Proteins altered in CS, including CSA and CSB, share a common function in transcription-coupled nucleotide excision DNA repair (TC-NER) in the nucleus and base excision repair (BER) of mitochondrial DNA (Sykora et al., 2012). TC-NER specifically removes lesions in DNA that obstruct RNA polymerases and block transcription, and is distinguished from global genome NER (GG-NER), which operates throughout the genome to eliminate helix-distorting DNA damage (Hoeijmakers, 2009). CSA and CSB are required for removal of RNA polymerase stalled at the site of a lesion for the repair reaction to occur (Fousteri & Mullenders, 2008). CS proteins are also localized to mitochondria and are thought to recruit and stabilize BER proteins to repair complexes in the inner mitochondrial membrane. Human CS cells exhibit an altered redox balance with increased intracellular ROS and mitochondrial dysfunction (Scheibye-Knudsen et al., 2012). Whether this reflects indirect effects of unrepaired DNA damage, or a more direct role of CS proteins in redox balance remains unknown. NER proteins also participate in a number of transcription-related processes including initiation and elongation (Brooks, 2013). Mutations in other NER-related genes (XPD, XPG, XPB) can also cause CS in combination with symptoms of xeroderma pigmentosum (XP), most notably pigmentation abnormalities and elevated skin cancer in UV-exposed skin, resulting in the combined disorder XPCS (Cleaver et al., 2009). CS includes a spectrum of severities with life expectancies ranging from 5 years for the most severe form (early onset type II including cerebro-oculo-facio-skeletal syndrome, COFS) (Suzumura & Arisaka, 2010), 16.1 for moderate (classical type I) and 30.3 for atypical (mild, late onset type III) (Natale, 2011).
Numerous mouse models of CS (Csb−/−, Csa−/−) and XPCS (XpdXPCS/XPCS) have been established, however, with a generally milder phenotype than CS (van de Ven et al., 2007). Further inhibition of NER by crossing these models into backgrounds lacking the GG-NER-specific protein XPC, or XPA which is common to both subpathways, results in a more severe disorder, NER progeria, that more closely resembles the small size, lipodystrophy, and neurological complications of CS (van de Ven et al., 2007). A hallmark of severe NER progeria in Csb−/− | Xpa−/− (Murai et al., 2001; van der Pluijm et al., 2006), XpdXPCS/XPCS | Xpa−/− (Andressoo et al., 2006), Csb−/− | Xpc−/− (Laposa et al., 2007) and Xpg−/− (Harada et al., 1999) mutant models is death before weaning at approximately 3 weeks of age. Attempts to extend lifespan past the weaning period by such interventions as reducing litter size, moistening the food (van der Pluijm et al., 2006), providing surrogate mothers (de Boer et al., 2002), or extending the nursing period (Andressoo et al., 2006) have so far proven unsuccessful.
We generated double homozygous mutants by intercrossing Csa−/− | Xpa+/− or Csa+/− | Xpa−/− mice. Csa−/− | Xpa−/− (CX) mice were born at Mendelian ratios and were indistinguishable at birth from controls. By postnatal day 5, CX mice were visually smaller in size compared to WT or single KO littermates. By postnatal day 12, clenching of the hind limbs, indicative of neurological dysfunction, was evident. As with other progeroid NER mice, CX mice reached a peak weight at approximately 2 weeks of age, followed by a decline in fitness, failure to thrive, and premature death with nearly 100% penetrance by postnatal day 28. Control heterozygous and single homozygous mutant CSA (Csa−/−) and XPA (Xpa−/−) mice developed normally as reported previously (de Vries et al., 1995; van der Horst et al., 2002).
Despite the fact that CX mutants nursed and had milk in their stomachs observable upon sacrifice between days 14 and 21, white adipose but not brown adipose tissue weights were reduced relative to body weight on postnatal day 10 (Fig. 1A). Serum triglycerides and blood glucose were also reduced (Fig. 1B,C). We hypothesized that a nutritional deficiency might contribute to death at weaning and reasoned that increasing the energy density of the milk by supplementing lactating dams with a high-fat diet may ameliorate this mutant phenotype. Feeding pregnant dams a high-fat diet (HFD, 60% calories from fat) vs. standard chow (22% calories from fat) increased triglycerides in their milk without significantly altering protein content (Fig. S1) and thus increasing total calories. Consistent with our hypothesis, HFD feeding of lactating dams rescued premature death in CX mice with high penetrance (Fig. 1D). However, counter to our hypothesis, a control low-fat diet (LFD) with only 10% calories from fat also rescued premature death (Fig. 1D) without significantly increasing triglycerides or total calories in the milk (Fig. S1).
