HDAC inhibition improves autophagic and lysosomal function to prevent loss of subcutaneous fat in a mouse model of Cockayne syndrome

See allHide authors and affiliations

Science Translational Medicine  29 Aug 2018:
Vol. 10, Issue 456, eaam7510
DOI: 10.1126/scitranslmed.aam7510

Autophagy and adipose

Hypersensitivity to sunlight and decreased subcutaneous fat are characteristics of Cockayne syndrome, a genetic disorder of premature aging. Majora and colleagues studied the role of autophagy underlying this skin phenotype using mice with a mutation in csb, a gene that causes Cockayne syndrome. They found an accumulation of autophagy-related proteins and lysosomal dysfunction in skin from csb mutant mice exposed to ultraviolet light. Treatment with a histone deacetylase inhibitor prevented loss of subcutaneous fat and rescued autophagic/lysosomal dysfunction. Further research is necessary to determine whether histone deacetylase inhibition can resolve other features of Cockayne syndrome, such as neurodegeneration.


Cockayne syndrome (CS), a hereditary form of premature aging predominantly caused by mutations in the csb gene, affects multiple organs including skin where it manifests with hypersensitivity toward ultraviolet (UV) radiation and loss of subcutaneous fat. There is no curative treatment for CS, and its pathogenesis is only partially understood. Originally considered for its role in DNA repair, Cockayne syndrome group B (CSB) protein most likely serves additional functions. Using CSB-deficient human fibroblasts, Caenorhabditis elegans, and mice, we show that CSB promotes acetylation of α-tubulin and thereby regulates autophagy. At the organ level, chronic exposure of csbm/m mice to UVA radiation caused a severe skin phenotype with loss of subcutaneous fat, inflammation, and fibrosis. These changes in skin tissue were associated with an accumulation of autophagic/lysosomal proteins and reduced amounts of acetylated α-tubulin. At the cellular level, we found that CSB directly interacts with the histone deacetylase 6 (HDAC6) and the α-tubulin acetyltransferase MEC-17. Upon UVA irradiation, CSB is recruited to the centrosome where it colocalizes with dynein and HDAC6. Administration of the pan-HDAC inhibitor SAHA (suberoylanilide hydroxamic acid) enhanced α-tubulin acetylation, improved autophagic function in CSB-deficient models from all three species, and rescued the skin phenotype in csbm/m mice. HDAC inhibition may thus represent a therapeutic option for CS.


Cockayne syndrome (CS) is a rare, hereditary multiorgan disease, which is predominantly caused by mutations in the csb gene. CS patients show a broad spectrum of pathological abnormalities such as postnatal developmental failure, neurodegeneration, and premature aging. Skin symptoms include ultraviolet (UV) hypersensitivity and a progressive loss of subcutaneous fat (1). There is currently no effective treatment for CS patients. Traditionally, CS is regarded as a DNA repair syndrome with a defect in transcription-coupled repair (impaired removal of UV-induced DNA lesions from actively transcribed DNA strands) (2). In addition, Cockayne syndrome group B (CSB) protein participates in the repair of various forms of oxidative DNA lesions by interacting with components of the base excision repair machinery (3, 4). Recent studies, however, indicate that CSB protein may have functions outside of the cell nucleus. Accordingly, it has been shown that CSB is present in mitochondria (4, 5) and that mitochondrial autophagy is impaired in vitro in CSB-deficient human fibroblasts (CSB-HF) (6).

Autophagy is an adaptive mechanism initiated by the cell in response to various forms of stress including UV radiation and is important for maintenance of cellular homeostasis. Currently, it is not known whether compromised autophagy is present and of pathogenic relevance in CSB-deficient organisms and, if relevant, how CSB promotes autophagy. Disturbed autophagy causes a variety of diseases (7), and regulation of autophagy has been intensively studied (8, 9). In the context of the present study, it is important that protein acetylation appears to be of particular relevance (10). Several histone acetyltransferases (HATs) and histone deacetylases (HDACs) are involved in the control of autophagy. Among these, HDAC6 promotes autophagy by interacting with microtubule proteins. By binding to polyubiquitinated proteins and dynein motor proteins, HDAC6 directs the retrograde transport of misfolded proteins toward the perinuclear microtubule-organizing center (MTOC) and thereby supports their aggresomal sequestration and subsequent clearance by autophagy (11, 12). HDAC6 also deacetylates α-tubulin at lysine 40, whereas MEC-17 acetylates this site (13, 14). Acetyl–α-tubulin facilitates the fusion of autophagosomes with lysosomes and promotes motor protein binding and motility to accelerate cargo transport along the tubulin network (1517).

Circumstantial evidence suggests that CSB might affect protein acetylation. A comparative microarray expression analysis revealed that many of the genes regulated by CSB can also be modulated by HDAC inhibitors (18), and impaired acetylation of histone H4 has been detected at the promoters of UV- and hypoxia-responsive genes in CSB-HF (19, 20).

Here, we provide evidence that dysfunctional autophagy due to CSB deficiency is of pathological relevance and may be due to insufficient α-tubulin acetylation. CSB might affect α-tubulin acetylation by interacting with both HDAC6 and MEC-17. Consequently, pharmacological HDAC inhibition improved autophagic function and prevented loss of subcutaneous fat in CSB-deficient mice, indicating a potential therapeutic option for treatment of CS patients.


UVA radiation enhances loss of subcutaneous fat in csbm/m mice

To learn more about the role of the CSB protein in autophagy, we used a csbm/m hairless mouse model in which the skin phenotype, that is, degeneration of subcutaneous fat, can be aggravated by chronic UVA exposure. We predicted that chronic irradiation would cause a considerable stress response, thereby increasing autophagy to maintain cellular homeostasis. Specifically, we used a UVA1 (340 to 400 nm) radiation source, which is relatively less effective in causing the formation of potentially cytotoxic and mutagenic DNA lesions like cyclobutane pyrimidine dimers, but highly effective in generating oxidative stress and inducing autophagy (21, 22).

Histological analysis of skin sections from unirradiated mice revealed partial degeneration of subcutaneous fat (Fig. 1, A to C), confirming our previous observation (4). Exposure of mice to low or high cumulative doses of UVA radiation dose-dependently enhanced loss of subcutaneous fat (Fig. 1, A to D). Degenerating adipocytes frequently contained large deposits of lipofuscin/ceroid pigments, suggesting impaired protein degradation (Fig. 1B, higher magnification). In addition to loss of subcutaneous fat, UVA-irradiated csbm/m mice developed skin fibrosis, as indicated by hematoxylin and eosin (H&E) and Picrosirius red staining (Fig. 1, A and B, and fig. S1), with accumulation of dense collagen fibers in the dermis and the subcutaneous layer of the skin. These fibrotic alterations were reflected by concomitant changes in gene expression. Quantitative real-time polymerase chain reaction (qRT-PCR) analysis indicated increased expression of fibrosis-associated genes including collagen types 1α1 and 1α2 in irradiated csbm/m mice (Fig. 1E). In the epidermis, histological analysis showed hyperplasia and hyperkeratosis in irradiated csbm/m mice [Fig. 1, B (higher magnification) and F]. In addition, skin physiological assessments revealed elevated transepidermal water loss (TEWL), decreased skin moisture, and increased sebum production (Fig. 1, G to I). Thus, chronic UVA irradiation of csbm/m mice caused a skin phenotype, which affected the subcutaneous fat, the dermis, and the epidermis.

