1. Cell Biology

Age-dependent changes in protein incorporation into collagen-rich tissues of mice by in vivo pulsed SILAC labelling

  1. Yoanna Ariosa-Morejon
  2. Alberto Santos
  3. Roman Fischer
  4. Simon Davis
  5. Philip Charles
  6. Rajesh Thakker
  7. Angus KT Wann
  8. Tonia L Vincent  Is a corresponding author
  1. Kennedy Institute of Rheumatology, Arthritis Research UK Centre for OA Pathogenesis, University of Oxford, United Kingdom
  2. Big Data Institute, Li-Ka Shing Centre for Health Information and Discovery, University of Oxford, United Kingdom
  3. Center for Health Data Science, Faculty of Health Sciences, University of Copenhagen, United Kingdom
  4. Nuffield Department of Medicine, Target Discovery Institute, University of Oxford, United Kingdom
  5. Academic Endocrine Unit, OCDEM, Churchill Hospital, University of Oxford, United Kingdom
https://doi.org/10.7554/eLife.66635
Share this article
Cite this article
  1. Yoanna Ariosa-Morejon
  2. Alberto Santos
  3. Roman Fischer
  4. Simon Davis
  5. Philip Charles
  6. Rajesh Thakker
  7. Angus KT Wann
  8. Tonia L Vincent
(2021)
Age-dependent changes in protein incorporation into collagen-rich tissues of mice by in vivo pulsed SILAC labelling
eLife 10:e66635.
https://doi.org/10.7554/eLife.66635
  2. Download BibTeX
  3. Download .RIS

Abstract

Collagen-rich tissues have poor reparative capacity that predisposes to common age-related disorders such as osteoporosis and osteoarthritis. We used in vivo pulsed SILAC labelling to quantify new protein incorporation into cartilage, bone, and skin of mice across the healthy life course. We report dynamic turnover of the matrisome, the proteins of the extracellular matrix, in bone and cartilage during skeletal maturation, which was markedly reduced after skeletal maturity. Comparing young adult with older adult mice, new protein incorporation was reduced in all tissues. STRING clustering revealed changes in epigenetic modulators across all tissues, a decline in chondroprotective growth factors such as FGF2 and TGFβ in cartilage, and clusters indicating mitochondrial dysregulation and reduced collagen synthesis in bone. Several pathways were implicated in age-related disease. Fewer changes were observed for skin. This methodology provides dynamic protein data at a tissue level, uncovering age-related molecular changes that may predispose to disease.

Introduction

As life expectancy extends, the societal burden of age-related diseases is predicted to increase substantially. Ageing, the natural decline of cellular and physiological processes during life, is particularly apparent in collagen-rich tissues such as the articular cartilage, bone, and skin, leading to osteoarthritis (OA), osteoporosis, and impaired cutaneous wound healing in the elderly adults. Each of these tissues is characterised by an abundance of extracellular matrix (ECM), a dynamic network composed of collagens, proteoglycans, glycoproteins, and ECM-associated proteins, defined collectively as the matrisome (Naba et al., 2016). The matrisome can constitute the majority of the tissue by volume, up to 95% in the case of articular cartilage (Sophia Fox et al., 2009), and is critically important in controlling the phenotype of the cells embedded within it. The cells responsible for matrix production and homeostasis include those that are post-mitotic after skeletal maturity, such as chondrocytes in articular cartilage and the osteocytes in bone, or those that may be renewable through repopulation (osteoclasts derived from blood mononuclear cells) or through proliferation, for example fibroblasts in skin and osteoblasts in bone (although this is often blunted with age). The ability to examine synthetic activity in these tissues in situ under normal physiological conditions, without disrupting the tissue, is highly valuable.

Several mechanisms have been proposed that account for changes to the ECM with age, including reduced synthesis, altered turnover (reduced and increased), and post-translational modifications. Of the latter, described processes include accumulation of advanced glycation end products (AGEs), deamidation and racemisation of amino acid residues, crosslinking of matrix macromolecules, and accumulation of protein aggregates (Hsueh et al., 2019; Verzijl et al., 2000; Ewald, 2019). These impact on the biophysical properties and regenerative capacity of the tissue and predispose to disease (Kim et al., 2015). In bone, the accumulation of AGEs, combined with collagen loss and cellular senescence, are regarded as critical drivers of osteoporosis (Sanguineti et al., 2014; Shuster, 2005). Loss of type II collagen turnover in mature cartilage is thought to predispose to poor regenerative capacity and OA. Similarly, reduced turnover of the matricellular proteins in dentin, tendon, and skin likely result in impaired regenerative responses and elsewhere may contribute to tissue fibrosis and cancer risk (Neill et al., 2016). In addition to structural and biomechanical impacts of matrisomal changes, signals from ECM protein fragments and growth factors associated with the ECM (Hynes, 2009) can bind directly to cell-surface receptors to modulate cell proliferation and survival, cell morphology, and tissue metabolism (Huang and Greenspan, 2012).

Quantitative proteomic studies provide valuable information regarding molecular composition and abundance of tissue protein, but this only represents a snapshot of the physiological state at a given time and does not accurately account for dynamic protein turnover within tissues (Claydon et al., 2012). A number of methodologies have been applied to investigate dynamic protein changes over time. Radiolabelling methods and methods that measure post-translational modifications (racemisation, AGE, and D-aspartate accumulation), in human skin, cartilage, dentin, and tendon, have identified major fibrillar collagens as very stable proteins with reduced incorporation after skeletal maturity and half-lives as long as 117 years in cartilage (type II collagen) and 15 years in skin (type I collagen) (Heinemeier et al., 2016; Libby et al., 1964; Verzijl et al., 2000; Maroudas et al., 1998; Verzijl et al., 2001). A limitation of these methods is that time is not the only factor influencing these changes. For example, high temperatures, common at injury sites, can accelerate racemisation (Dyer et al., 1993; Maroudas et al., 1992; Stabler et al., 2009). Metabolic labelling is seen as a more direct method to estimate protein lifespans. Incorporation of 14C using the bomb-pulse method takes advantage of an increase in atmospheric 14C that resulted from nuclear testing in the 1950s (Libby et al., 1964; Lynnerup et al., 2008; Nielsen et al., 2016). This method is suitable for estimating turnover of bulk tissue proteins or select protein types but is incompatible with more agnostic approaches using mass spectrometry. Another method using deuterated water (2H2O) has been used to estimate protein synthesis and turnover rates in several animal models and tissues (Choi et al., 2020; Kim et al., 2012). With this approach, animals incorporate deuterium into C-H bonds, thus labelling newly synthesised proteins and other biomolecules. The increase in peptide mass over time indicates neosynthesis. A limitation of this method is that because 2H2O is incorporated into multiple biomolecules, labelled amino acids can derive directly from the diet during the labelling period or from recycled, previously labelled, biomolecules (Alevra et al., 2019).

Stable isotope labelling (Lys (6)-SILAC-Mouse Diet, SILANTES GmbH) incorporated into the diet overcomes many of the aforementioned problems and has been used to determine incorporation rates of proteins from blood cells, organelles, and organs in vivo (Kruger et al., 2008). In this diet, the six naturally abundant 12C molecules in Lysine have been replaced with 13C, conferring a molecular mass shift of 6 Daltons for each lysine in a proteolytically cleaved peptide. This method was originally used to perform quantitative comparisons of the proteome of fully labelled (F2 generation) mice to determine protein dynamics in vivo (Kruger et al., 2008). Quantification of the ‘heavy’ and ‘light’ protein fractions by mass spectrometry over time provides a measure of the incorporation rates of individual proteins (Ebner and Selbach, 2011; Schwanhausser et al., 2009). In the study, we pulse label mice with SILAC diet for 3 weeks at three different post-natal periods: weeks 4–7 (maximum skeletal growth), weeks 12–15 (young adult), and weeks 42–45 (older adult), then apply mass spectrometry to quantify new protein incorporation into three collagen-rich tissues (articular cartilage, bone, and skin) and plasma, which was used as a high-turnover reference tissue, allowing us to estimate and compare new protein incorporation rates at proteome-wide scale across the healthy life course.

Results

Proteome turnover in cartilage, skin, bone, and plasma during skeletal growth

To measure new protein synthesis and incorporation during the period of maximal skeletal growth, we pulse labelled two groups of mice (groups A and B, Figure 1) by feeding them heavy SILAC diet for 3 weeks from 4 weeks of age. Group A was culled for tissue collection immediately after the feeding period, whilst group B was fed the light SILAC diet for another 3 weeks (to week 10). Plasma, knee articular cartilage, ventral skin, and tibial bone were collected for analysis at week 7 (group A) or week 10 (group B). Tissues were extracted and analyzed by mass spectrometry (Figure 1).

Figure 1
Download asset Open asset
Experimental design.

Four groups of four C57BL/6 J male mice were fed with heavy SILAC diet (13C6-Lys) or light SILAC diet (12C6-Lys) for 3 weeks at different ages. Two groups of mice (A and B) were fed with the heavy diet from weeks 4 to 7. Group A was culled for tissue collection, and group B was switched to light diet from weeks 7 to 10, then culled for tissue collection. Groups C and D were fed with the heavy SILAC diet for 3 weeks until week 15 and week 45 respectively. Plasma, knee articular cartilage, tibial bone, and ventral skin were collected from fixed anatomical positions shown above. Left-hand panel shows Safranin O stained coronal sections of a murine knee joint before and after micro-dissection of the articular cartilage. Tissues were processed according to tissue specific protocols, trypsinised, and peptides analysed by liquid chromatography–tandem mass spectrometry. Peptides and protein identification and heavy/light (H/L) ratios were obtained by Maxquant software.