Because both diets (HFD, LFD) that rescued death before weaning were composed of refined ingredients, we next hypothesized that a contaminant or toxin present in the standard chow, a grain-based diet made with unrefined ingredients, might interact with the congenital DNA repair deficiency and cause premature death. To address this, we pulverized refined and unrefined pellets and prepared them in a final 1% semi-solid agar form separately or in a 3:1 ratio. Regardless of the composition of the diet, CX mutant litters receiving agar-based diets survived weaning with high penetrance (Fig. 1D). These data are inconsistent with a toxin in unrefined chow or differences in macronutrient content between refined and unrefined diets as causative of premature death in CX mice. Instead, they suggest a defect in the ability to transition from mother's milk to solid food unless that food is soft. Consistent with this, HFD pellets are soft due to their high fat content, and LFD pellets, although initially hard, readily absorbed water (Fig. S2) and became soft in the hopper within days, likely due to the absorbance of humidity from the cage. To directly test the potential role of soft food during the transition from milk to solids, we presented litters with standard chow in soft agar form from postnatal day 14, coincident with this transition in C57BL/6 mice (http://jaxmice.jax.org/literature/factsheet/LT0001_Pups.pdf). Twelve of 19 CX mice survived weaning at 28 days (Fig. 1D). We conclude that CX mutants that cannot transition to solid food during this developmental window die of malnutrition (Fig. 1E).
CX mutants weaned at 4 weeks of age onto chow in soft agar form had an average lifespan of ~16 weeks (Fig. 2A). Five CX mutants weaned onto hard pelleted chow displayed similar longevity (data not shown); the requirement for soft food is thus likely to occur only during an early developmental window. CX mutants weaned onto HFD had slightly extended mean but not maximal lifespan relative to LFD or chow (Fig. 2A). CX mutants remained smaller than control littermates throughout their lifespans (Fig. 2B,C). Similar to preweaned mice, CX mutants weaned onto a chow diet had reduced adiposity, consisting of ~2% of body weight through most of their adult life. Mutants on a HFD from 7 weeks of age achieved higher adiposity of ~5%, but failed to maintain it toward the end of their lifespan (Fig. 2D). Lean percent body mass was maintained independent of diet or loss of adiposity; organ weights were also maintained relative to body weight except for a reduction in WAT, while heart, lung, and brain were slightly increased (Fig. S3).
Neurological involvement was observed on several levels, including abnormal gait consistent with cerebellar ataxia, inappropriate hindlimb clasping in a tail-hang test (Fig. S4) and progressive reduction in grip strength (Fig. 2E) culminating in uncoordinated hindlimb movement, dystonia, and eventual paralysis (Fig. 2F, Fig. S5 movies). Other symptoms typical of progeroid disorders were also observed, including kyphosis (Fig. S6). The reasons underlying these progressive progeroid symptoms are unknown, and they could not be rescued by feeding refined or high energy-density diets (Fig. 2A). Taken together, our data indicate that early postnatal death hampering existing models of severe CS resulted from malnutrition and could be rescued with soft food during the critical developmental transition from milk to solid food (Fig. 1E). Surviving animals were smaller in size and had progressive adipose and neurological features reminiscent of CS.
Why were CX mice unable to transition normally to solid food? Prolonged survival on a soft-food diet after weaning was previously observed in an inducible mouse model of Hutchinson–Gilford progeria syndrome with apparent dental abnormalities (Sagelius et al., 2008). Oral anomalies are frequently noted in patients with CS(Schneider, 1983; Sorin, 1994; Horbelt, 2010); however, analysis of bone density in CX teeth revealed no defects in tooth development, enamel, emergence of incisors or molars, or differences in root length during the developmental window of P14 to P19 or later in life (Fig. S7). Patients with CS typically have problems in eating, possibly due to a neuromuscular deficit, such as restricted mandibular range of motion that can hinder chewing (Boraz, 1991), or coordination of chewing and swallowing (E. Nielan, personal communication). As a result, many are fed soft food and/or receive nutrition via a gastrojejunostomy feeding tube. Future experiments will be required to address the underlying causes of feeding problems in patients with CS, as well as the developmental defect in eating hard chow pellets observed in CX mice, and any potential connection between the two.
Why is rescue of death before weaning in CX mice important? This is the first severe CS model to survive weaning with high penetrance, allowing for the analysis of key progeroid features of CS, including progressive loss of adiposity and neurological dysfunction, and their potential relationship to DNA repair insufficiency outside of the window of early development. It also cautions against the use of lifespan extension beyond weaning as a means to score efficacy of experimental interventions until the cause of the inability to switch to solid food is better understood (van der Pluijm et al., 2006).