Fig. 1 UVA radiation causes enhanced loss of subcutaneous fat in csbm/m mice.

(A and B) Macroscopic appearance of UVA-irradiated skin of csbm/m and wild-type (WT) mice evaluated by H&E staining after low-dose (A) or high-dose UVA (B). Representative pictures are shown (n = 7). Scale bars, 100 (A) and 200 μm (B). The boxed regions in (B) are shown in higher magnification. (C and D) Relative area of subcutaneous fat (SF) determined from skin sections after low-dose (C) or high-dose (D) UVA (means ± SEM; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, one-way analysis of variance (ANOVA) with Tukey’s multiple comparisons tests; n = 6 to 7). (E) qRT-PCR analysis of Col1α1, Col1α2, Col3α1, FN-1, Timp-1, and MMP-3 mRNA expression in the skin of csbm/m and WT mice after low-dose UVA. Expression was normalized to 18S ribosomal RNA (rRNA) expression (means ± SEM; **P < 0.01, ***P < 0.001, ****P < 0.0001, two-way ANOVA with Tukey’s multiple comparisons tests; n = 6). (F) Epidermal thickness of csbm/m and WT mouse skin (means ± SEM; ****P < 0.0001, one-way ANOVA with Tukey’s multiple comparisons tests; n = 7). (G) TEWL measured on the back skin of csbm/m and WT mice after low-or high-dose UVA (means ± SEM; *P < 0.05, ***P < 0.001, ****P < 0.0001, one-way ANOVA with Tukey’s multiple comparisons tests; n = 7 low dose and n = 5 high dose). (H) Skin humidity of csbm/m and WT mice after low-dose UVA, measured with a corneometer (means ± SEM; **P < 0.01, ***P < 0.001, two-way ANOVA with Tukey’s multiple comparisons tests; n = 7). a.u., arbitrary units. (I) Sebum content of the skin of csbm/m and WT mice after low-dose UVA was measured using a sebumeter (means ± SEM; *P < 0.05, **P < 0.01, two-way ANOVA with Tukey’s multiple comparisons tests; n = 7).

The skin phenotype of csbm/m is associated with an accumulation of autophagy-related proteins

We next asked whether the development of this skin phenotype was associated with signs of defective autophagy. We noted that UVA-irradiated csbm/m mice developed a yellowish skin color, which was visible with the naked eye and was confirmed by chromametry (Fig. 2A). This change in skin color might result from an accumulation of nondegraded material in the skin due to impaired proteolysis. csbm/m mice showed increased expression of ubiquitinated proteins (Fig. 2B), which were most prominent at the sites of subcutaneous fat degeneration in UVA-irradiated animals (Fig. 2C). We also detected high amounts of 4-hydroxynonenal (4-HNE)–modified proteins indicating oxidative damage in the subcutaneous tissue of UVA-irradiated animals (Fig. 2D). To assess whether the observed phenotype is linked to dysfunctional autophagy, we next analyzed murine skin sections for the autophagy markers LC3B and p62. LC3B is an integral component of the autophagosomal membrane after its truncation, whereas p62 is a ubiquitin-binding adapter protein that is thought to shuttle target proteins destined for degradation to the autophagosomes (23). Because p62 is also a substrate for autophagic degradation, its accumulation is regarded as a marker for autophagic dysfunction (24). In the subcutaneous tissue of UVA-irradiated csbm/m mice (and, to a lower extent, in unirradiated animals), we detected an accumulation of LC3B and p62 (Fig. 2, E and F), thus corroborating and extending previous findings (6).

Fig. 2 Ubiquitinated and autophagy-related proteins accumulate in the skin of UVA-irradiated csbm/m mice.

(A) Skin color (yellow/blue) of csbm/m and WT mice after low-dose UVA measured with a chromameter (means ± SEM; *P < 0.05, two-way ANOVA with Tukey’s multiple comparisons tests; n = 7). (B) Western blot analysis of ubiquitinated proteins in whole-skin lysates of csbm/m and WT mice after low-dose UVA. β-Actin was used as a loading control. (C to F and H and I) Immunohistochemistry of ubiquitin (C), 4-HNE (D), LC3B (E), p62 (F), IL-1β (H), or TGF-β1 (I) in the subcutis of skin sections from csbm/m and WT mice after low-dose UVA radiation. Representative images are shown (n = 3). Scale bars, 50 μm. (G) qRT-PCR analysis of IL-1β, TNF-α, KC, TGF1, COX2, and HMOX-1 mRNA expression in the skin of csbm/m and WT mice after low-dose UVA. Expression was normalized to 18S expression (means ± SEM; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, two-way ANOVA with Tukey’s multiple comparisons tests; n = 6).

Impaired autophagy is implicated in the onset of inflammatory responses and various forms of pathological conditions including fibrosis (25). Dysfunctional autophagy has been shown to trigger the expression of inflammatory cytokines including tumor necrosis factor–α (TNF-α) and interleukin-1β (IL-1β) in a p62-dependent manner (26, 27). Pharmacological inhibition of autophagy in human and murine adipose tissue explants resulted in increased expression of IL-1β, IL-8, and keratinocyte chemoattractant (KC) (28). Autophagy blockade by deletion of LC3B is reported to result in collagen deposition and increased amounts of the profibrotic factor transforming growth factor–β1 (TGF-β1) in murine kidneys after unilateral ureteral obstruction (29). Likewise, we observed that impaired autophagy in whole-skin samples prepared from UVA-irradiated csbm/m mouse skin was associated with an up-regulation of mRNA levels for IL-1β, TNF-α, KC, COX2, and TGF1 (Fig. 2G). Heme oxygenase 1 (HMOX-1), which is known to be involved in the regulation of inflammatory processes and induced by 4-HNE, was also highly expressed in csbm/m skin. Additional immunohistological analysis of skin sections showed that in the degenerating subcutaneous tissue of UVA-irradiated csbm/m mice, IL-1β and TGF-β1 were also increased at the protein level (Fig. 2, H and I). These results suggest that impaired autophagy in UVA-irradiated csbm/m mice triggered an inflammatory and profibrotic response, which led to the observed loss of subcutaneous fat and fibrosis.

