Although patterns of tooth wear are crucial in palaeo-reconstructions, and dental wear abnormalities are important in veterinary medicine, experimental investigations on the relationship between diet abrasiveness and tooth wear are rare. Here, we investigated the effect of four different pelleted diets of increasing abrasiveness (due to both internal [phytoliths] and external abrasives [sand]) or whole grass hay fed for 2 weeks each in random order to 16 rabbits (Oryctolagus cuniculus) on incisor and premolar growth and wear, and incisor and cheek tooth length. Wear and tooth length differed between diets, with significant effects of both internal and external abrasives. While diet abrasiveness was linked to tooth length for all tooth positions, whole forage had an additional effect on upper incisor length only. Tooth growth was strongly related to tooth wear and differed correspondingly between diets and tooth positions. At 1.4–3.2 mm/week, the growth of cheek teeth measured in this study was higher than previously reported for rabbits. Dental abnormalities were most distinct on the diet with sand. This study demonstrates that concepts of constant tooth growth in rabbits requiring consistent wear are inappropriate, and that diet form (whole vs. pelleted) does not necessarily affect cheek teeth. Irrespective of the strong effect of external abrasives, internal abrasives have the potential to induce wear and hence exert selective pressure in evolution. Detailed differences in wear effects between tooth positions allow inferences about the mastication process. Elucidating feedback mechanisms that link growth to tooth-specific wear represents a promising area of future research. J. Exp. Zool. 321A:283–298, 2014. © 2014 Wiley Periodicals, Inc.
Teeth are essential for most mammals. Dental adaptations to diets are well documented. In herbivorous mammals, various morphological adaptations (such as enamel ridge alignment) or chewing muscle size and physiological adaptations (such as rumination) have evolved to maximize chewing efficiency (Clauss et al., 2008; Fritz et al., 2009; Schwarm et al., 2009; Kaiser et al., 2010). Various adaptations are directed towards ensuring continuous tooth function in the presence of abrasion, most notably high tooth crowns (hypsodonty) (Damuth and Janis, 2011; Damuth and Janis, 2014), the ever-growing (hypselodont) teeth of many rodents and lagomorphs (Ungar, 2010), and increased enamel thickness (Rabenold and Pearson, 2011). Durable teeth are important for longevity and hence lifetime reproductive output (Skogland, 1988; Kojola et al., 1998; Loe et al., 2006; Veiberg et al., 2007).
Differences in the patterns of tooth wear have been described at the macroscopic and microscopic level as mesowear (Fortelius and Solounias, 2000) and microwear (Walker et al., 1978), respectively, and microtexture (Ungar et al., 2003; Schulz et al., 2010). These patterns are interpreted in relation to the diets usually reported for the species in question (e.g., Rodrigues et al., 2009; Scott, 2012; Scott et al., 2012; Kaiser et al., 2013), but the resulting interpretations about the processes of wear remain mostly untested. Various analyses of tooth wear on populations with different diets or management practices (Ward and Mainland, 1999; Mainland, 2000; Mainland, 2003a; Mainland, 2006; Clauss et al., 2007; Kaiser et al., 2009; Merceron et al., 2010; Taylor et al., 2014) help formulate such testable concepts of tooth wear. Controlled feeding experiments that test hypotheses about tooth wear on different diets have, in contrast, only been rarely reported so far (Mainland, 2003b; Schulz et al., 2013; Solounias et al., 2014).
Processes of tooth wear are still not well understood. “Wear” of dental tissue can be caused by both tooth-to-tooth contact, that is, attrition, and by abrasion due to internal abrasives, such as phytoliths in grasses, or external abrasives such as dust or grit (Butler, 1972; Fortelius, 1985; Kaiser et al., 2013). However, to what degree internal and external abrasives cause dental tissue loss is a matter of ongoing debate (Mainland, 2003a; Sanson et al., 2007; Damuth and Janis, 2011; Lucas et al., 2013; Erickson, 2014; Rabenold and Pearson, 2014). The degree of hypsodonty in herbivores is linked to the excretion (and hence intake) of abrasive elements in their faces (Hummel et al., 2011), but clarification of whether these elements represent internal or external abrasives is still lacking. That the ingestion of external abrasives induces wear was demonstrated in several studies (Healy and Ludwig, 1965; Ludwig et al., 1966; Healy et al., 1967; Mainland, 2003a) and is usually not debated, but the contribution of internal abrasives remains largely untested.