Figure 2 shows the percentage of heavy isotope incorporated into individual proteins at the end of the labelling period (week 7) compared with the amount of labelled protein remaining in that tissue after a 3 week washout period of light isotope diet (week 10). In essence, reflecting proteins that are newly incorporated into the tissue and their stability over time. Each tissue was considered separately. We used the plasma profile to define a fast turnover protein group as previously described (Price et al., 2010). As the animals are growing during this period, the amount of label remaining after the 3 week washout period reflects both protein turnover and label dilution due to tissue expansion. Proteins that showed over 80% incorporation during the heavy diet period and retained less than 20% after the light diet period were considered fast turnover proteins (hashed blue area, Figure 2), while proteins that showed little change in the percentage of isotope incorporation between the end of the heavy diet (week 7) and the end of the light diet (week 10) were regarded as ‘stable proteins’ (hashed pink area, Figure 2). Over 83% (142/171) of plasma proteins fell within the fast turnover group, while only 3.5% (6/171) proteins were within the stable group. Cartilage showed the lowest proportion of fast turnover proteins (18%, 115/634) when compared with bone (48%, 338/712) and skin (39%, 136/352), and was also the tissue that contained the most proteins with less than 40% isotope incorporation during the heavy diet period (bottom left-hand corner of graphs, Figure 2). The patterns of heavy isotope incorporation were similar across the four biological replicates for each tissue and time point (Figure 2—figure supplement 1). The full profile of the four tissue proteomes during skeletal growth are presented in Figure 2—source data 1.

Figure 2 with 1 supplement see all
Download asset Open asset
Synthesis rate and incorporation of protein in plasma, skin, bone, and articular cartilage during skeletal growth.

Each dot represents the mean of the percentage of incorporation of the heavy isotope for an individual protein, n = 4. The x axis represents percentage of the heavy isotope incorporation into proteins from weeks 4 to 7 of age (group A), and the y axis the heavy isotope subsequently lost during the light diet, weeks 7–10 (group B). Collagens are highlighted in red, proteoglycans in yellow, and glycoproteins in blue. The three most stable collagens, proteoglycans, and glycoproteins in each tissue are named. Percentage of stable (red hashed area) and fast turnover proteins (grey hashed areas) for each tissue are indicated.

Figure 2—source data 1

Full profile of incorporation rates (H/L), iBAQ L and iBAQ H values for plasma, skin, bone, and cartilage protein groups during skeletal growth.

Groups A and B shown in Figure 2. Values in red were imputed as follows, missing iBAQ light in >2 samples in group A was assumed to be a consequence of fully labelled protein while missing iBAQ heavy in group B >2 samples as a result of complete turnover. Missing iBAQ light was imputed with the maximum H/L value +0.1 in that column and missing iBAQ heavy with the minimum H/L value –0.01. Only proteins with lysine containing peptides were imputed.

https://cdn.elifesciences.org/articles/66635/elife-66635-fig2-data1-v1.xlsx

When considering the stability of proteins across the whole proteome by examining the ratio (fold change) of % heavy label for each protein at time B compared with time A, a broad range of protein stabilities was demonstrated in each of the tissues (fold change range 0.05–1.06 [plasma], 0.11–1.0 [cartilage], 0.10–0.95 [bone], and 0.08–1.14 [skin]). Of these, matrisome proteins dominated the most stable proteins (ratios ≥ 0.8) particularly in cartilage and skin (12/18 proteins in skin, 14/25 proteins in cartilage, 2/14 proteins in bone, and 0/6 in plasma). The three most stable collagens (red dots), proteoglycans (yellow dots), and glycoproteins (blue dots) are highlighted for each tissue (Figure 2). During this period of skeletal growth, collagens exhibited variable stability across all tissues, and this included the fibrillar collagens (types I, II, III, V, and XI), which are generally regarded as being the most stable. Proteoglycan synthesis was more dynamic, showing lower fold change levels of the B/A ratio. Glycoproteins spanned a wide turnover range in all tissues.

Age-dependent remodelling of the tissue proteome

To investigate how new protein synthesis and incorporation changes during ageing, we pulse labelled another two groups of four male mice: groups C and D. Group C represented skeletally mature (young adult) mice fed heavy diet between 12 and 15 weeks of age. Group D mice were fed heavy diet between 42 and 45 weeks of age (older adult). For each animal, plasma, knee articular cartilage, ventral skin, and tibial bone were processed immediately after the 3 week heavy diet as in A (Figure 1). Histograms showing the frequency of incorporation of the heavy isotope into proteins in groups A, C, and D are shown in Figure 3. Protein synthesis generally decreased in all four tissues with age, as shown by a shift to the left of the histograms. As expected, this shift was more apparent for bone and cartilage between skeletally immature (group A) and young adult animals (group C) than between young (group C) and older adult mice (group D), commensurate with cessation of skeletal maturation. Although all tissues showed an age-related decline in new protein incorporation, protein synthesis was significantly different between the four tissues and at each age group, Kruskal–Wallis test, p<0.0001 for all comparisons. As expected, collagens (shown in red, Figure 3) dominated the proteins incorporating at the lowest levels (lowest % label).

Figure 3
Download asset Open asset
Percentage of incorporation of heavy isotope into newly synthesised proteins in plasma, ventral skin, knee articular cartilage, and tibial bone at different ages.

The percentage of newly synthesised proteins during each heavy diet period is reflected by the percentage of heavy isotope (13C6-Lys) incorporated into proteins. Collagens are labelled in red, proteoglycans in yellow, glycoproteins in blue, and other proteins in grey. Corresponding numbers of a chains are located next to each colour code in legends. New protein synthesis was significantly different between the four tissues and at all age groups, Kruskal–Wallis test, p<0.0001 for all comparisons.

We were able to determine protein synthesis rates in 27 different collagens α chains (Figure 4, Figure 4—source data 1), 21 different proteoglycans (Figure 5, Figure 5—source data 1), and 66 glycoproteins (Figure 6, Figure 6—source data 1) across all tissues and age groups. Fibrillar collagens (I, II, III, V, and XI) showed the greatest decline of incorporation rates between skeletal immaturity and maturity, compared with non-fibrillar: FACIT (fibrillar associated with interrupted triple helices) (IX, XII, XIV, XVI, XXII), network (IV, X), and beaded filament (VI) collagens. Major collagens, type II in cartilage and type I in bone and skin, had the most striking decrease in new incorporation, falling to <2% in cartilage upon reaching skeletal maturity. Conversely, a small number of matrisome proteins were incorporated for the first time in adult tissues such as versican in cartilage and type XV collagen in bone. Compared with collagens, proteoglycans and glycoproteins maintained a higher incorporation rate across the healthy life course.

Figure 4
Download asset Open asset
New collagen incorporation into different tissues during ageing.

The percentage of newly synthesised collagen incorporated into articular cartilage, tibial bone, and ventral skin is estimated by the incorporation of the heavy isotope (13C6-Lys) into proteins during the 3 weeks of heavy diet. The three periods of heavy diet were compared: skeletal growth (4–7 weeks old), young adults just after skeletal maturity (12–15 weeks old), and older adults (42–45 weeks old). Error bars represent the standard deviation (n = 4). ND = not detected, protein group not present in the group dataset. NQ = not quantified, protein group present, but with <2 samples quantified. NL = not labelled, protein group with only light isotope quantified. Statistical significance was determined by pairwise comparisons using Perseus software, Student’s t-test reporting Benjamini-Hochberg adjusted p<0.05. Statistically significant differences were denoted as follows: &between groups A and C, $between groups A and D, ^between groups C and D, #between groups A–C and A–D, *between all three groups.

Figure 4—source data 1

Data supporting Figure 5.

Percentage of new collagen incorporation in the three collagenous tissues. Cartilage, bone, and skin of skeletally growing mice (group A), young adults (group C), and older adults (group D). Mean ± SD, n = 4, Pairwise comparisons using Student’s t-test, p<0.05 significant.

https://cdn.elifesciences.org/articles/66635/elife-66635-fig4-data1-v1.xlsx
Figure 5
Download asset Open asset
New proteoglycan incorporation rate into different tissues during ageing.

The percentage of newly synthesised proteoglycans incorporated into articular cartilage, tibial bone, and ventral skin is estimated by the percentage of incorporation of the heavy isotope (13C6-Lys) into the proteins during the 3 weeks of heavy diet. Protein synthesis and incorporation was estimated across the healthy life span, covering skeletal growth (4–7 weeks old), young adults just after skeletal maturity (12–15 weeks old), and older adults (42–45 weeks old). Error bars represent the standard deviation (n = 4). ND = not detected, protein group not present in the group dataset. NQ = not quantified, protein group present, but with <2 quantified samples. BMPg, bone marrow proteoglycan. Statistical significance was determined by pairwise comparisons using Perseus software, Student’s t-test reporting Benjamini-Hochberg adjusted p<0.05. Statistically significant differences were denoted as follows: &between groups A and C, $between groups A and D, ^between groups C and D, #between groups A–C and A–D, *between all three groups.

Figure 5—source data 1

Data supporting Figure 5.

Percentage of new proteoglycan incorporation in the three collagenous tissues. Cartilage, bone, and skin of skeletally growing mice (group A), young adults (group C), and older adults (group D). Mean ± SD, n = 4, Pairwise comparisons using Student’s t-test, p<0.05 significant.

https://cdn.elifesciences.org/articles/66635/elife-66635-fig5-data1-v1.xlsx
Figure 6
Download asset Open asset
Heatmap of new glycoproteins incorporation rates into different tissues during ageing.

The percentage of newly synthesised glycoproteins incorporated into articular cartilage, tibial bone, and ventral skin is estimated by the percentage of incorporation of the heavy isotope (13C6-Lys) into the proteins during the 3 weeks of heavy diet. Protein synthesis and incorporation was estimated across life, covering skeletal growth (4–7 weeks old), young adults just after skeletal maturity (12–15 weeks old), and older adults (42–45 weeks old).

Figure 6—source data 1

Data supporting Figure 6.