Autophagic/lysosomal function is impaired in CSB-HF and csbm/m mouse skin

We next asked whether autophagy would also be dysfunctional in irradiated CSB-HF. Primary human CS1AN fibroblasts were used because these cells express a truncated CSB protein similar to that of the csbm/m mouse strain used in the in vivo experiments. CSB-HF and control fibroblasts from healthy donors [normal human fibroblasts (NHFs)] were irradiated with a single dose of UVA, and autophagy was evaluated by Western blot analysis. Increased accumulation of autophagosomal membrane-bound LC3BII (the lower band) and p62 protein aggregates was already detectable in unirradiated CSB-HF as compared to NHF, and this difference was further augmented by UVA irradiation (Fig. 3, A, B, D, and E). The applied UVA dose did not affect cell viability (fig. S2). Moreover, we observed an increase in lysosomal proteins lysosomal-associated membrane protein 1 (LAMP1) and LAMP2, which was independent of UVA exposure (Fig. 3A). Increased uptake of LysoTracker dye in CSB-HF, which was further enhanced by UVA, indicated increased lysosomal mass in these cells (Fig. 3C). In subsequent immunofluorescence analysis of LAMP2 expression, we observed an accumulation of dilated lysosomes, which formed extended perinuclear aggregates in CSB-HF (Fig. 3D). Costaining of p62 and LC3B (fig. S3) confirmed the presence of large numbers of p62 aggregates and autophagosomes in CSB-HF, which often localized in the periphery of lysosomes.

Fig. 3 Autophagy is inhibited in human CSB-HF.

(A) Western blot analysis of LC3B, p62, LAMP1, and LAMP2 in NHF and CS1AN cells 20 hours after UVA irradiation. β-Actin was used as a loading control (n = 3). (B) p62 protein amount normalized to β-actin quantified by densitometry (means ± SEM; *P < 0.05, two-way ANOVA with Tukey’s multiple comparisons tests; n = 4). (C) NHF and CS1AN cells irradiated with UVA, stained with LysoTracker 20 hours after UVA irradiation, and analyzed by flow cytometry (n = 3). (D) Immunofluorescence of LAMP2 and p62 in NHF and CS1AN cells 20 hours after UVA (n = 3). Scale bar, 50 μM. (E) Intracellular p62 aggregates in NHF and CS1AN fibroblasts (means ± SD; ****P < 0.001, one-way ANOVA with Tukey’s multiple comparisons tests; n = 46 to 71). (F) Electron microscopic analysis of NHF and CS1AN cells 20 hours after UVA irradiation. (I) NHF, arrows point at lysosomes/autolysosomes. (II, III, and IV) CS1AN cells; arrows point at autophagosome structures adjacent to lysosomes (III) or lysosomes with leaky membranes (IV). Scale bar, 500 nm (n = 2). (G) NHF and CS1AN cells irradiated with UVA, stained with LysoSensor 4 hours later, and analyzed by flow cytometry (n = 3). (H) Western blot analysis of cathepsin B and pan-cathepsin in NHF and CS1AN cells at various time points after UVA irradiation. β-Actin was used as a loading control (n = 3). DAPI, 4′,6-diamidino-2-phenylindole; FCS, forward scatter; SSC, side scatter.

We next assessed lysosomes at the ultrastructural level by electron microscopy. In UVA-irradiated NHF, only a few small lysosomes and autolysosomes were detectable. In contrast, CSB-HF contained large quantities of abnormally dilated and irregularly shaped lysosomes, which often fused with each other and sometimes were marked by multilamellar bodies (Fig. 3F). Numerous autophagosomes or autophagosome-like structures were located distant from lysosomes, thus confirming our fluorescence microscopy results. We also detected lysosomes with ruptured membranes in CSB-HF. Because function of lysosomal enzymes depends on a low pH, we next evaluated lysosomal acidification in CSB-HF by analysis of LysoSensor dye staining. UVA induced lysosomal acidification in both NHF and CSB-HF, but staining was generally more intense in CSB-HF, suggesting a lower lysosomal pH in these cells (Fig. 3G).

Accumulation of LC3BII, p62, and LAMP1 has previously been observed in NHF that had been either UVA-irradiated or treated with an inhibitor of cathepsin B, a known target for UVA radiation (30). Here, Western blot analysis showed increased amounts of procathepsin B protein, its truncated active form, and enhanced staining using a pan-cathepsin antibody in unirradiated and in UVA-irradiated CSB-HF (Fig. 3H).

Lysosomal parameters were also altered in the skin of csbm/m mice. Expression of cathepsins and other lysosomal genes was higher in the irradiated skin of csbm/m mice, as compared with irradiated wild-type mice (fig. S4A). Similarly, immunohistochemical analysis revealed increased cathepsin B, cathepsin D, pan-cathepsin, and LAMP2 staining in the subcutis of irradiated csbm/m mice (fig. S4, B to E). Increased cathepsin protein amounts in csbm/m skin were verified by Western blot using antibodies detecting cathepsin B and pan-cathepsin (fig. S4F). These observations indicate that autophagy is impaired in CSB-HF and in the skin of UVA-irradiated csbm/m mice and that this impairment is associated with morphological, structural, and functional alterations of lysosomes, which are reminiscent of lysosomal storage diseases (LSDs).

CSB locates to the centrosome and regulates acetylation of α-tubulin

Niemann-Pick type C (NP-C), an LSD, is characterized by lysosomal accumulation of cholesterol, autophagic dysfunction, and a heterogeneous spectrum of symptoms primarily affecting the brain and the liver (31). In NP-C fibroblasts, lysosomal dysfunction was linked to increased expression and function of HDACs (32). We analyzed the skin of csbm/m mice for HDAC proteins and saw higher amounts of various HDAC proteins including HDAC6, which is critically involved in autophagy regulation (Fig. 4, A and B), in UVA-irradiated csbm/m skin compared to wild-type. Histological analysis of skin sections for HDAC1 and HDAC6 revealed that these proteins were particularly enriched in the degenerated subcutaneous tissue of UV-irradiated csbm/m mice (Fig. 4, C and D). At the mRNA level, expression of most HDAC genes (including HDAC1 and HDAC6) was also increased in csbm/m mouse skin (fig. S5).

Fig. 4 CSB interacts with HDAC6 and MEC-17, locates to the centrosome, and regulates acetylation of α-tubulin.

