Senescent cells (SCs) have been considered a source of age-related chronic sterile systemic inflammation and a target for anti-aging therapies. To understand mechanisms controlling the amount of SCs, we analyzed the phenomenon of rapid clearance of human senescent fibroblasts implanted into SCID mice, which can be overcome when SCs were embedded into alginate beads preventing them from immunocyte attack. To identify putative SC killers, we analyzed the content of cell populations in lavage and capsules formed around the SC-containing beads. One of the major cell types attracted by secretory factors of SCs was a subpopulation of macrophages characterized by p16(Ink4a) gene expression and β-galactosidase activity at pH6.0 (β-galpH6), thus resembling SCs. Consistently, mice with p16(Ink4a) promoter-driven luciferase, developed bright luminescence of their peritoneal cavity within two weeks following implantation of SCs embedded in alginate beads. p16(Ink4a)/β-galpH6-expressing cells had surface biomarkers of macrophages F4/80 and were sensitive to liposomal clodronate used for the selective killing of cells capable of phagocytosis. At the same time, clodronate failed to kill bona fide SCs generated in vitro by genotoxic stress. Old mice with elevated proportion of p16(Ink4a)/β-galpH6-positive cells in their tissues demonstrated reduction of both following systemic clodronate treatment, indicating that a significant proportion of cells previously considered to be SCs are actually a subclass of macrophages. These observations point at a significant role of p16(Ink4a)/β-galpH6-positive macrophages in aging, which previously was attributed solely to SCs. They require re-interpretation of the mechanisms underlying rejuvenating effects following eradication of p16(Ink4a)/β-galpH6-positive cells and reconsideration of potential cellular target for anti-aging treatment.

Constitutive p16Ink4a expression, along with senescence-associated β-galactosidase (SAβG), are commonly accepted biomarkers of senescent cells (SCs). Recent reports attributed improvement of the healthspan of aged mice following p16Ink4a-positive cell killing to the eradication of accumulated SCs. However, detection of p16Ink4a/SAβG-positive macrophages in the adipose tissue of old mice and in the peritoneal cavity of young animals following injection of alginate-encapsulated SCs has raised concerns about the exclusivity of these markers for SCs. Here we report that expression of p16Ink4a and SAβG in macrophages is acquired as part of a physiological response to immune stimuli rather than through senescence, consistent with reports that p16Ink4a plays a role in macrophage polarization and response. Unlike SCs, p16Ink4a/SAβG-positive macrophages can be induced in p53-null mice. Macrophages, but not mesenchymal SCs, lose both markers in response to M1- [LPS, IFN-α, Poly(I:C)] and increase their expression in response to M2-inducing stimuli (IL-4, IL-13). Moreover, interferon-inducing agent Poly(I:C) dramatically reduced p16Ink4a expression in vivo in our alginate bead model and in the adipose tissue of aged mice. These observations suggest that the antiaging effects following eradication of p16Ink4a-positive cells may not be solely attributed to SCs but also to non-senescent p16Ink4a/SAβG-positive macrophages.

During polyamine catabolism, spermine and spermidine are first acetylated by spermidine/spermine N(1)-acetyltransferase (SSAT) and subsequently oxidized by polyamine oxidase (PAO) to produce spermidine and putrescine, respectively. In attempting to clone the PAO involved in this back-conversion pathway, we encountered an oxidase that preferentially cleaves spermine in the absence of prior acetylation by SSAT. A BLAST search using maize PAO sequences identified homologous mammalian cDNAs derived from human hepatoma and mouse mammary carcinoma: the encoded proteins differed by 20 amino acids. When either cDNA was transiently transfected into HEK-293 cells, intracellular spermine pools decreased by 75% while spermidine and N (1)-acetylspermidine pools increased, suggesting that spermine was selectively and directly oxidized by the enzyme. Substrate specificity using lysates of oxidase-transfected HEK-293 cells revealed that the newly identified oxidase strongly favoured spermine over N (1)-acetylspermine and that it failed to act on N (1)-acetylspermidine, spermidine or the preferred PAO substrate, N (1), N (12)-diacetylspermine. The PAO inhibitor, MDL-72,527, only partially blocked oxidation of spermine while a previously reported PAO substrate, N (1)-( n -octanesulphonyl)spermine, potently inhibited the reaction. Overall, the data indicate that the enzyme represents a novel mammalian oxidase which, on the basis of substrate specificity, we have designated spermine oxidase in order to distinguish it from the PAO involved in polyamine back-conversion. The identification of an enzyme capable of directly oxidizing spermine to spermidine has important implications for understanding polyamine homoeostasis and for interpreting metabolic and cellular responses to clinically relevant polyamine analogues and inhibitors.

In the polyamine back-conversion pathway, spermine and spermidine are first acetylated by spermidine/spermine N1 -acetyltransferase (SSAT) and then oxidized by polyamine oxidase (PAO) to produce spermidine and putrescine respectively. Although PAO was first purified more than two decades ago, the protein has not yet been linked to genomic sequences. In the present study, we apply a BLAST search strategy to identify novel oxidase sequences located on human chromosome 10 and mouse chromosome 7. Homologous mammalian cDNAs derived from human brain and mouse mammary tumour were deduced to encode proteins of approx. 55 kDa having 82% sequence identity. When either cDNA was transiently transfected into HEK-293 cells, intracellular spermine pools decreased by approx. 30%, whereas spermidine increased 2-4-fold. Lysates of human PAO cDNA-transfected HEK-293 cells, but not vector-transfected cells, rapidly oxidized N1-acetylspermine to spermidine. Substrate specificity determinations with the lysate assay revealed a preference ranking of N1-acetylspermine= N1-acetylspermidine> N1,N12-diacetylspermine>>spermine; spermidine was not acted upon. This ranking is identical to that reported for purified PAO and distinctly different from the recently identified spermine oxidase (SMO), which prefers spermine over N1-acetylspermine. Monoethyl- and diethylspermine analogues also served as substrates for PAO, and were internally cleaved adjacent to a secondary amine. We deduce that the present oxidase sequences are those of the FAD-dependent PAO involved in the polyamine back-conversion pathway. In Northern blot analysis, PAO mRNA was much less abundant in HEK-293 cells than SMO or SSAT mRNA, and all three were differentially induced in a similar manner by selected polyamine analogues. The identification of PAO sequences, together with the recently identified SMO sequences, provides new opportunities for understanding the dynamics of polyamine homoeostasis and for interpreting metabolic and cellular responses to clinically-relevant polyamine analogues and inhibitors.