Percentage of new glycoprotein incorporation in the three collagenous tissues. Cartilage, bone, and skin of skeletally growing mice (group A), young adults (group C), and older adults (group D). Mean ± SD, n = 4, Pairwise comparisons using Student’s t-test, p<0.05 significant.

https://cdn.elifesciences.org/articles/66635/elife-66635-fig6-data1-v1.xlsx

To examine changes in new protein incorporation across the whole tissue proteome with age, we compared the whole proteome profile in skeletally mature groups D and C. Data are represented in volcano plots (Figure 7). Significantly regulated proteins with –1.5 > FC > 1.5, and with Benjamini–Hochberg (BH) correction FDR p<0.05, for each tissue, are highlighted (bold points). Overall, of those proteins whose incorporation changed with age, most showed reduced incorporation. In connective tissues these were 37 from 452 in skin, 291 from 572 in bone, and 175 from 597 in cartilage. In plasma, 83 of 180 proteins exhibited statistically significant regulation upon ageing (proteins shown in Figure 7—source data 1). A small number of proteins had increased incorporation with age. Only two regulated proteins were common among the three collagenous tissues; H1f0, a histone protein, down-regulated in cartilage and bone but upregulated in skin, and Rpl12, a structural constituent of the ribosome 60 S subunit, down-regulated in the three tissues. Articular cartilage and tibial bone proteomes shared 9.6% of the statistically significant, regulated proteins while overlap between bone and skin was 1.1%, and cartilage and skin 0.7%. Total protein levels in each tissue, assessed by iBAQ measurements, remained unchanged post-skeletal maturity with just seven proteins changing significantly in articular cartilage (Figure 7—figure supplement 1). This likely reflects the substantially enhanced sensitivity of the pulse SILAC methodology to detect alterations in protein homeostasis.

Figure 7 with 3 supplements see all
Download asset Open asset
Changes in protein synthesis and incorporation rates during tissue remodelling.

(A–D) Differential protein incorporation rates in young (15 weeks) versus older (45 weeks) adult tissues. (A) Cartilage, (B) bone, (C) skin, and (D) plasma. Volcano plots, unpaired Student’s t-test with BH correction FDR < 0.05, FC > 1.5, n = 4. Full list of proteins available in Figure 7—source data 1. (E–H) STRING cluster analysis of differentially incorporated proteins in each tissue. (E) Cartilage, (F) bone, (G) skin, and (F) plasma. Edges show high confidence interaction score = 0.7. Networks clustered to MCL inflation parameter = 2. Cluster elements are listed in Figure 7—source data 2.

Figure 7—source data 1

Data supporting Figure 7A–D.

New protein incorporation across whole tissue proteomes with age. Unpaired Student's t-test statistics with BH corrected p values (p<0.05 significant) of all protein groups included in the volcano plots, plus list of proteins that were recovered using iBAQ heavy or light data.

https://cdn.elifesciences.org/articles/66635/elife-66635-fig7-data1-v1.xlsx
Figure 7—source data 2

Data supporting Figure 7E–H.

STRING clusters with protein description. Proteins with statistically significant changes between groups C and D (unpaired Student's t-test BH corrected p>0.05) were clustered using STRING software. Networks were built with edges high confidence interaction score of 0.7 and clustered applying MCL algorithm with inflation parameter = 2.

https://cdn.elifesciences.org/articles/66635/elife-66635-fig7-data2-v1.xlsx

STRING protein interaction networks of statistically significant, regulated proteins for the four tissues are shown in Figure 7, with full list of cluster elements in Figure 7—source data 2. Each of the tissues shared a common cluster relating to regulatory elements involving ribosomal and histone proteins, albeit with different constituent proteins in each tissue. The largest cluster of interaction networks in cartilage (Figure 7) was mainly formed by ECM proteins such as collagens, collagen processing-related, and proteoglycans. Also notable, were two clusters abundant in proteoglycans and growth factors, and another cluster rich in myosin family members. Fifteen proteins with decreased incorporation have previously been implicated in OA development (Figure 7—source data 2, final column).

In bone, the largest cluster represented several myosin family members, involved in muscle contraction, some structural proteins, and proteins with roles in ATP transfer. Two unique bone clusters were involved in energy metabolism; one formed by 29 mitochondrial proteins such as ATP synthases, cytochrome oxidases, and NADH dehydrogenases, and the other largely by enzymes involved in glycolysis and fatty acid metabolism (Figure 7, Figure 7—source data 2). Five proteins with a statistically significant decrease in incorporation have previously been linked to osteoporosis including biglycan, Col1a2 and troponin C2, YWHAE and Cd44 (Figure 7—source data 2, final column).

One striking result was the relative paucity of the percentage of age-regulated proteins in skin (8%) compared with cartilage (29%) and bone (51%) despite a similar number of total labelled proteins quantified (Figure 7—figure supplement 2). In skin, a number of weak clusters were identified with a couple involving energy metabolism (Figure 7, Figure 7—source data 2). In plasma, one strong cluster comprising 40 proteins with roles in lipid transport (apolipoproteins), proteins involved in the innate immune response, and several serpin family members, was identified (Figure 7, Figure 7—source data 2). Further bioinformatic pathway analysis was performed using Database for Annotation, Visualization and Integrated Discovery (DAVID), which produced similar results to STRING (data not shown), and Ingenuity Pathway Analysis (IPA). IPA identified a number of enriched cellular pathways (Figure 7—figure supplement 2), but none of these reached corrected statistical significance.

Discussion

To extend our knowledge of connective tissue renewal during growth and ageing, we used metabolic labelling with stable isotopes in combination with quantitative proteomics to estimate protein incorporation rates in live animals across the healthy life course. Here, we describe tissue proteome incorporation rates for proteins in skin, cartilage, bone, and plasma at three post-natal stages of life: during maximum skeletal growth, and at young and older adulthood. Our study shows that although new protein incorporation changes significantly in all tissues after skeletal maturity, it displays distinct temporal and molecular tissue signatures.

Tissue turnover can be divided into tissue modelling, which occurs during growth and development, and tissue remodelling, which is required to repair or adapt the tissue to maintain homeostasis (Frantz et al., 2010; Karsdal et al., 2016). Collagen-rich tissues are believed to have a limited reparative capacity, mainly due to the inability to incorporate new fibrillar collagens into mature matrices (Verzijl et al., 2000). Our results quantified incorporation of multiple collagens in each of the connective tissues, with almost all of them showing decreases in incorporation with age. This was most marked for bone and cartilage. These tissues are largely post-mitotic after skeletal maturity, and although they retain global synthetic activity in adult life, this is blunted in a sizeable fraction of proteins (29% bone, 51% cartilage) by 45 weeks of age compared with skin (8%), in which the tissue is known to renew throughout adult life (Visscher et al., 2015). Bone was the only tissue that showed a strong collagen cluster (enriched in 12 collagens) when comparing young with older adult tissue, although, as expected, incorporation of new type II collagen fell to <2% in cartilage in the older group. In skin, significant decreases were also observed in fibrillar collagens, but the transition between skeletally immature and mature tissues was less marked, with substantial on-going incorporation of fibrillar collagens through young and older adulthood.

Previous age-related research has identified several cellular pathways that are consistently found to be dysregulated. Cell energy and metabolism, including mitochondrial dysfunction, is one of the top processes (Campisi et al., 2019; Javadov et al., 2020; Sprenger and Langer, 2019). Proteins such as clusterin, superoxide dismutase 1 (SOD1), and apolipoprotein E have previously been found to be dysregulated in ageing (Sabaretnam et al., 2010; Sentman et al., 2006; Trougakos and Gonos, 2006). Although we were underpowered to gain significant insights from IPA and DAVID pathway analysis packages, by STRING analysis bone displayed a very strong metabolic, mitochondrial signature, and we found clusterin dysregulation in plasma, cartilage, and bone of older adult mice, and down-regulation of SOD1 in cartilage. Modulators of protein synthesis such as heterogeneous nuclear ribonucleoprotein D, which affects stability and metabolism of mRNAs, were also down-regulated in cartilage and have previously been linked to ageing (Pont et al., 2012). Our data showed only two proteins where reduced incorporation was seen in all three collagenous tissues. These included the ribosomal protein, RPL12 and Histone H1.0. Moreover, strong down-regulated clusters, rich in ribosomal proteins, were present in bone, cartilage, and skin, indicating a decline in incorporation rates of ribosomal structural proteins across all tissues as the animal ages. This finding agrees with several studies in which progressive decreases in the expression of ribosomal proteins or rRNA has been observed with age (D’Aquila et al., 2017; Jung et al., 2015).

In cartilage, 15 of the down-regulated proteins have been associated with OA. Several of these have known protective functions in cartilage tissue homeostasis, such as Prg4 (also known as lubricin) (Flannery et al., 2009), Timp3, an inhibitor of disease-associated metalloproteinases (Nakamura et al., 2020), and LDL receptor-related protein 1 (Lrp1), an important chondrocyte scavenger receptor (Yamamoto et al., 2013). In addition, a number of pro-repair and chondrogenic growth factors, all known to be sequestered in the pericellular matrix, were reduced in older adult cartilage such as fibroblast growth factor 2 (FGF2), connective tissue growth factor, transforming growth factor beta (TGFβ1), and Frzb, a Wnt antagonist (Chia et al., 2009; Tang et al., 2018; Khan et al., 2011; Van Der Kraan, 2017; Evangelou et al., 2009; Huang et al., 2019; Zhong et al., 2017). These results point towards loss of repair pathways in the pathogenesis of age-related OA, a conclusion supported by recent genome-wide association studies (Tachmazidou et al., 2019).

The balance between bone formation and resorption is compromised with increasing age, resulting in progressive net bone loss and osteoporosis (Javaheri and Pitsillides, 2019). Three proteins whose incorporation rates decreased with age are regarded as biomarkers of osteoporosis. One of these was biglycan, a bone modifying protein in murine osteoporosis (Cui et al., 2019) and collagen 1a2, the most abundant protein in bone matrix (Nagy et al., 2020). In addition, we found decreased incorporation in two proteins, 14-3-3 protein epsilon and Cd44. SNPs in these two protein coding genes, YWHAE and CD44, are associated with low bone mineral density in two recent GWAS studies (Morris et al., 2019; Medina-Gomez et al., 2018).

In skin, only 8% of proteins had altered incorporation rates in the older group, considerably lower than in cartilage and bone (29% and 51%, respectively). Six clusters comprising a maximum of four proteins in each were discerned. Wound healing is known to be impaired in the elderly adults, but this phenotype is probably less apparent in ‘middle aged’ individuals which would better represent our older adult mice (Gardiner, 2011). One down-regulated protein in the skin, solute carrier family 3 (SLC3A2) (Figure 7—source data 1), has been implicated in wound healing (Falanga, 2005; Grande-Garcia et al., 2007; Boulter et al., 2013), and ageing (Tissot et al., 2018). Although there were several proteins with statistically significant differences in incorporation rates in older skin, the clusters did not predict an overall negative effect on wound healing although this tissue signature might not be expected in non-wounded tissues.