(A) HDAC protein amount in csbm/m and WT mouse skin were analyzed by Western blot (WB). (B) Amount of HDAC6 protein in mouse skin was quantified densitometrically (n = 2). (C and D) Immunohistochemistry of HDAC1 and HDAC6 in the subcutis of csbm/m and WT mice (n = 3). Scale bars, 50 μm. (E) WB analysis of acetyl–α-tubulin in the skin of csbm/m and WT mice. (F) WB analysis of acetyl–α-tubulin, HDAC6, and MEC-17 in NHF and CS1AN fibroblasts (n = 3). (G) Amount of MEC-17 protein in NHF and CS1AN fibroblasts was quantified densitometrically (n = 3). (H) Immunofluorescence of MEC-17 and acetyl–α-tubulin in NHF and CS1AN cells 6 hours after UVA (n = 3). Scale bar, 50 μM. (I) WB analysis of acetyl–α-tubulin in C. elegans lysates from csb-1 and control small interfering RNA (siRNA)–treated animals (n = 3). (J) Immunofluorescence analysis of UVA-irradiated NHF and CS1AN fibroblasts stained for CSB in combination with dynein, acetyl–α-tubulin, or HDAC6 (n = 3). Scale bar, 10 μm. The centrosome is shown in higher magnification in the lower right of each picture. (K) Proximity ligation assay for CSB/HDAC6 (n = 3) and CSB/MEC-17 (n = 2) NHF and CS1AN fibroblasts. Scale bar, 50 μM. Nuclear spots were quantified (means ± SD; **P < 0.01, ****P < 0.0001, one-way ANOVA with Tukey’s multiple comparisons tests; n = 193 to 333 nuclei for CSB/HDAC6 and n = 129 to 277 nuclei for CSB/MEC-17). (L) Immunofluorescence analysis of UVA-irradiated NHF and CS1AN fibroblasts stained for CSB in combination with MEC-17 (n = 3). Scale bar, 20 μm. The centrosome and a cytoplasmic region are shown in higher magnification. (M) Co-immunoprecipitation (IP) of CSB protein and MEC-17 in CS1AN fibroblasts transfected with a plasmid expressing HA-tagged WT CSB (n = 3). (N and O) WB analysis of acetyl-H4K16 in NHF and CSB-HFs (N) and C. elegans lysates from csb-1 and control siRNA–treated animals (O) (n = 3). IgG, immunoglobulin G.

A major target for HDAC6-catalyzed deacetylation is α-tubulin. Increased acetylation of this cytoskeletal protein at K40 is required for autophagy stimulation after nutrient deprivation and facilitates the fusion of autophagosomes with lysosomes (15, 33). Therefore, we next analyzed the acetylation status of α-tubulin at K40. Although unirradiated wild-type and csbm/m animals displayed similar amounts of acetyl–α-tubulin, a decrease was noted in the skin of UVA-irradiated csbm/m mice compared to wild-type littermates (Fig. 4E). In vitro, UVA radiation time-dependently increased α-tubulin acetylation in NHF within 6 hours, which correlated with the elevated expression of the tubulin acetyltransferase MEC-17 (Fig. 4, F and G); HDAC6 expression was not altered (Fig. 4F). Although protein content of MEC-17 and HDAC6 was comparable to NHF, acetyl–α-tubulin amounts in CSB-HF were markedly decreased, suggesting an imbalance of acetylation versus deacetylation activity in the absence of functional CSB. In line with that, MEC-17 accumulated in the nucleus of CS1AN fibroblasts (Fig. 4H). Senescent NHF did not show decreased α-tubulin acetylation, indicating that it is not a feature of cellular senescence (fig. S6). We next analyzed acetylation of α-tubulin in Caenorhabditis elegans lysates. CSB deficiency, caused by csb-1 knockdown, decreased acetyl–α-tubulin levels in this model organism (Fig. 4I).

To better understand how CSB could affect acetylation of α-tubulin, we next costained UVA-irradiated and unirradiated NHF and CSB-HF for CSB and dynein, acetyl–α-tubulin, or HDAC6. Dynein is accumulated at the centrosome of NHF and CSB-HF where it formed perinuclear spots (Fig. 4J). As expected, CSB protein, which could also be detected in its mutant form in CS1AN fibroblasts, was mainly located in the nucleus. We also note that, however, a distinct signal at the centrosome which partially colocalized with dynein, and which was more intense in irradiated cells, indicating UV stress-dependent recruitment of CSB to the centrosome. CSB also partially colocalized with acetyl–α-tubulin and HDAC6 at the centrosome (Fig. 4J). Staining of the centrosome for CSB was also seen with an alternative anti-CSB antibody (fig. S7).

To test whether CSB directly interacts with HDAC6, we performed proximity ligation assays after confirming that CSB interacts with HDAC1 (fig. S8A) (34). As is shown in Fig. 4K, wild-type and mutant CSB interact with HDAC6 in UVA-irradiated and in unirradiated fibroblasts. This interaction predominantly occurred in the nucleus and was more frequently detected for the mutant CSB protein. Moreover, CSB also interacted with MEC-17 (Fig. 4K). In unirradiated NHF, the signal was mostly found in the nucleus, whereas it shifted to the cytoplasm after UVA irradiation. In contrast to HDAC6, mutant CSB interacted less frequently with MEC-17 in the nucleus compared to the wild-type protein. These data demonstrate that impaired acetylation of α-tubulin in CS1AN cells correlates with an altered interaction between mutant CSB and HDAC6/MEC-17.

Immunofluorescence analysis additionally revealed that MEC-17 was also present at the centrosome where it colocalized with CSB in NHF and CS1AN fibroblasts (Fig. 4L). Colocalization of both proteins was also detected at perinuclear and peripheral cytoplasmic sites within the cells (Fig. 4L, magnification). To confirm the interaction between CSB and MEC-17, we immunoprecipitated CSB from immortalized CS1AN fibroblasts, which had been transfected with a vector expressing hemagglutinin (HA)–tagged wild-type CSB. Western blot detection of MEC-17 in the immunoprecipitated samples from UVA-irradiated and unirradiated cells indicated the presence of a double band corresponding to the size of MEC-17 (Fig. 4M). These results indicate that CSB and MEC-17 can interact with each other. A modulation of this interaction by UVA, however, is only suggested by the proximity ligation assay data. It should be noted that UVA increases the abundance of both proteins, making a quantitative interpretation of this interaction difficult. Interaction with HDAC6 was not only specific for CSB but also detectable for other DNA repair enzymes including CSA, XPA, and XPC. Also, decreased acetylation of α-tubulin was detected in human fibroblasts deficient for these nucleotide excision repair (NER) proteins (fig. S8, A to F).

An important nuclear target for HDAC1 is histone H4K16, which is linked to cellular life span in general and chromatin remodeling, transcription, autophagy, and DNA repair in particular. We therefore next determined the amount of acetyl-H4K16 in histones isolated from NHF, CS1AN, and CSB-deficient fibroblasts from a second CS patient. UVA induced a time-dependent increase in H4K16 acetylation (Fig. 4N). Constitutive and UVA-induced amounts of acetylated H4K16 were decreased in CSB-HF as compared to NHF (Fig. 4N). This finding was confirmed in csb-1–deficient C. elegans (Fig. 4O), suggesting that impairment of protein acetylation due to CSB deficiency is not limited to α-tubulin.

