Changes in the expression of splicing factor transcripts and variations in alternative splicing are associated with lifespan in mice and humans

Abstract

Dysregulation of splicing factor expression and altered alternative splicing are associated with aging in humans and other species, and also with replicative senescence in cultured cells. Here, we assess whether expression changes of key splicing regulator genes and consequent effects on alternative splicing are also associated with strain longevity in old and young mice, across 6 different mouse strains with varying lifespan (A/J, NOD.B10Sn-H2(b) /J, PWD.Phj, 129S1/SvlmJ, C57BL/6J and WSB/EiJ). Splicing factor expression and changes to alternative splicing were associated with strain lifespan in spleen and to a lesser extent in muscle. These changes mainly involved hnRNP splicing inhibitor transcripts with most changes more marked in spleens of young animals from long-lived strains. Changes in spleen isoform expression were suggestive of reduced cellular senescence and retained cellular proliferative capacity in long-lived strains. Changes in muscle isoform expression were consistent with reduced pro-inflammatory signalling in longer-lived strains. Two splicing regulators, HNRNPA1 and HNRNPA2B1, were also associated with parental longevity in humans, in the InCHIANTI aging study. Splicing factors may represent a driver, mediator or early marker of lifespan in mouse, as expression differences were present in the young animals of long-lived strains. Changes to alternative splicing patterns of key senescence genes in spleen and key remodelling genes in muscle suggest that correct regulation of alternative splicing may enhance lifespan in mice. Expression of some splicing factors in humans was also associated with parental longevity, suggesting that splicing regulation may also influence lifespan in humans.

Keywords: isoforms; lifespan; longevity; mRNA splicing; mouse; splicing factors.

Figure 1

Schematic of study design. This figure shows the experimental strategy employed to assess the effects of strain longevity and mouse age on the expression of an a priori panel of splicing factors (SFs).

Figure 2

Splicing factor expression according to mouse lifespan. This plot illustrates association between median lifespan and slicing factor expression in total RNA from spleen (A) or muscle (B) tissues in mice of 6 strains of different longevities, as assessed by linear regression against median strain lifespan. The identity of specific splicing factors is given on the x‐axis. The log10 P‐values for associations between lifespan and splicing factor expression from mice of strains with different lifespans and of different ages are given on the y‐axis. Direction of effect is also indicated; data appearing above the zero line on the y‐axis represent positive associations, whilst data appearing below the zero line represent negative associations. Analysis including all animals in the sample is given by dark grey bars, in young animals only by medium grey bars and in old animals only by light grey bars. The dotted line refers to a P‐value cut‐off for statistical significance of P = 0.05.

Figure 3

The expression of alternative isoforms of key genes according to mouse lifespan. This plot illustrates association between median lifespan and the expression of alternatively expressed isoforms of key genes in total RNA from spleen (A) or muscle (B) tissues in mice of 6 strains of different longevities as assessed by linear regression against median strain lifespan. The identity of specific splicing factors is given on the x‐axis. The log10 P‐values for associations between lifespan and splicing factor expression from mice of strains with different lifespans and of different ages are given on the y‐axis. Direction of effect is also indicated; data appearing above the zero line on the y‐axis represent positive associations, whilst data appearing below the zero line represent negative associations. Analysis including all animals in the sample is given by dark grey bars, in young animals only by medium grey bars and in old animals only by light grey bars. The dotted line refers to a P‐value cut‐off for statistical significance of P = 0.05.

Figure 4

Splicing factor expression according to mouse age. This plot illustrates association between age and splicing factor expression in total RNA from spleen (A) or muscle (B) tissues in young (6 months) vs. old (20–22 months) mice. The identity of specific splicing factors is given on the x‐axis. The log10 P‐values for associations between age and splicing factor expression from mice of different ages and of different strains are given on the y‐axis. Direction of effect is also indicated; data appearing above the zero line on the y‐axis represent positive associations, whilst data appearing below the zero line represent negative associations. Analysis including all animals in the sample is given by dark grey bars, in animals of average‐lived strains only by medium grey bars and in long‐lived animals only by light grey bars. The dotted line refers to a P‐value cut‐off for statistical significance of P = 0.05.

Figure 5

The expression of alternative isoforms of key genes according to mouse age. This plot illustrates association between the expression of age and the expression of alternatively expressed isoforms of key genes in total RNA from spleen (A) or muscle (B) tissues in young (6 months) vs. old (20–22 months) mice. The identity of isoforms is given on the x‐axis. The log10 P‐values for associations between age and the expression of alternative isoforms from mice of different ages and of different strains are given on the y‐axis. Direction of effect is also indicated; data appearing above the zero line on the y‐axis represent positive associations, whilst data appearing below the zero line represent negative associations. Analysis including all animals in the sample is given by dark grey bars, in animals of average‐lived strains only by medium grey bars and in long‐lived animals only by light grey bars. The dotted line refers to a P‐value cut‐off for statistical significance of P = 0.05.