We recognise a number of limitations in this study. For cartilage and bone, an insoluble pellet remained after sample processing despite multiple refinements to the protocol for each tissue. It is likely that the insoluble fraction is mainly composed of the oldest, highly cross-linked portion of fibrillar collagens that increases with age (Robins, 2007; Verzijl et al., 2000). Therefore, in the older group, the proportion of newly synthesised collagens might be overestimated. It is also the case that the insoluble pellet may have been influenced by a change in the ratio of cortical to trabecular bone with age. Another limitation was that we only studied mice until 45 weeks of age, approximately half the lifetime of a laboratory mouse. We did not include older mice because they frequently develop spontaneous OA, which would have been a co-founding variable in cartilage, preventing us from discerning effects due to age and those due to disease. Likewise, in bone, insidious loss of tissue density occurs with age (Jilka, 2013). In this study, we only examined male mice to keep our initial analysis simple. There is a clear gender difference in the risk of OA and osteoporosis in both humans and mice, so a male/female comparison would be a valuable future proposal to uncover important sex-dependent biological differences that contribute to disease risk. Finally, we limited this analysis to only three contrasting connective tissues, each of which is distinct in terms of cell type and number. As the labelled mouse carcasses are currently frozen, the authors welcome collaborative approaches to extend this analysis to other tissues in line with the three R principles (Percie du Sert et al., 2020).

Compared with protein abundance, which only captures a snapshot of a protein at a particular time, this methodology provides insights into dynamic protein synthesis in different tissues and at different ages. Whilst we do not measure the stability of individual proteins, we reveal distinct age-dependent synthetic activities in dividing and post-mitotic connective tissues and identify shared and tissue-specific protein clusters that could represent novel insights into age-dependent disease.

Materials and methods

Key resources table
Reagent type (species) or resourceDesignationSource or referenceIdentifiersAdditional information
C57BL/6 J (Mus musculus, male)C57BL/6 J4 weeks old mice were from in house breeding colony and 11 and 41 weeks old from Charles rivers laboratories, Oxford, EnglandRRID:MGI:5655520Wild type
Chemical compound, drug13C6-Lysine-SILAC (97 % atom 13C)SILANTES GmbH, Munich, Germany13C-aa-Lys(6)-SILAC-Mouse diet-252923926
Chemical compound, drug12C6-Lysine-SILAC dietsSILANTES GmbH, Munich, Germany12C-aa-Lys(0)-SILAC-Mouse diet-230004600
Software, algorithmGraphPad Prism 7GraphPad Software Inc, San Diego, CA, USAPrism: RRID:SCR_015807GraphPad Prism 7
Software, algorithmMaxQuantMaxQuantMaxQuant: RRID:SCR_014485MaxQuant v1.5.7.4
Software, algorithmPerseusPerseusPerseus: SCR_015753Perseus v1.6.1.1
Software, algorithmClinical Knowledge Graph’s analytics coreCKG
Software, algorithmSTRINGSTRINGSTRING: SCR_005223STRING v11
Software, algorithmIPAQIAGEN IncQIAGEN IPA: SCR_008653
Software, algorithmThe Database for Annotation, Visualization and Integrated Discovery (DAVID ) v6.8DAVIDDAVID v6.8

Samples and reagents

All animal experiments were carried out with Institutional Ethical Approval Under Animals (Scientific Procedures) Act 1986 (Licence Number 30-3129)

Request a detailed protocol

Wild-type C57BL/6 J mice, ages 4, 11, and 41 weeks, were obtained originally from Charles River, Oxford, England. Mice from experimental groups A and B were taken from a breeding colony of these mice within our animal unit. Groups C and D were purchased directly for the experiment. 13C6-Lysine-SILAC (97% atom 13C) (heavy diet) and 12C6-Lysine-SILAC (light diet) were purchased from SILANTES GmbH, Munich, Germany. Animals’ care was in accordance with institutional guidelines.

Pulsed SILAC labelling experimental design

Request a detailed protocol

Our labelling strategy was aimed at labelling long lived ECM proteins. Therefore, four groups of four male mice were fed with a stable isotope diet for a period of 3 weeks at different ages that spanned from weaning to late adulthood. The labelling scheme (Figure 1A) was as follows: Two groups of mice (A and B) were fed with the heavy diet from weeks 4 to 7. Then, one of these groups was culled for tissue collection and the other was changed to the light isotope diet (12C6-Lys) from weeks 7 to 10, then culled for tissue collection. This was used to calculate proteome turnover rates during skeletal growth. The third group (C) was fed with the heavy isotope diet from weeks 12 to 15 (young adult), and the fourth group (D) was fed with the heavy isotope diet from weeks 42 to 45 (late adult). All mice were acclimatised to the SILAC diet formulation by feeding them with the light isotope diet for 1 week before introducing the heavy diet. All mice gained weight in accordance with reference laboratory data for C57BL/6 mice (data not shown).

Tissue harvest

Request a detailed protocol

At the end of each labelling period, corresponding to ages 7, 10, 15, and 45 weeks, animals were culled injecting a terminal dose (30 mg/animal) of the anaesthetic Pentoject. Blood, knee articular cartilage, tibial bone, and ventral skin were collected from each animal (n = 4 mice per time point) as follows: blood was collected by cardiac puncture and mixed with 0.5 M anticoagulant EDTA to obtain a final concentration of 5 mM EDTA. The buffered blood was centrifuged for 15 min at 3000 rpm at 4°C. The resulting supernatant (plasma) was transferred to a 0.5 ml tube to be used for analysis. Knee articular cartilage was harvested using a micro-dissection technique previously developed described by our group (Gardiner, 2011). Articular cartilage from the femoral and tibial surfaces of both knees of one mouse were micro-dissected under the stereo microscope and collected in 1.5 ml micro-centrifuge tubes containing 50 μl PBS. Tibial bone samples were sectioned between the crest of the tibia and the insertion of the fibula. The bone marrow was flushed three times with PBS to remove cells using a 25 G needle. The final wash was checked for lack of cells using a stereo microscope. Approximately 0.8 cm2 of skin was cut from the lower flank of the ventral surface. Hair was removed using hair removal cream (Veet sensitive skin). Adipose tissue and blood vessels were removed from the subcutaneous region under the stereo microscope, and a portion of clean skin was collected using a 6 mm biopsy punch. All tissues were washed with phosphate-buffered saline (PBS [pH 7.4]) and stored frozen at −80°C until further processing.

Mass spectrometry sample preparation

View detailed protocol

Bone and cartilage samples were placed in 180 µl of 5 mM dithiothreitol (DTT) and heated at 65°C for 15 min. After cooling to room temperature, samples were alkylated with 20 mM iodoacetamide (IAA) for 30 min. To quench remaining IAA, DTT was added and samples were incubated for 30 min. Then, samples were incubated with 4 M GuHCl for 2 hr. The samples were adjusted to pH 8 with 400 mM Tris base. Proteins were digested with 1 µg of trypsin overnight at 37°C. Digestion was terminated adding trifluoroacetic acid (TFA) to a final concentration of 0.5–1%. Peptides were purified using C18 solid phase extraction cartridges (SOLA HRP SPE cartridges, Thermo Fisher Scientific).

Five microlitres of plasma was reduced with 5 mM DTT, alkylated with 20 mM IAA, and proteins precipitated with methanol/chloroform (Wessel and Flugge, 1984). Precipitated proteins were solubilised in 6 M urea buffer. Urea was diluted to <1 M with milli-Q water and proteins digested with trypsin overnight at 37°C at an enzyme to substrate ratio of 1:25. Digestion was terminated adding TFA to a final concentration of 0.5–1%. Peptides were purified using C18 solid-phase extraction cartridges as above.

Skin samples were grinded in a Cryomill, then incubated in 8 M urea, 3% SDS with protease inhibitors (cOmplete Mini EDTA-free Protease Inhibitor Cocktail, Roche) at room temperature for 1 hr in a total volume of 3 ml. As there was still a substantial pellet, samples were shaken overnight at 4°C to solubilise the pellet. Samples were centrifuged at 2500 g for 10 min and 100 µl of sample was taken for further processing. Proteins were reduced with 5 mM DTT for 30 min at room temperature, alkylated with 20 mM IAA for 30 min at room temperature, and precipitated with methanol/chloroform. Precipitated proteins were solubilised in 6 M urea buffer. Urea was diluted to <1 M with milli-Q water and proteins digested with trypsin overnight at 37°C at an enzyme to substrate ratio of 1:25. Digestion was terminated adding TFA to a final concentration of 0.5–1%. Peptides were purified using C18 solid-phase extraction cartridges as above.

Liquid chromatography–tandem mass spectrometry (LC–MS/MS)

Request a detailed protocol

Tissues from groups A and B as well as skin samples were analysed on a LC–MS/MS platform consisting of Orbitrap Fusion Lumos coupled to an UPLC ultimate 3000 RSLCnano (Thermo Fisher Scientific) and samples from groups C and D with similar platform but coupled to a Q-Exactive HF. Samples were loaded in 1% acetonitrile and 0.1% TFA and eluted with a gradient from 2% to 35% acetonitrile in 0.1% formic acid and 5% DMSO in 60 min with a flow rate of 250 nl/min on a 50 cm EASY-Spray column (ES803, Thermo Fisher Scientific).

Orbitrap Fusion Lumos

Request a detailed protocol

The survey scan was acquired at a resolution of 120,000 between 400 and 1500 m/z and an AGC target of 4E5. Selected precursor ions were isolated in the quadrupole with a mass isolation window of 1.6 Th and analysed after CID fragmentation at 35% normalised collision energy in the linear ion trap in rapid scan mode. The duty cycle was fixed at 3 s with a maximum injection time of 35 ms, AGC target of 4000, and parallelisation enabled. Selected precursor masses were excluded for the following 27 s.

Q-Exactive HF

Request a detailed protocol

The survey scan was acquired at a resolution of 60,000 between 375 and 1500 m/z with an AGC target of 3E6, up to the top 12 most abundant ions were selected for fragmentation from each scan. Selected precursor ions were isolated with a mass isolation window of 1.2 Th and fragmented by HCD at 28% normalised collision energy. Fragment scans were acquired at 30,000 resolution with an AGC target of 5E4 and a maximum injection time of 100 ms. Selected precursor masses were excluded for the following 27 s.