HDAC inhibition enhances α-tubulin acetylation and improves autophagic/lysosomal function in CSB-HF and C. elegans

Next, we assessed whether dysfunctional autophagy and decreased acetylation of α-tubulin are interrelated and whether pharmacological inhibition of HDACs can improve autophagic/lysosomal functions in CSB-deficient models. Lysosomal function could be improved in human NP-C fibroblasts upon treatment with the HDAC inhibitor suberoylanilide hydroxamic acid (SAHA), as seen by reduction of lysosomal filipin and sphingolipid accumulation (32). Western blot analysis showed that SAHA treatment efficiently diminished accumulation of p62 in unirradiated and UVA-irradiated CSB-HF (Fig. 5A), indicating improvement of autophagy by enhancement of protein acetylation. This effect depended on lysosomes as bafilomycin A1 (BAF), which inhibits lysosomal function, abolished SAHA-mediated clearance of p62. LC3BII protein amounts were not reduced, indicating that SAHA affected autophagic flux rather than formation of autophagosomes. Treatment of cells with SAHA also increased amounts of lysosomal proteins LAMP1, LAMP2, and cathepsin B in irradiated and unirradiated CSB-HF. Enhancement of the active cathepsin form was again abolished by BAF. SAHA also enhanced cathepsin B, K, and L activities and further increased lysosomal mass in unirradiated and irradiated CSB-HF (Fig. 5B).

Fig. 5 SAHA improves impaired protein acetylation, autophagic flux, and lysosomal function in human CSB-HF and C. elegans.

(A) WB analysis of autophagy- and lysosome-related markers in NHF and CS1AN cells exposed to UVA with or without SAHA or BAF (n = 3). (B) Cathepsin activity and LysoTracker fluorescence in NHF and CSB-HF with or without SAHA and UVA exposure (n = 3). (C) p62 puncta were quantified in p62-GFP–expressing csb-1 knockdown C. elegans with or without SAHA (box and whiskers + minimum to maximum; ****P < 0.0001, one-way ANOVA with Tukey’s multiple comparisons tests; n = 3). (D) Fluorescence intensity of csb-1 knockdown C. elegans treated with Nile red with or without SAHA (box and whiskers + minimum to maximum; ****P < 0.0001, one-way ANOVA with Tukey’s multiple comparisons tests; n = 3). (E) WB analysis of acetyl–α-tubulin in csb-1 knockdown C. elegans with or without SAHA (n = 3). (F) WB analysis of acetyl–α-tubulin in the skin of csbm/m mice treated with or without SAHA. (G) Immunofluorescence analysis of p62 and acetyl–α-tubulin in NHF and CS1AN cells either treated with SAHA or tubastatin A (TUBA) (n = 3). Scale bar, 10 μm. (H and I) WB analysis of p62 in CS1AN fibroblasts treated with SAHA, tubastatin A, vinblastine (H), or ciliobrevin D (I) (n = 5). (J and K) WB analysis of acetyl-H4K16 in CS1AN fibroblasts (J) or csb-1 knockdown C. elegans (K) with or without SAHA (n = 3). DMSO, dimethyl sulfoxide; RNAi, RNA interference; GFP, green fluorescent protein.

Similar to csbm/m mice and CSB-HF, csb-1 knockdown in C. elegans showed an increased accumulation of p62 aggregates (Fig. 5C) and accumulated more Nile red as compared to control animals (Fig. 5D), suggesting increased lipid storage in lysosome-related organelles. SAHA treatment reduced p62 aggregates and Nile red fluorescence in CSB-deficient worms to amounts observed in wild-type animals (Fig. 5, C and D); this indicates improvement of autophagic/lysosomal function by pharmacological HDAC inhibition.

We next tested whether SAHA treatment would enhance acetylation of α-tubulin in our models. Increased acetyl–α-tubulin was detected in csb-1 knockdown C. elegans treated with SAHA (Fig. 5E), and in skin lysates obtained from UVA-irradiated and unirradiated csbm/m mice treated with SAHA-containing drinking water (Fig. 5F). To assess whether increased α-tubulin acetylation was involved in SAHA-induced improvement of autophagy, we treated CSB-HF with SAHA or the HDAC6 inhibitor tubastatin A and UVA-irradiated them. Immunofluorescence analysis confirmed impaired α-tubulin acetylation in solvent-treated CSB-HF, which was associated with accumulation of p62 aggregates distributed throughout the cell (Fig. 5G). In contrast, α-tubulin acetylation in NHF was more intense and predominantly located to the perinuclear region in proximity to the p62 aggregates, which were also reduced in number. Treatment with the pan-HDAC inhibitor SAHA or the HDAC6 inhibitor tubastatin A not only improved α-tubulin acetylation in CSB-HF but also stimulated the clearance of p62 aggregates. Remaining p62 bodies were often sequestered into perinuclear aggresome–like structures, which are known to be targets for autophagy-mediated degradation (35).

Degradation through this pathway requires dynein-dependent transport of the cargo via the microtubule network to the centrosome where aggresome formation, engulfment into autophagosomes, fusion with lysosomes, and subsequent clearance occurs (11, 36). To verify that SAHA- or tubastatin A–mediated clearance of p62 requires both an intact tubulin network and dynein function, we exposed CSB-HF treated with HDAC inhibitors to the microtubule disrupting agent vinblastine or the dynein inhibitor ciliobrevin D. As expected, SAHA and tubastatin A reduced the p62 load in UVA-irradiated and unirradiated fibroblasts, and this effect was inhibited by cotreatment with vinblastine or ciliobrevin D (Fig. 5, H and I). In contrast, the transcriptional inhibitor α-amanitin did not abrogate this SAHA effect in UVA-irradiated CS1AN cells (fig. S9A). SAHA did not induce expression of autophagy-related genes or of genes controlled by the transcription factor ATF3, a known target of CSB, in the skin of UVA-irradiated csbm/m mice and CS1AN fibroblasts (fig. S9, B and C).

Together, we did not obtain evidence that de novo gene expression is of relevance for SAHA-mediated clearance of p62 aggregates. Our results suggest that SAHA induces autophagic clearance of p62 by enhancing α-tubulin acetylation via HDAC6 inhibition and improving lysosomal function. SAHA also enhanced acetylation of H4K16 in CS1AN cells and csb-1 knockdown C. elegans, indicating a general improvement in protein acetylation in the absence of functional CSB (Fig. 5, J and K).

SAHA rescues the skin phenotype in csbm/m mice

SAHA is a registered drug, which was approved by the Food and Drug Administration for the treatment of cutaneous T cell lymphoma, and has previously been used in mouse experiments (37, 38). We therefore asked whether SAHA treatment could rescue the skin phenotype in csbm/m mice.

