reflects the difference in incubation times, i.e., 6 hr. for CM and 24 hr. for EVs. Nevertheless, we demonstrated that PAPP-A was found on the surface of EVs from senescent preadipocytes and was proteolytically active.

Is PAPP-A expression induced in senescence driven by increases in pro-inflammatory cytokines?

Interleukin (IL)-1β and tumor necrosis factor (TNF)-α are potent stimulators of PAPP-A expression in human pre-adipocytes, with response of Om cells > Sub cells⁴⁰. However, these pro-inflammatory cytokines were below the level of detection in 72 hr. CM (Table 6). Common SASP factors, IL-6 and Monocyte Chemoattractant protein (MCP)-1 were significantly elevated in CM from senescent Sub and Om cells. Other recognized SASP factors IL-8, Hepatocyte Growth Factor (HGF) and Plasminogen Activator Inhibitor (PAI)-1 were detectable in the Sub and Om CM but not significantly changed with induced senescence.

DISCUSSION

It is known that visceral pre-adipocytes are relatively resistant to differentiation⁵², which affects function, but it is not known whether this also applies to senescence. Therefore, we first asked ‘Are subcutaneous- and visceral fat-derived pre-adipocytes differentially responsive to etoposide-induced senescence?’. The answer was no. We found very similar induction and extent of SA-β-gal staining, elevation in p21 expression, and induction of common SASP factors, IL-6 and MCP-1, in Sub and Om pre-adipocytes. To our knowledge, this is the first study to directly compare senescence induction and SASP in primary human subcutaneous and omental preadipocytes. There are studies looking at senescence in subcutaneous pre-adipocytes and one which showed that omental pre-adipocytes are capable of replicative senescence^{17, 53}, but there are no paired studies in Sub and Om cells from the same subject and cultured under identical conditions. We acknowledge that there may be differences in Sub and Om with other inducers of senescence.

The answer to our next question ‘Does senescence affect PAPP-A expression and proteolytic activity?’ was yes. There were highly significant effects of senescence on PAPP-A secretion and activity in both Sub and Om CM. PAPP-A protein was increased 8- to 15-fold in Sub CM and 9- to 11-fold in Om CM from senescent cells compared to non-senescent counterparts. Overall PAPP-A levels in Om CM were ~ 3-fold higher than in Sub CM from senescent cells. Large increases in IGF-dependent proteolytic activity against IGFBP-4 in CM were also seen following induced senescence. Specificity of PAPP-A-mediated IGFBP-4 proteolysis was confirmed with mAb-PA1/41, which inhibits this activity^{38, 39}. Thus, active PAPP-A that can enhance local IGF bioavailability and inhibit apoptosis is part of soluble SASP in pre-adipocytes. Resistance to apoptosis is a hallmark of cellular senescence and can promote accumulation of senescent cells with deleterious effects on the tissue^1–3.

The answer to our third question ‘Is PAPP-A found on extracellular vesicles (EVs) released by senescent preadipocytes?’ was also a yes. We found that senescent pre-adipocytes secreted EVs with active PAPP-A on their membrane surface, which could impact non-senescent cells. Little is known about active components on the surface of EVs and their effect on signaling. EVs released by senescent human pre-adipocytes could be a novel mechanism by which PAPP-A contributes to spread of premature senescence in adipose tissue, likely through enhanced paracrine IGF signaling.

We also asked whether PAPP-A expression could be regulated by pro-inflammatory cytokines co-secreted in the SASP. IL-1β and TNF-α are known potent stimulators of PAPP-A expression in a variety of cells, including pre-adipocytes^{40, 54}. However, neither IL-1β nor TNF-α was detected in pre-adipocyte CM even upon induction of senescence. IL-6 is generally the most abundant pro-inflammatory cytokine in various SASPs, including our Sub and Om pre-adipocyte SASP. However, we have shown that IL-6 does not stimulate PAPP-A expression in human pre-adipocytes⁴⁰. The other analyte that was significantly increased with senescence in these cells was MCP-1, a regulator of migration and infiltration of immune cells. It was found to be elevated during replicative senescence of mesenchymal stromal cells⁵⁵. There is no known association of MCP-1 with PAPP-A expression. IL-8 and HGF have been considered markers of senescence in different cell types^{56, 57}, but levels were low in Sub and Om pre-adipocyte CM with no significant elevation with senescence. PAI-1 levels in CM were high and showed considerable variability. Its relationship to the senescent phenotype is unclear as it is multifunctional protein⁵⁸. Thus, increased PAPP-A expression with senescence does not appear to be secondary to other factors in Sub and Om SASP and, therefore, may be an independent biomarker of senescence at least in adipose tissue.

We acknowledge that senescence is an interactive process involving several cell types, various growth factors and cytokines, crosstalk and feedback loops that cannot be fully appreciated in cell cultures studies. Nevertheless, the experiments were carefully designed and controlled to answer the specific questions that relate to a proposed new pathologic mechanism underlying cellular senescence. Answers to these questions are important for recognizing PAPP-A as a novel SASP component that could contribute to the spread of cellular senescence and for providing rationale for potential PAPP-A-targeted therapies in aging and age-related disorders where senescence is known to play a role.

HIGHLIGHTS.

• Senescence was induced in primary cultures of adult human pre-adipocytes

• Senescence significantly increased PAPP-A secretion and proteolytic activity

• Senescence increased active PAPP-A on the surface of extracellular vesicles (EV)

• Soluble and EV-PAPP-A are novel components of SASP in human pre-adipocytes

ABBREVIATIONS

PAPP-A

pregnancy-associated plasma protein-A

Sub

subcutaneous

Om

omental

CM

conditioned medium

SASP

senescence-associated secretory phenotype

EV

extracellular vehicle

IGF

insulin-like growth factor

IGFBP-4

insulin-like growth factor binding protein-4