Protein identification

Request a detailed protocol

Raw mass spectral data files were searched using MaxQuant software (V1.5.7.4, Tyanova et al., 2016b) using SILAC (Lys6) quantitation. Fixed modification was carbamidomethylation of cysteine, and variable modifications were oxidised methionine, deamidation of asparagine and glutamine, acetylation at protein N-terminal, and hydroxylation of proline. The data was searched against the mouse canonical Uniprot database (29/07/2015). FDR on peptide and protein level were set to 1%. Second peptide and ‘match between runs’ options were enabled, all other parameters were left at default settings. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE (Perez-Riverol et al., 2019) partner repository with the dataset identifier PXD023180. Reviewer account details are as follows: Username: reviewer_pxd023180@ebi.ac.uk, Password: iSH0ppKX.

Statistical analysis

Request a detailed protocol

After MaxQuant analysis, Excel version 1.5. 21.1, GraphPad Prism 7 (GraphPad Software Inc, San Diego, CA), STRING, Perseus software version 1.6.1.1 (Tyanova et al., 2016a), DAVID, IPA (QIAGEN Inc), and Python libraries from the Clinical Knowledge Graph’s analytics core (Santos et al., 2020) were used for data visualisation, statistical analysis, and pathway enrichment and protein network analysis.

Proteins were examined for normality against the Shapiro–Wilk normality test and for variance against the Levene test and demonstrated that >80% of proteins were normally distributed. Pairwise comparisons were therefore by Student’s t-tests. For group wise comparisons, where the data were not normally distributed, non-parametric tests were used. For those proteins with a heavy to light ratio (H/L), the percentage of label incorporation was estimated as follows: (H/L/((H/L) + 1)) * 100. To discern protein turnover profiles in tissues, we plotted the percentage of label present in group A versus group B. In those cases where H/L ratios were not reported, we used iBAQ light (iBAQ L) and heavy (iBAQ H) data to impute H/L ratios. Missing iBAQ light in >2 samples in group A was assumed to be a consequence of fully labelled protein while missing iBAQ heavy in >2 samples in group B as a result of complete turnover. H/L ratios from samples with missing iBAQ L values were imputed with the maximum H/L value +0.1 in that column and those with missing iBAQ H values with the minimum H/L value – 0.01 (only proteins with lysine containing peptides were imputed). To define what proteins were differentially incorporated when comparing groups D and C (unpaired t-test), we selected proteins with at least two valid values for H/L, and imputed missing values using the K-nearest neighbour algorithm. Furthermore, to account for full labelling or absence of new protein incorporation, we identified proteins where H/L ratios were reported in one group, but not in any of the four replicates in the other group. In those cases, we used heavy or light iBAQ values present in at least two replicates and imputed H/L ratio using uniform random values near the maximum (maximum, maximum+ std) or the minimum (0, minimum) of the dataset, respectively.

The recovered proteins that showed changes in incorporation between groups were added to the STRING networks, DAVID, and IPA analyses.

The STRING networks were built using high confidence interaction score = 0.7, and networks clustered to MCL inflation parameter = 2 (Szklarczyk et al., 2019).

Data availability

Proteomics raw data and Maxquant output files have been deposited to the ProteomeXchange Consortium via the PRIDE (Perez-Riverol et al., 2019) partner repository with the dataset identifier PXD023180. The rest of the data generated or analysed are included in the manuscript and supporting files. Source data files have been provided for all figures.

The following data sets were generated
    1. Roman F
    (2021) PRIDE
    ID PXD023180. Age-dependent changes in protein incorporation into collagen-rich tissues of mice by in vivo pulsed SILAC labelling.
    https://www.ebi.ac.uk/pride/archive/projects/PXD023180

References

    1. Jilka RL
    (2013) The relevance of mouse models for investigating age-related bone loss in humans
    The Journals of Gerontology. Series A, Biological Sciences and Medical Sciences 68:1209–1217.
    https://doi.org/10.1093/gerona/glt046

Decision letter

  1. Cheryl Ackert-Bicknell
    Reviewing Editor; University of Colorado, United States
  2. Matt Kaeberlein
    Senior Editor; University of Washington, United States

Our editorial process produces two outputs: (i) public reviews designed to be posted alongside the preprint for the benefit of readers; (ii) feedback on the manuscript for the authors, including requests for revisions, shown below. We also include an acceptance summary that explains what the editors found interesting or important about the work.

Acceptance summary:

Capturing the rate and degree of protein half-life in tissues rich in collagen, proteoglycans and glycoproteins over the life span provides valuable information about how these tissues age and can increase our understanding of age-related disease. Using a Stable Isotope Labeling (SILAC) method to examine new protein Lys6 incorporation rates in three high collage content tissues, the authors show that protein-change is low in older mice in these tissues, but the depth of the data generated provide a detailed examination of the impact of age at a level we have not previously had. This paper would be of interest to a broad range of scientists studying connective tissues in the context of development and ageing.

Decision letter after peer review:

Thank you for submitting your article "Age-dependent changes in protein incorporation into collagen-rich tissues of mice by in vivo pulsed SILAC labelling" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, one of whom is a member of our Board of Reviewing Editors, and the evaluation has been overseen by Matt Kaeberlein as the Senior Editor. The reviewers have opted to remain anonymous.

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

Essential revisions:

1. A major weakness is the tissue/cell turnover and protein turnover aspect. The authors compared plasma with cartilage, bone and skin which contain a diverse set of distinct cell populations. In addition, labelled proteins from plasma are derived from the liver, which represents a proliferating tissue. Similarly, skin is also a proliferating tissue, while cartilage and bone tissue are most likely post-mitotic. It is not surprising that proliferating tissues have higher Lys6 incorporation compared to non-dividing cells. Unfortunately, the authors did not address this issue, and the comparison unfortunately remains unclear.

2. A more focused analysis of a particular cell type and its Lys6 incorporation rate at different time points or disease states is a viable alternative approach, which would also be highly informative, if not more so for selected questions. In light of the comment above, a more complete discussion of the limitations of this work is needed. It would be worth it to add a justification of why the methods chosen where chosen over other strategies to address the main thesis of the paper.

3. Overall, the statistical power is difficult to estimate and the authors should carefully consider the statistical analysis of a non-normal distributed data set. The comparison of young animals with the older stages simply represents growing and dividing tissues, which is naturally more labelled than "mature" tissue.

4. Throughout this paper, it is stated that trabecular bone was examined. The methods for collecting this bone are spartan. It appears that the part of the tibia collected is very trabecular bone poor and is mostly cortical bone. The turnover rates of these two bone compartments are very different. Please clarify what kind of bone you had and the implications of looking at only trabecular/ only cortical/ a mix for the interpretations of these results. Considering the tissue of interest appears to be trabecular bone, how did you ensure there was no contaminating marrow. Trabeculae are very small and hard to "clean".

5. This study is only male mice. Bone turnover is higher in females. In humans, females are more likely to get Osteoarthritis than males and wound healing is slower in females than males. The implications of sex on the interpretation of the results and the limitations of this study design need to be included in the discussion.

6. The authors claim that incorporation is lower in aged animals (week 45) compared to young animals (week 15). For example, Table S2 shows an incorporation rate of 2.97 +/- 2.13 at week 45 and 3.38+/-0.98 at week 15 for the cartilage tissue and protein collagen 1a1. The ANOVA calculation showed a significant change for this candidate. Do the authors believe that a fold-change of w45/w15 = 1:1.14 is biologically relevant? The same is true for most of the proteins listed in the table. Is it possible to calculate accurate incorporation rates in % of SILAC ratios of H/L = 0.01? This corresponds to an H/L ratio of 1:100 in the MS spectra. It would be helpful to show a real MS spectrum of such incorporation rates at least in the SI data. This would also show if the "heavy" peak is represented by a true isotopic pattern and whether this peak was selected for fragmentation. Alternatively, it might be that the "low" peak represents only a background signal. The calculation of the coefficient of variation could also help to document the variability of the data set. A more complete comparison of this work to similar previous studies in the field would provide both context for these results and validation of the expected ranges of turnover.

7. The median lifespan of C57BL/6J male mice is ~128 weeks. The growth plate of mice does not fuse and the femur will increase in length past 26 weeks of age and periosteal expansion is still happening at 12 months (PMID: 13678781). A sound rational for using mice under one year of age is provided. But please be careful when describing these mice. The oldest group are at best mature adult male mice and are certainly not "late adulthood" (line 229-230). The middle group are far from finished growing.

8. Could it be that older animals move less and eat less? This would also explain a slightly reduced Lys6 incorporation. Did the authors observe any possible weight gain or loss during feeding? Did the animals always take-up the same amount of Lys6 during the experiment?

9. There is a large difference between wound healing and normal skin maintenance/turnover. This this distinction is often blurred when discussing the results for skin and the implications of the results for skin.

10. Line 329-332. Mice very much lose trabecular bone without losing estrogen (PMID: 17488199) and naturally occurs without immobilization. Please reframe this section reflecting the reality of you model.

11. It might be better to focus the analysis on a specific tissue/cell type that is either proliferating or post-mitotic. Why did the authors not calculate absolute turnover rates? This would be a better comparison and could help to better assess the statistical power of their method.

12. Introduction – for completeness, please briefly also describe the use of deuterium labelling to measure proteome dynamics in tissues:

https://pubmed.ncbi.nlm.nih.gov/22915825/, https://pubmed.ncbi.nlm.nih.gov/32393437/ and https://pubmed.ncbi.nlm.nih.gov/32403418/

13. It is strongly suggested that the authors establishing if there were any alterations in relative protein abundance between age groups assessed. This will allow identification of any proteins that show both dysregulated turnover and abundance with ageing and are therefore most likely involved in age-related diseases.

14. Have the authors using Ingenuity Pathway Analysis (or similar software) to identify upstream regulators predicted to result in the differential protein synthesis observed with ageing? This could provide potential targets for age-related diseases that could be investigated further in future studies. The detailed bioinfomatic STRING analysis is certainly helpful but many points raised in the discussion are unfortunately very speculative and should be verified by experimental work. All reviewers were in agreement that a more comprehensive bioinformatics data analysis of some type is mandatory.