UVA-irradiated mice received SAHA via the drinking water [2-hydroxypropyl-β-cyclodextrin (HOP-β-CD) was added to improve solubility and bioavailability]. This regimen was well tolerated because both genotypes did not show weight loss or any signs of behavioral abnormalities. As previously reported (39), csbm/m mice showed reduced weight gain throughout the experiment, and this phenotype was not altered by SAHA treatment (fig. S10). Unirradiated csbm/m mice, which received drinking water containing HOP-β-CD but not SAHA, showed a phenotype that was less pronounced than that observed in previous experiments in which mice received standard drinking water (Fig. 1E versus Fig. 6C, Fig. 2G versus Fig. 7E, and fig. S4A versus Fig. 7F). This discrepancy between experiments may be explained by the fact that HOP-β-CD can ameliorate the lysosomal storage phenotype in NP-C models, most likely by improving autophagic/lysosomal functions (40, 41). Nevertheless, irradiated csbm/m mice again showed a skin phenotype that was essentially identical to the one observed before (Figs. 1 and 2, and fig. S4 versus Figs. 6 and 7). SAHA treatment not only improved autophagic/lysosomal functions in irradiated csbm/m mice but also rescued their skin phenotype. Histological analysis revealed that loss of subcutaneous fat, fibrosis, inflammation, and epidermal hyperplasia in UVA-irradiated csbm/m mice was almost completely prevented in mice supplemented with SAHA, but not in mice receiving control drinking water (Fig. 6, A and B, and fig. S11A). Concomitant changes in skin physiological parameters (fig. S11B) and gene expression, where induction of fibrosis-associated genes was inhibited by SAHA (Fig. 6C), supported these results. These data show that treatment with a small-molecule HDAC inhibitor can prevent loss of subcutaneous fat, a cardinal symptom of csbm/m mice.

Fig. 6 The HDAC inhibitor SAHA prevents degeneration of SF in csbm/m mice.

(A) Macroscopic appearance of UVA-irradiated skin of csbm/m and WT mice treated with SAHA (+) or control water (CW, −) evaluated by H&E staining after low-dose UVA. Representative pictures are shown (n = 5). Scale bar, 200 μm. (B) Relative area of SF determined from murine skin sections from conditions listed in (A) (means ± SEM; *P < 0.05, **P < 0.01, two-way ANOVA with Tukey’s multiple comparisons tests; n = 4). (C) qRT-PCR analysis of Col1α1, Col1α2, Col3α1, FN-1, Timp-1, and MMP-3 mRNA expression in the skin of csbm/m and WT mice treated with SAHA or CW (−) after low-dose UVA. Expression was normalized to 18S rRNA expression (means ± SEM; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, two-way ANOVA with Tukey’s multiple comparisons tests; n = 4).

Moreover, the SAHA-mediated rescue of the skin phenotype in csbm/m mice was associated with an improvement in autophagy and lysosomal functions. As expected, irradiated csbm/m mice supplied with control drinking water had altered skin coloring, an accumulation of ubiquitinated proteins, LC3B and p62, and increased expression of genes including IL-1β, KC, TNF-α, TGF1, and HMOX-1. These changes were reduced or even completely abrogated by SAHA treatment (Fig. 7, A to F). Treatment also normalized overexpression of lysosomal genes at the mRNA level and reduced cathepsin B, cathepsin D, and pan-cathepsin at the protein level in the skin of irradiated csbm/m mice (Fig. 7, G and H). These results indicate that SAHA treatment can rescue the skin phenotype of UVA-irradiated csbm/m mice and that this effect is associated with increased α-tubulin acetylation and improvement of autophagic/lysosomal function.

Fig. 7 SAHA prevents accumulation of lysosomal proteins in the skin of UVA-irradiated csbm/m mice.

(A) Skin color (yellow/blue) of SAHA- or CW-treated csbm/m and WT mice after low-dose UVA measured with a chromameter (n = 5). (B to D) Immunohistochemistry of ubiquitin (B), LC3B (C), and p62 (D) in the subcutis of skin sections from csbm/m mice treated with SAHA (+) or CW (−) after low-dose UVA radiation. Representative images are shown (n = 3). Scale bars, 50 μm. (E, F) qRT-PCR analysis of IL-1β, TNF-α, KC, TGF1, COX2, HMOX-1 mRNA (E), and lysosomal gene (F) expression in the skin of csbm/m and WT mice treated with SAHA or CW after low-dose UVA (n = 4). (G) Immunohistochemistry of pan-cathepsin in the subcutis of skin sections from csbm/m mice treated with SAHA or CW after low-dose UVA radiation. Representative images are shown (n = 3). Scale bar, 50 μm. (H) Cathepsin expression in whole-skin lysates from csbm/m mice treated with SAHA or CW after low-dose UVA radiation was analyzed by Western blot using cathepsin B, cathepsin D, and pan-cathepsin antibody. β-actin was used as a loading control. ns, not significant.


Functions of the CSB protein may extend beyond its well-established role in DNA repair. Accordingly, autophagy was recently found to be impaired in vitro in CSB-HF (6). Here, we corroborate this finding and show that in csbm/m mice, impairment of autophagy is present in vivo and linked to degeneration of subcutaneous fat, a clinical hallmark of CS. By using three different species, we also provide evidence that autophagic dysfunction in CSB deficiency is due to reduced protein acetylation of the cytoskeletal protein α-tubulin. Enhancement of protein acetylation in CSB-HF or organisms by HDAC inhibitor treatment improved autophagy and rescued the skin phenotype in csbm/m mice. Our data thus indicate that this newly described function of the CSB protein contributes to the development of the clinical phenotype present in CS and that it may be therapeutically targeted to treat this disease. The HDAC inhibitor used in our study is a registered drug, and it will thus be possible in future studies to assess the translation of our results to CS patients.

We observed that dysfunctional autophagy in CSB-deficient mice, worms, and human fibroblasts was accompanied by signs of lysosomal dysfunction. We interpreted the accumulation of lysosomes and activated cathepsins in CSB-HF and in the skin of UVA-irradiated csbm/m mice as a compensatory response to cope with autophagic dysfunction and deposition of protein aggregates; this is similar to neurodegenerative diseases including Alzheimer, the LSD NP-C, or even normal aging (42). In a rat hippocampal slice model, accumulation of protein aggregates generated by treatment with the lysosomal inhibitor chloroquine not only resulted in increased expression of lysosomal enzymes but was also associated with decreased α-tubulin acetylation and concomitant destabilization of the tubulin network (43). Similarly, here, we observed that stimulation of autophagy upon UV stress was accompanied by increased α-tubulin acetylation in NHF, whereas total amounts of this posttranslational modification were reduced in CSB-HF. In our study, the administration of the pan-HDAC inhibitor SAHA or the HDAC6 inhibitor tubastatin A enhanced α-tubulin acetylation, increased the expression of lysosomal proteins, and allowed perinuclear sequestration and reduction of p62 aggregates. Treatment with the lysosome, tubulin, and dynein inhibitors BAF, vinblastine, and ciliobrevin D, respectively, revealed that HDAC inhibitor–mediated clearance of p62 aggregates in CSB-HF depended on lysosomal function, dynein, and an intact tubulin network. Thus, SAHA seems to stimulate autophagic flux by increasing α-tubulin acetylation and by improving protein aggregate trafficking toward the lysosome.

Upon chronic exposure to UVA radiation, the csbm/m mice used in the current study developed a marked skin phenotype, which involved the subcutis, dermis, and epidermis. It was characterized by a marked loss of subcutaneous fat, functional impairment of the epidermal barrier, and the development of skin fibrosis and increased expression of proinflammatory genes. Similarly, genetic ablation of the DNA repair factor XP-V in mice was previously reported to induce a proinflammatory response in the visceral fat, suggesting a pivotal role for NER-related factors in maintaining integrity of adipose tissue (44). Loss of subcutaneous fat is a hallmark of CS patients (1). Interstitial renal, leptomeningeal, and spinal root fibrosis have been described in CS patients (45, 46). Moreover, regional skin fibrosis seems to constitute a hallmark of progeroid syndromes (47). There is evidence that disturbance of autophagy and lysosomal function is involved in fibrotic disorders in various organs including lung, kidney, and liver (29, 48, 49). We therefore believe that an imbalance of HDAC and HAT activity, resulting in impaired acetylation of cellular proteins including α-tubulin, at least partially contributes to autophagic dysfunction, which then drives the development of the skin phenotype in our animal model.

We found that CSB interacts with both HDAC6 and MEC-17 in human fibroblasts. The interaction between HDAC6 and MEC-17 with the CSB protein varied for wild-type and mutated CSB. In particular, HDAC6 interacted more strongly with the mutated CSB protein, whereas MEC-17 interacted more strongly with the wild-type protein. Although these differences could be best seen in the nucleus, they are most likely of general relevance. Furthermore, CSB localizes to the centrosome, which is the main cellular MTOC and which is closely associated with the nucleus where CSB is predominantly located. UVA radiation promoted CSB recruitment to the centrosome along with induction of α-tubulin acetylation, indicating a causal relationship. At the centrosome, wild-type and mutant CSB partially colocalized with dynein, HDAC6, and acetyl–α-tubulin, which was still abundant at CSB-HF centrosomes, although it was reduced throughout the cytoplasm. These findings suggest a role for CSB in the regulation of α-tubulin acetylation and microtubule dynamics due to its capacity to interact with HDAC6, MEC-17, and potentially other binding partners. Other NER proteins might have a similar function because CSA, XPA, and XPC also interacted with HDAC6, and because α-tubulin acetylation was decreased in CSA-, XPA-, and XPC-deficient cells.

At this stage, we cannot exclude, however, that reduced acetylation of other target proteins or additional mechanisms might also be involved. Accordingly, it has been shown that UV- and hypoxia-mediated acetylation of histone H4 at the promoters of certain responsive genes is impaired in CSB-HF, and that this is due to defective recruitment of the HAT p300 to the acetylation site (19, 20). A similar function of CSB has also been demonstrated for another HAT, p300/CBP-associated factor (PCAF), because it is recruited to rRNA promoters in a CSB-dependent manner to enable transcription initiation (50). We believe that the impact of CSB on acetylation of nuclear targets might be relevant for its role in regulating chromatin remodeling (51, 52). CSB is capable of repositioning nucleosomes in an adenosine 5′-triphosphate (ATP)–dependent manner in cooperation with the nucleosome assembly protein 1 (NAP1)–like histone chaperone to enable efficient conduction of transcription-coupled repair (53). In addition to its role in transcription-coupled repair, NAP1 has also been shown to regulate histone acetylation at least at histone H3K9 during transcription (54), suggesting a functional interplay between histone acetylation status and chromatin remodeling. In keeping with this, hyperacetylation of histones H3 and H4 was detected at a recombination hotspot in Schizosaccharomyces pombe. Deletion of the HAT SpGCN5, which was required for most of the H3 acetylation at this hotspot, caused a significant delay in chromatin remodeling (55). Likewise, chromatin fibers containing acetylated H4K16 preferentially associated with Imitation Switch (ISWI), a catalytic subunit of ATP-dependent chromatin remodeling complexes (56), indicating that posttranslational histone modifications like acetyl-H4K16 support binding of chromatin remodeling proteins.

Here, we observed decreased global amounts of acetyl-H4K16 in CSB-HFs and csb-1–deficient C. elegans, indicating that CSB might affect chromatin remodeling by regulating acetylation of nuclear proteins. The epigenetic marker acetyl-H4K16, which has also been linked to cellular life span, affects basal cellular functions including transcription, DNA repair, and autophagy (5761), processes that are impaired in CS. Acetyl-H4K16 is associated with active and open chromatin structure, which will facilitate recruitment of additional factors that can in part directly bind to this modified lysine residue. Genome-wide analysis of H4K16 acetylation sites revealed an enrichment for DNA binding motifs of CTCF (62), a multifunctional protein involved in organization of chromatin architecture and transcription. CSB directly interacts with CTCF and facilitates its interaction with DNA upon exposure to oxidative stress. In contrast, CTCF conversely promotes CSB binding to certain genomic sites under these conditions (63). These results suggest that cooperation of CTCF and CSB might be critically involved in acetylation of H4K16 and mediation of associated downstream signaling. Impaired H4K16 acetylation may thus contribute to CS-specific alterations in chromatin remodeling and may also affect chromatin displacement of spliceosomes caused by unrepaired transcription-blocking DNA lesions (51, 64). Accordingly, acetyl-H4K16 underlies cross-talk with methylated H3K36, which has been shown to regulate pre-mRNA splicing in Saccharomyces cerevisiae (65, 66).

Here, we show that SAHA treatment rescued the skin phenotype of csbm/m mice. Given the importance of nonskin symptoms, in particular those concerning the central nervous system (CNS) in CS patients, it will be interesting in future studies to determine the effects of SAHA on noncutaneous phenotypes. Unfortunately, the neurological phenotype is absent or only weakly present in the CSB mouse model (67). Accordingly, we did not notice any signs of abnormal behavior in our mice. Addressing this question might therefore require human clinical studies. We believe that these are warranted because the mechanisms that we found to be improved by SAHA treatment in skin might also be involved in the pathogenesis of the neurological phenotype of CS patients. Accordingly, dysfunction in autophagy has been reported in various forms of neurodegenerative disorders including Alzheimer’s, Parkinson’s, and Huntington’s disease (68). Moreover, switching off autophagy by genetic ablation of atg7 in the CNS of mice causes a severe phenotype marked by neurodevelopmental disturbances and neurodegeneration (69). Decreased amounts of acetylated α-tubulin have been detected in these diseases as well (7072), and enhancement of α-tubulin acetylation by pharmacological inhibition of HDAC6 or SIRT2, respectively, improved phenotypic features in models of Huntington’s and Parkinson’s disease. In conclusion, our study provides evidence that enhancement of autophagy via the HDAC inhibitor SAHA is effective in improving the skin phenotype of UVA-irradiated csbm/m mice and thus may represent a novel therapeutic option for CS.