15. Is it possible to calculate and report individual protein half-lives/fractional synthesis rates from the data? This would enable more direct comparison of half-lives measured in previous studies.

16. Materials and methods – different extraction protocols were used for skin, compared to cartilage and bone samples. Please provide a rationale for this. Is this likely to have had an impact of the proteins identified? If so, this needs to be more completely mentioned in the discussion as a limitation of this work.

17. Results, Ln 160, Figures 3 and 4. It is apparent that incorporation rates decline with ageing across tissues, with most obvious differences between immature and mature animals and much smaller changes between mature and old animals. Are you able to present statistical analysis of these results on the graphs to identify significant differences with ageing?

18. Although the authors achieved a dilution of the label by a label switch to Ly0, this aspect was unfortunately only very little evaluate. The reviewers felt that this point should be elaborated on.

19. Protein turnover is always regulated by synthesis and degradation. The labelling of living animals with stable isotopes is a useful technique, but one should be very careful which tissues one compares and whether the incorporation rates are regulated by instrinsic protein stability or by cell division.

[Editors' note: further revisions were suggested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "Age-dependent changes in protein incorporation into collagen-rich tissues of mice by in vivo pulsed SILAC labelling" for further consideration by eLife. Your revised article has been evaluated by 3 peer reviewers, one of whom is a member of our Board of Reviewing Editors, and the evaluation has been overseen by Matt Kaeberlein as the Senior Editor. The reviewers have opted to remain anonymous.

All reviewers agreed that this manuscript is substantially clearer and very much improved after this extensive revision. Some matters however remain, and it was agreed after much discussion among the reviewers that an additional revision would be requested

1. The figure provided for A6 would be of benefit as a supplemental figure. However, in its current form it is not usable as the font size is impossible to read.

2. The largest concern from the reviewers was that the study remains descriptive and the underlying experiments comparing Lys-6 incorporation rates between different tissues or developmental stages remain questionable. While the authors have addressed this to some degree in the discussion (ln 457), further comment on the specific effect of cell division on Lys-6 incorporation would more clearly highlight the limitations of the study. The reviewers would have liked to see an integration of the absolute half-lives as a function of proliferation. Thus, the comparison always remains dependent on cell division, which unfortunately prevents a clear statement of the absolute protein half-life or protein turnover. Therefore, the wording "turnover" should not be used in the manuscript at all, but rather Lys6 incorporation rates (i.e. Figure 2A) in proliferating or postmitotic cells. Lys6 incorporation rates do not necessarily reflect the turnover of a protein and hence this term, while used sparingly, should be avoided overall.

https://doi.org/10.7554/eLife.66635.sa1

Author response

Essential revisions:

1. A major weakness is the tissue/cell turnover and protein turnover aspect. The authors compared plasma with cartilage, bone and skin which contain a diverse set of distinct cell populations. In addition, labelled proteins from plasma are derived from the liver, which represents a proliferating tissue. Similarly, skin is also a proliferating tissue, while cartilage and bone tissue are most likely post-mitotic. It is not surprising that proliferating tissues have higher Lys6 incorporation compared to non-dividing cells. Unfortunately, the authors did not address this issue, and the comparison unfortunately remains unclear.

Thank you for raising this important point which we failed to highlight in our manuscript. Yes, it is perfectly true that changes that we observe may well be related to the mitotic status of the cells perhaps commensurate with the renewable nature of each tissue. One might imagine that tissues that continue to turn over during life, such as the skin, will be synthetically more active and this will be reflected in the results. In fact, the total number of proteins incorporating Lys6 (i.e. number of proteins with a measured H/L ratio) across the three mature connective tissues are broadly the same (skin 450, bone 550 and cartilage 599 proteins, also Figure 3) and their range of label incorporation also similar (20-90% in plasma, 10-90% in skin, 2-90% in bone and 2-80% in cartilage), so our results do not appear to be biased by large differences in the global synthetic rates across tissues. Rather, it appears that the difference is in select groups of proteins only.

It is, of course, difficult to be definitive about how important this is when considering complex and simple tissues as the tissue analyses do not specifically indicate which cells are driving the observed phenotype. In the case of cartilage, we can be fairly confident that most (possibly all) cells are post-mitotic chondrocytes. In bone the cells will be a mixture of post mitotic (osteocytes), renewable (blood derived osteoclasts) and proliferating (osteoblasts) cells (although the latter are thought to fail with age). It is perhaps a surprise that post mitotic tissues like cartilage manage to maintain this level of synthesis at all. The post-mitotic tissues have striking effects in select protein groups suggesting that these cells do have a restricted synthetic lifespan. Our thesis is that this possibly explains age-related disease risk.

Action (A) 1. We have inserted a paragraph into the introduction (lines 62-67) and further commented on this in the discussion (lines 539-542). In the limitations section we also mention that we are comparing tissues that have a disparate number and type of cells (line 791). We have also substituted old Figure S2 with a new figure (Figure 3) that shows the incorporation levels between tissues and over time more clearly, and also shows the position of key matrisomal proteins.

2. A more focused analysis of a particular cell type and its Lys6 incorporation rate at different time points or disease states is a viable alternative approach, which would also be highly informative, if not more so for selected questions. In light of the comment above, a more complete discussion of the limitations of this work is needed. It would be worth it to add a justification of why the methods chosen where chosen over other strategies to address the main thesis of the paper.

We agree that understanding which cell types are driving the ageing phenotype across different tissues would be interesting and valuable. However, this type of analysis is severely limited in collagen rich tissues (in particular) as the cell behaviour is greatly influenced by the native tissue environment i.e. extracellular matrix. For example when articular chondrocytes are isolated they rapidly de-differentiate, start to proliferate and are highly synthetic. Therefore, one of the significant advantages of using this approach is that cellular activities can be examined in situ. It might be possible to combine our proteomic analysis with spatial proteomics/transcriptomics or single nuclear RNA sequencing, but this is technically very challenging in murine connective tissues and has not yet been optimised by our group. With regards to disease states, these studies are ongoing for osteoarthritis in our group and will follow, but we feel detract from the main messages of the current paper.

We have incorporated a comment to stress the advantage of being able to analyse synthetic activity in complex tissues without disrupting the cells and matrix in the introduction (line 62 and lines 67-68).

3. Overall, the statistical power is difficult to estimate and the authors should carefully consider the statistical analysis of a non-normal distributed data set. The comparison of young animals with the older stages simply represents growing and dividing tissues, which is naturally more labelled than "mature" tissue.

Our ‘within group’ datasets where we examined the normality of each individual protein against the Shapiro-Wilk normality test and for variance against the Levene test demonstrated that >80% of proteins were normally distributed. Therefore for protein:protein comparisons we applied Student t-tests for pairwise comparisons. We performed group wise comparisons between young and older adults using non-parametric tests.

We agree that the differences observed between skeletally immature and mature groups likely represents growing tissues. These data (presented in Figures 2-6) were intended to demonstrate that the in vivo SILAC methodology was fit for purpose i.e. to demonstrate a proof of concept before going on to look in more detail at the differences in skeletally mature tissues with age. In addition we also feel that these data have intrinsic interest. For instance we were able to define a number of matrisomal proteins that appear to switch off after skeletal maturity e.g. keratocan in all three tissues, as well as those that switch on in adult tissues and are not apparently synthesised during growth e.g. type XV collagen in bone and versican in articular cartilage.

We have added detail to the statistical methods section to clarify the statistical tests used (lines 941-948). We have added in the results that whilst most matrisomal proteins go down after skeletal maturity, some proteins are incorporated for the first time in adult tissues (lines 392-393).

4. Throughout this paper, it is stated that trabecular bone was examined. The methods for collecting this bone are spartan. It appears that the part of the tibia collected is very trabecular bone poor and is mostly cortical bone. The turnover rates of these two bone compartments are very different. Please clarify what kind of bone you had and the implications of looking at only trabecular/ only cortical/ a mix for the interpretations of these results. Considering the tissue of interest appears to be trabecular bone, how did you ensure there was no contaminating marrow. Trabeculae are very small and hard to "clean".

Thank you for pointing out this error. The bone that we collected was mainly cortical bone, not trabecular. We chose this region firstly because we wanted to avoid parts of the bone that were near cartilage/cartilaginous tissue such as subchondral bone and the growth plate, and also because it had clear tissue landmarks, which enabled consistent tissue collection in each animal irrespective of age. The diaphyseal area of tibia from which bone samples were taken has a high cortex-to-trabecular bone ratio. In humans, trabecular bone has a higher turnover rate than cortical bone and it is possible that changes in the ratio of cortical to trabecular bone with age could have influenced our results.

The bone marrow was washed through with PBS injected with a fine needle and syringe and washes were checked to ensure that they were acellular using a stereo microscope. It is correct that we cannot guarantee 100% marrow-free bone.

We have corrected reference to trabecular bone throughout the text and have added a sentence in the limitations section of the discussion to highlight the potential change of cortical to trabecular bone ratio with age that may have influenced the results (lines 731-732). Bone preparation has been clarified in the methods (lines 854-856).

5. This study is only male mice. Bone turnover is higher in females. In humans, females are more likely to get Osteoarthritis than males and wound healing is slower in females than males. The implications of sex on the interpretation of the results and the limitations of this study design need to be included in the discussion.

We agree, which is why we controlled for gender. Time and funding allowing we would have liked to do this in age-matched female mice also. We have added this as a limitation in discussion (737-788).

6. The authors claim that incorporation is lower in aged animals (week 45) compared to young animals (week 15). For example, Table S2 shows an incorporation rate of 2.97 +/- 2.13 at week 45 and 3.38+/-0.98 at week 15 for the cartilage tissue and protein collagen 1a1. The ANOVA calculation showed a significant change for this candidate. Do the authors believe that a fold-change of w45/w15 = 1:1.14 is biologically relevant? The same is true for most of the proteins listed in the table. Is it possible to calculate accurate incorporation rates in % of SILAC ratios of H/L = 0.01? This corresponds to an H/L ratio of 1:100 in the MS spectra. It would be helpful to show a real MS spectrum of such incorporation rates at least in the SI data. This would also show if the "heavy" peak is represented by a true isotopic pattern and whether this peak was selected for fragmentation. Alternatively, it might be that the "low" peak represents only a background signal. The calculation of the coefficient of variation could also help to document the variability of the data set. A more complete comparison of this work to similar previous studies in the field would provide both context for these results and validation of the expected ranges of turnover.