Study design

We used csbm/m mice, csb-1 knockdown C. elegans, and CSB-HF to study the role of autophagy in the pathogenesis of CS with particular emphasis on the skin. All mice used throughout this study were exclusively on the hairless SKH1 background because this model is widely used to investigate the skin phenotype of irradiated mice. In csbm/m mice and wild-type littermates, autophagy was induced in the skin by chronic exposure to UVA radiation. For some experiments, mice were treated with the HDAC inhibitor SAHA or the solvent HOP-β-CD, both applied via drinking water. Mice were randomly selected for the respective treatment groups. Determination of sample sizes was based on previous experiences with mouse experiments. Irradiation was continued until UVA-irradiated csbm/m mice not receiving SAHA showed visible skin alterations. Because of the conspicuous skin phenotype of the mice, these studies could not be blinded. All data were included. All animal studies were approved by the local ethical committee of the Federal State of North Rhine-Westphalia of the Landesamt für Natur, Umwelt und Verbraucherschutz in Düsseldorf, Germany. The collection of human skin biopsies for isolation of fibroblasts from healthy volunteers was approved by the local ethical committee of the Medical Faculty of the Heinrich-Heine University of Düsseldorf, Germany and was conducted after informed consent of the volunteers.

Mice and treatment

CSB-deficient (csbm/m) mice generated by van der Horst et al. (67) were crossed onto SKH1 hairless mice. Together with wild-type littermates, these mice were subjected to a chronic UV irradiation protocol using an UVA1 irradiation device (340 to 400 nm; Sellamed Systems) at an age of 5 to 9 weeks. Mice were irradiated five times a week with a single daily dose of 10 J/cm2 for 24 weeks (low dose) or 20 J/cm2 for 42 weeks (high dose).

Measurement of TEWL, sebum content, skin humidity, and skin color

TEWL, sebum content, and skin humidity were analyzed by using a Tewameter (TM 300), a sebumeter, and a corneometer, respectively (all instruments are from Courage & Khazaka). Skin color was measured by using a chromameter (CR-300, Konica Minolta). The b value representing the ratio of yellow and blue color was determined.

Cell culture and treatment

CSB-mutated CS1AN fibroblasts were obtained from the Coriell Institute. A mutation in one of the csb alleles of CS1AN cells results in generation of a truncated CSB protein. A similar mutation was introduced into the murine csb gene for the generation of the csbm/m mice (67). NHFs were isolated from skin samples obtained from healthy donors. Twenty-four hours before irradiation, cultivation medium was exchanged to phenol red–free medium containing 2% fetal calf serum. SAHA (2 μM), tubastatin A (5 μM), or DMSO as solvent control was added to the medium. For some experiments, cells were additionally treated with BAF (20 nM), vinblastine sulfate (20 nM), ciliobrevin D (20 μM), or α-amanitin (2.5 μM), which was added 2 hours before irradiation. For UV irradiation, cells were washed and covered with phosphate-buffered saline before being placed under the UVA1 source (Sellamed Systems; same model as used for mouse irradiation) and irradiated with a single dose of 20 J/cm2. After irradiation, cells were supplied with fresh phenol red–free medium containing the respective ingredients and cultured for the indicated time before further processing for the required applications.

qRT-PCR analysis

For RNA isolation from murine skin, an RNeasy Fibrous Tissue kit (Qiagen) was used. Initially, skin samples were disrupted in a TissueLyser (Qiagen). All successive steps were conducted according to the manufacturer’s instructions. RNA from primary human fibroblasts was isolated using TRIzol reagent (Life Technologies) and purified by deoxyribonuclease digestion. Complementary DNA synthesis was performed using 500 ng of RNA and M-MLV Reverse Transcriptase (Promega). qRT-PCR was conducted using a CFX Connect Real-Time PCR Detection System and iQ SYBR Green Supermix (Bio-Rad). Expression of genes of interest was normalized to expression of 18S rRNA. Primer sequences are listed in table S1.

Statistical analysis

Preparation of figures and statistical analysis was performed using GraphPad Prism software. For all experiments, data are expressed as means ± SEM or SD. The statistical significance of the observed differences was calculated using one- or two-way ANOVA with Tukey’s multiple comparisons tests. Results were considered significant when P < 0.05.


Materials and Methods

Fig. S1. UVA radiation causes collagen accumulation in the skin of csbm/m mice.

Fig. S2. UVA does not affect viability of CS1AN fibroblasts.

Fig. S3. Autophagy is inhibited in CS1AN fibroblasts.

Fig. S4. UVA causes abnormal accumulation of lysosomal markers in the skin of csbm/m mice.

Fig. S5. HDAC genes are dysregulated in the skin of csbm/m mice.

Fig. S6. Senescence does not affect acetylation of α-tubulin in NHF.

Fig. S7. CSB protein is located at the centrosome in NHF and CS1AN fibroblasts.

Fig. S8. NER factors interact with HDAC6 and affect acetylation of α-tubulin.

Fig. S9. SAHA does not induce expression of ATF3-regulated or autophagy-related genes to improve autophagy.

Fig. S10. SAHA does not affect weight gain of csbm/m mice.

Fig. S11. SAHA improves the skin phenotype of UVA-irradiated csbm/m mice.

Table S1. List of primers used in this study.


Acknowledgments: We are grateful to B. van der Horst and H. van Steeg for providing csbm/m mice. We thank I. Haußer-Siller for electron microscopic analysis. We are grateful to M. Berneburg for discussion and critical reading of the manuscript. Some nematode strains used in this work were provided by the Caenorhabditis Genetics Center, funded by the NIH Office of Research Infrastructure Programs (P40 OD010440). We also would like to thank Thorsten Hoppe Laboratory for providing the PP1545 strain. Funding: Funding for this study was provided by the Deutsche Forschungsgemeinschaft (SFB 728, Kr 871/5-1 to J.K.; VE 663/3-1 to N.V.), the Bundesministerium für Bildung und Forschung, the Italian Association for Cancer Research (AIRC; MFAG11509 to N.V.), and a Symrise fellowship (to M.M.). Author contributions: M.M. and J.K. designed the experiments and wrote the manuscript. M.M. and I.U. conducted in vivo and in vitro experiments. K.S. and M.K. conducted in vitro experiments. N.V. designed the C. elegans experiments, which were conducted by A.S. C.E. was responsible for backcrossing and breeding of the mice. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data associated with this study are present in the paper or the Supplementary Materials.

Stay Connected to Science Translational Medicine

Navigate This Article