Thank you for the comment. We are not in a position to claim biological, nor indeed, clinical relevance from individual protein results, but think that percentages of new protein incorporation are valuable when interpreted as a trend of incorporation rates across protein groups. Even though we will always be limited by the sensitivity of protein incorporation rates, we can claim some confidence by showing that we were able to measure a wide range of H/L ratios for a diverse range of proteins and time points in all tissues and in a reproducible fashion between replicates.

We include an example spectra of two col1a1 peptides showing distinct peaks in labelled and unlabelled peptide for the 15 and 45 week data. We would be happy to include this as an extra supplementary figure if felt necessary. (Editors note: this has now been added as Figure 4—figure supplement 1).

7. The median lifespan of C57BL/6J male mice is ~128 weeks. The growth plate of mice does not fuse and the femur will increase in length past 26 weeks of age and periosteal expansion is still happening at 12 months (PMID: 13678781). A sound rational for using mice under one year of age is provided. But please be careful when describing these mice. The oldest group are at best mature adult male mice and are certainly not "late adulthood" (line 229-230). The middle group are far from finished growing.

We have now replaced the terms to read “young adult” and “older adult”. We avoided “mature adult” as we felt that this was confusing as we use the term skeletal maturity for both groups C and D.

8. Could it be that older animals move less and eat less? This would also explain a slightly reduced Lys6 incorporation. Did the authors observe any possible weight gain or loss during feeding? Did the animals always take-up the same amount of Lys6 during the experiment?

We recorded animal weight throughout the experiment (see Author response image 1). Mice gained weight in the same way that they do when feeding on standard diet (based on reference growth curves). We cannot exclude that older adult mice moved less as we didn’t formally assess their movement. This might be expected from other ageing studies and could contribute to the changes we are seeing. We didn’t observe any obvious differences in intake between young and older adults although this wasn’t measured formally.

Author response image 1
Download asset Open asset
Weight gain for each of the experimental mice over the period of SILAC heavy diet (3-10 weeks, 11-15 weeks or 42-45 weeks).

These weight curves are consistent with reference growth curves in C57BL/6 mice.

We have added a comment about gaining expected weight in the methods section (lines 842-843).

9. There is a large difference between wound healing and normal skin maintenance/turnover. This this distinction is often blurred when discussing the results for skin and the implications of the results for skin.

We recognise this in our efforts to extrapolate the data to an age-related clinical problem in skin. We have added a comment about not necessarily being able to extrapolate to wounding as this was not a wounding study (lines 725-726).

10. Line 329-332. Mice very much lose trabecular bone without losing estrogen (PMID: 17488199) and naturally occurs without immobilization. Please reframe this section reflecting the reality of you model.

Thank you. We have clarified this in the text. See changes (lines 736-737).

11. It might be better to focus the analysis on a specific tissue/cell type that is either proliferating or post-mitotic. Why did the authors not calculate absolute turnover rates? This would be a better comparison and could help to better assess the statistical power of their method.

One of the primary purposes for doing this project was to ask whether change in synthetic rates alter between different collagen rich tissues and whether this could account for very poor repair responses in select tissues that predispose to age-related disease (e.g. cartilage and osteoarthritis). This question is not the same as asking about individual turnover rates of proteins within tissues (reflecting the balance between synthesis and degradation), which is what the field has largely focused on using other methodologies. In order to address protein turnover rates we would have needed to do a second experimental group after a 3 week washout (similar to comparison of group A and B in Figure 2). We hope we have stressed this now in the substantially revised manuscript.

12. Introduction – for completeness, please briefly also describe the use of deuterium labelling to measure proteome dynamics in tissues:

https://pubmed.ncbi.nlm.nih.gov/22915825/, https://pubmed.ncbi.nlm.nih.gov/32393437/ and https://pubmed.ncbi.nlm.nih.gov/32403418/

This has now been added to introduction (lines 202-207).

13. It is strongly suggested that the authors establishing if there were any alterations in relative protein abundance between age groups assessed. This will allow identification of any proteins that show both dysregulated turnover and abundance with ageing and are therefore most likely involved in age-related diseases.

This is an important consideration. We therefore re-examined raw data and used total iBAQ values to calculate an abundance score for each protein. Interestingly, very few proteins showed regulated abundance levels when considering total protein levels. Articular cartilage was the only tissue where a few proteins were regulated significantly after multiple correction (five out of six of these were also regulated in the labelled analysis). These results most likely reflect the enhanced sensitivity that the SILAC labelling method affords, although clearly this could also be related to other factors that contribute to protein turnover.

We have added a new supplementary figure showing raw data correlation and iBAQ volcano plots and have mentioned these briefly in the results (lines 429-433).

14. Have the authors using Ingenuity Pathway Analysis (or similar software) to identify upstream regulators predicted to result in the differential protein synthesis observed with ageing? This could provide potential targets for age-related diseases that could be investigated further in future studies. The detailed bioinfomatic STRING analysis is certainly helpful but many points raised in the discussion are unfortunately very speculative and should be verified by experimental work. All reviewers were in agreement that a more comprehensive bioinformatics data analysis of some type is mandatory.

We did consider this type of pathway analysis but felt that we would likely be underpowered to draw robust conclusions (these analyses being best validated for transcriptomic datasets that are usually substantially bigger). However, in response to reviewers’ suggestion we have performed pathway enrichment using various proprietary software, including IPA and DAVID. Cluster enrichment using DAVID showed similar results to STRING. Using IPA, several pathways were identified although none reaching statistical significance after correction.

We have included the IPA pathway enrichment analysis as supplementary information (Figure 7—figure supplementary 3). The results are mentioned in results (lines 472-476) and discussion (559-560).

15. Is it possible to calculate and report individual protein half-lives/fractional synthesis rates from the data? This would enable more direct comparison of half-lives measured in previous studies.

Unfortunately, we are unable to calculate individual protein half-lives from single time points. As mentioned above, our focus was more on the synthetic activity of the tissue rather than the turnover of individual proteins, although we accept that having this information would have helped to validate results from similar studies using different methodologies.

16. Materials and methods – different extraction protocols were used for skin, compared to cartilage and bone samples. Please provide a rationale for this. Is this likely to have had an impact of the proteins identified? If so, this needs to be more completely mentioned in the discussion as a limitation of this work.

Different connective tissues often require different protocols to solubilise the tissue due to variable degrees of collagen cross-linking, mineralisation etc. For each tissue we established protocols to solubilise the tissue prior to doing the labelling study. This is why they have different protocols. Articular cartilage solubilisation was, as expected, most challenging, presumed to be due to accumulation of long-lived cross-linked fibrillar collagens. It is likely that this will underestimate the calculated % collagen that is labelled but it is unlikely to account for the % labels of other proteins as these are generally soluble under such conditions.

This limitation is discussed in the limitation section of the discussion (line 728).

17. Results, Ln 160, Figures 3 and 4. It is apparent that incorporation rates decline with ageing across tissues, with most obvious differences between immature and mature animals and much smaller changes between mature and old animals. Are you able to present statistical analysis of these results on the graphs to identify significant differences with ageing?

The statistical significant differences are presented in new figure 4 and 5, with explanation to code in the figure legends. The information can also be found in Figure 4-source data and Figure 5-source data.

18. Although the authors achieved a dilution of the label by a label switch to Ly0, this aspect was unfortunately only very little evaluate. The reviewers felt that this point should be elaborated on.

We accept this criticism. When designing these experiments we did the first part of this paper (groups A and B in skeletally immature animals) as a proof of concept that in vivo SILAC would provide a valid robust measure of protein synthesis in tissues in vivo. We agree that these results deserve more attention, so have (conscious of space restraints) added further discussion around these results.

Further granularity of these results is given in results, lines 339-344.

19. Protein turnover is always regulated by synthesis and degradation. The labelling of living animals with stable isotopes is a useful technique, but one should be very careful which tissues one compares and whether the incorporation rates are regulated by instrinsic protein stability or by cell division.

We take the point that incorporation rates will be in part regulated by intrinsic protein stability (which could change with age), although are not clear how cell division affects incorporation rates except in proteins specifically implicated in replicative activities. Our data clearly show that cartilage and bone continue to synthesise proteins across the healthy lifecourse albeit that the reduction in synthetic activity appears to be more age-sensitive and selective for long-lived proteins such as collagen type II.

We stress that we are only measuring incorporation, not stability of proteins in the final paragraph of the discussion (lines 795-796). [Editors' note: further revisions were suggested prior to acceptance, as described below.]

1. The figure provided for A6 would be of benefit as a supplemental figure. However, in its current form it is not usable as the font size is impossible to read.

We have created a new supplementary figure to Figure 4 which we have spread out over 4 pages to improve clarity, with reference to this in the text lines 287 and 572-3.

2. The largest concern from the reviewers was that the study remains descriptive and the underlying experiments comparing Lys-6 incorporation rates between different tissues or developmental stages remain questionable. While the authors have addressed this to some degree in the discussion (ln 457), further comment on the specific effect of cell division on Lys-6 incorporation would more clearly highlight the limitations of the study. The reviewers would have liked to see an integration of the absolute half-lives as a function of proliferation. Thus, the comparison always remains dependent on cell division, which unfortunately prevents a clear statement of the absolute protein half-life or protein turnover. Therefore, the wording "turnover" should not be used in the manuscript at all, but rather Lys6 incorporation rates (i.e. Figure 2A) in proliferating or postmitotic cells. Lys6 incorporation rates do not necessarily reflect the turnover of a protein and hence this term, while used sparingly, should be avoided overall.

We accept that the term “turnover” is misleading when we have not measured the half-lives of individual proteins, and where we are comparing tissues that have different cell renewal rates which may also change with age. We have therefore gone through the manuscript carefully and have edited “turnover” to 13C6 Lys incorporation or equivalent in each of these places. These changes can be found on lines 34, 37, 165, 192, 485, 590, 594, 871, 1031, and are highlighted in green for ease of reading.

There are two cases where we feel the term “turnover” is still appropriate. Firstly in the introduction, where we refer to other published studies, and secondly in Figure 2 where we broadly categorise high and low turnover protein groups within an individual tissue. As mitotic activity is controlled for within each tissue at baseline and after the washout period, we believe this to be a reasonable interpretation of the data. We reinforce this distinction by adding “high turnover for a given tissue” to line 187.

Conversely, later in Figure 2 where we compare profiles between tissues we have removed the term “turnover” and substituted it for 13C6 Lys incorporation or equivalent. We hope this is acceptable.

We have further tried to clarify this by adding the following to line 315 of the discussion:

“The tissue turnover, which is likely to reflect both protein turnover and cellular renewal, can be divided into tissue modelling….”.

https://doi.org/10.7554/eLife.66635.sa2

Article and author information

Author details

  1. Yoanna Ariosa-Morejon

    Kennedy Institute of Rheumatology, Arthritis Research UK Centre for OA Pathogenesis, University of Oxford, Oxford, United Kingdom
    Contribution
    Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Visualization, Writing - original draft
    Competing interests
    none

  2. Alberto Santos

    1. Big Data Institute, Li-Ka Shing Centre for Health Information and Discovery, University of Oxford, Oxford, United Kingdom
    2. Center for Health Data Science, Faculty of Health Sciences, University of Copenhagen, Copenhagen, United Kingdom
    Contribution
    Data curation, Formal analysis, Software, Validation, Visualization, Writing – review and editing
    Competing interests
    none

  3. Roman Fischer

    Nuffield Department of Medicine, Target Discovery Institute, University of Oxford, Oxford, United Kingdom
    Contribution
    Conceptualization, Formal analysis, Resources, Supervision, Visualization, Writing – review and editing
    Competing interests
    none
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-9715-5951

  4. Simon Davis

    Nuffield Department of Medicine, Target Discovery Institute, University of Oxford, Oxford, United Kingdom
    Contribution
    Formal analysis, Methodology, Writing – review and editing
    Competing interests
    none

  5. Philip Charles

    Nuffield Department of Medicine, Target Discovery Institute, University of Oxford, Oxford, United Kingdom
    Contribution
    Formal analysis, Visualization, Writing – review and editing
    Competing interests
    None
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-5278-5354

  6. Rajesh Thakker

    Academic Endocrine Unit, OCDEM, Churchill Hospital, University of Oxford, Oxford, United Kingdom
    Contribution
    Investigation, Writing – review and editing
    Competing interests
    none
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-1438-3220

  7. Angus KT Wann

    Kennedy Institute of Rheumatology, Arthritis Research UK Centre for OA Pathogenesis, University of Oxford, Oxford, United Kingdom
    Contribution
    Visualization, Writing – review and editing
    Competing interests
    none

  8. Tonia L Vincent

    Kennedy Institute of Rheumatology, Arthritis Research UK Centre for OA Pathogenesis, University of Oxford, Oxford, United Kingdom
    Contribution
    Conceptualization, Funding acquisition, Project administration, Resources, Supervision, Writing – review and editing
    For correspondence
    tonia.vincent@kennedy.ox.ac.uk
    Competing interests
    none
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-3412-5712

Funding

Daphne Jackson Trust

  • Yoanna Ariosa-Morejon

Versus Arthritis (Centre for Osteoarthritis Pathogenesis 20205)

  • Tonia L Vincent

Versus Arthritis (Centre for Osteoarthritis Pathogenesis 21621)

  • Tonia L Vincent

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank H Liao and N Ternette for useful discussions and advice with proteomics data analysis. We also thank G Wilson for helping with figure and table preparations and revisions. YA-M had a fellowship from the Daphne Jackson trust that was co-funded by the Kennedy Trust for Rheumatology Research and University of Oxford. The study was also supported by the Centre for Osteoarthritis Pathogenesis Versus Arthritis (grant nos. 20205 and 21621).

Ethics

Animals' care and experimentation was performed in strict accordance with the Animal (Scientific Procedures) Act 1986 and institutional guidelines of the University of Oxford. (licence number 30-3129).

Senior Editor

  1. Matt Kaeberlein, University of Washington, United States

Reviewing Editor

  1. Cheryl Ackert-Bicknell, University of Colorado, United States

Publication history

  1. Preprint posted: January 13, 2021 (view preprint)
  2. Received: January 26, 2021
  3. Accepted: September 3, 2021
  4. Version of Record published: September 28, 2021 (version 1)

Copyright

© 2021, Ariosa-Morejon et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 766
    Page views
  • 114
    Downloads
  • 9
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Yoanna Ariosa-Morejon
  2. Alberto Santos
  3. Roman Fischer
  4. Simon Davis
  5. Philip Charles
  6. Rajesh Thakker
  7. Angus KT Wann
  8. Tonia L Vincent
(2021)
Age-dependent changes in protein incorporation into collagen-rich tissues of mice by in vivo pulsed SILAC labelling
eLife 10:e66635.
https://doi.org/10.7554/eLife.66635

Categories and tags

Research organism

Further reading

    1. Cell Biology

    Actin-regulated Siglec-1 nanoclustering influences HIV-1 capture and virus-containing compartment formation in dendritic cells

    Enric Gutiérrez-Martínez, Susana Benet Garrabé ... Maria F Garcia-Parajo
    Research Article Updated

    The immunoglobulin-like lectin receptor CD169 (Siglec-1) mediates the capture of HIV-1 by activated dendritic cells (DCs) through binding to sialylated ligands. These interactions result in a more efficient virus capture as compared to resting DCs, although the underlying mechanisms are poorly understood. Using a combination of super-resolution microscopy, single-particle tracking and biochemical perturbations we studied the nanoscale organization of Siglec-1 on activated DCs and its impact on viral capture and its trafficking to a single viral-containing compartment. We found that activation of DCs leads to Siglec-1 basal nanoclustering at specific plasma membrane regions where receptor diffusion is constrained by Rho-ROCK activation and formin-dependent actin polymerization. Using liposomes with varying ganglioside concentrations, we further demonstrate that Siglec-1 nanoclustering enhances the receptor avidity to limiting concentrations of gangliosides carrying sialic ligands. Binding to either HIV-1 particles or ganglioside-bearing liposomes lead to enhanced Siglec-1 nanoclustering and global actin rearrangements characterized by a drop in RhoA activity, facilitating the final accumulation of viral particles in a single sac-like compartment. Overall, our work provides new insights on the role of the actin machinery of activated DCs in regulating the formation of basal Siglec-1 nanoclustering, being decisive for the capture and actin-dependent trafficking of HIV-1 into the virus-containing compartment.

    1. Cell Biology

    Autofluorescence imaging permits label free cell type assignment in unmodified airway and reveals the dynamic formation of airway secretory cell associated antigen passages (SAPs)

    Viral S Shah, Jue Hou ... Jayaraj Rajagopal
    Tools and Resources

    The specific functional properties of a tissue are distributed amongst its component cell types. The various cells act coherently, as an ensemble, in order to execute a physiologic response. Modern approaches for identifying and dissecting novel physiologic mechanisms would benefit from an ability to identify specific cell types in live tissues that could then be imaged in real-time. Current techniques require the use of fluorescent genetic reporters that are not only cumbersome, but which only allow the simultaneous study of 3 or 4 cell types at a time. We report a non-invasive imaging modality that capitalizes on the endogenous autofluorescence signatures of the metabolic cofactors NAD(P)H and FAD. By marrying morphological characteristics with autofluorescence signatures, all seven of the airway epithelial cell types can be distinguished simultaneously in mouse tracheal explant in real-time. Furthermore, we find that this methodology for direct cell type specific identification avoids pitfalls associated with the use of ostensibly cell type-specific markers that are, in fact, altered by clinically relevant physiologic stimuli. Finally, we utilize this methodology to interrogate real-time physiology and identify dynamic secretory cell associated antigen passages (SAPs) that form in response to cholinergic stimulus. The identical process has been well documented in the intestine where the dynamic formation of secretory and goblet cell associated antigen passages (SAPs and GAPs) enable luminal antigen sampling. Given that airway secretory cells can be stimulated to make mucous within hours, we suspect that both SAPs and GAPs are also used for luminal antigen sampling in the airway. This hypothesis is supported by our observation that secretory cells with airway SAPs are frequently juxtaposed to antigen presenting cells.

    1. Cell Biology

    Nuclear SUN1 stabilizes endothelial cell junctions via microtubules to regulate blood vessel formation

    Danielle B Buglak, Pauline Bougaran ... Victoria L Bautch
    Research Article

    Endothelial cells line all blood vessels, where they coordinate blood vessel formation and the blood-tissue barrier via regulation of cell-cell junctions. The nucleus also regulates endothelial cell behaviors, but it is unclear how the nucleus contributes to endothelial cell activities at the cell periphery. Here, we show that the nuclear-localized linker of the nucleoskeleton and cytoskeleton (LINC) complex protein SUN1 regulates vascular sprouting and endothelial cell-cell junction morphology and function. Loss of murine endothelial Sun1 impaired blood vessel formation and destabilized junctions, angiogenic sprouts formed but retracted in SUN1-depleted sprouts, and zebrafish vessels lacking Sun1b had aberrant junctions and defective cell-cell connections. At the cellular level, SUN1 stabilized endothelial cell-cell junctions, promoted junction function, and regulated contractility. Mechanistically, SUN1 depletion altered cell behaviors via the cytoskeleton without changing transcriptional profiles. Reduced peripheral microtubule density, fewer junction contacts, and increased catastrophes accompanied SUN1 loss, and microtubule depolymerization phenocopied effects on junctions. Depletion of GEF-H1, a microtubule-regulated Rho activator, or the LINC complex protein nesprin-1 rescued defective junctions of SUN1-depleted endothelial cells. Thus, endothelial SUN1 regulates peripheral cell-cell junctions from the nucleus via LINC complex-based microtubule interactions that affect peripheral microtubule dynamics and Rho-regulated contractility, and this long-range regulation is important for proper blood vessel sprouting and junction integrity.