The incisors of rodents and lagomorphs, and in some species also the cheek teeth, are ever-growing, which is an evident adaptation to compensate for tooth wear (Rensberger, 1975; Rensberger, 1986; Williams and Kay, 2001; Schmidt-Kittler, 2002). It is obvious that the rates of wear and growth must match in ever-growing teeth in order to maintain proper occlusion (Schmidt-Kittler, 2002; Ungar, 2010). However, to what degree growth actually responds to wear, and whether a certain growth occurs at a fixed rate that cannot be modified, is mostly unexplored. In veterinary clinical practice, and in common veterinary textbooks, constant growth is assumed that will lead to dental abnormalities if wear is insufficient (Crossley, 2000; Harkness et al., 2010). This concept is supported by findings in abnormal pet rabbits where the incisor teeth are not in occlusion: normal tooth wear cannot occur, and the incisors grow uncontrolled, with the maxillary incisors typically curling inward into the oral cavity or flaring out laterally, and the mandibular incisors protruding from the mouth (Van Caelenberg et al., 2008; Harcourt-Brown, 2009).
However, common sense suggests the existence of some regulatory mechanism that matches growth to abrasion, because free-ranging animals will face a variety of resources that are not identical in their abrasiveness, and will also undergo a variety of metabolic states that require different levels of food intake. For example, during hibernation, tooth growth is decreased distinctively in ground squirrels (Spermophilus tridecemlineatus) (Sarnat and Hook, 1942). In beavers (Castor canadensis), growth rates of incisor teeth vary between seasons, being higher during summer and lower during winter (Rinaldi and Cole, 2004). Actually, in a study on tooth wear and growth in pet rabbits (Wolf and Kamphues, 1995; Wolf and Kamphues, 1996), incisor growth rates varied on different diets, apparently compensating for differences in tooth wear, but this flexibility was not emphasized.
An important aspect of a putative regulatory mechanism for tooth growth is that beyond the findings on variation with season or diet, this mechanism must be tooth-specific, because wear will differ between teeth of different positions, especially between incisors and cheek teeth. In the veterinary literature, growth rates of rabbit teeth are given as 1.3–3.0 mm/week for incisors (Wolf and Kamphues, 1996) and 2.0–3.0 mm/month for cheek teeth (Meredith, 2007; Lord, 2011; Schumacher, 2011), with the implication that this growth needs to be matched by constant wear (induced by an appropriate diet) to avoid overgrowth. Evidence for tooth-specific growth regulation comes from examples in rats and rabbits where a single incisor was broken off and showed an increased growth rate as compared to its contralateral neighbor that was unbroken and hence in continuous occlusion with its antagonist (Schour and Medak, 1951; Ness, 1956).
Rabbits as Model Animals
Rabbits are attractive model animals to study tooth wear and growth because they are natural herbivores accepting a variety of feeds, and are comparatively easy to maintain. Furthermore they have continuously growing incisors and cheek teeth, which can be manipulated for macroscopic inspection as well as for computed tomography (CT), and their dental health has been studied extensively in the veterinary literature (Meredith, 2007; Capello and Cauduro, 2008; Van Caelenberg et al., 2010; Van Caelenberg et al., 2011; Jekl and Redrobe, 2013). Dental problems are one of the most important conditions for presenting pet rabbits to veterinary clinics, with frequencies for dental disease in rabbits ranging from 6.7% (Mosallanejad et al., 2010), to 14% (Langenecker et al., 2009), 30% (Mullan and Main, 2006), and even 38.1% (Jekl et al., 2008). Although these surveys show how significant dental problems are, the etiology of this disease complex is still not fully understood.
Malocclusion is the common denominator of dental problems in rabbits, but the reasons for malocclusion to occur in the first place are debated (Jekl and Redrobe, 2013). They are divided into congenital and acquired causes (Lennox, 2008). Congenital or hereditary tooth abnormalities are often diagnosed in younger animals. The most common condition is incisor overgrowth. This may be due to maxillary brachygnathia or brachycephalism, and is particularly seen in dwarf breeds (Crossley, 1995). The underlying cause is an autosomal recessive gene (Harkness et al., 2010). In most cases the incisor teeth are not in occlusion and the mandibular incisors deviate labially from the maxillary ones.
Acquired causes may include traumatic injuries that lead to malocclusion (Capello, 2008), but are mainly related to diet, and to management factors that determine exposure to UVB light (Jekl and Redrobe, 2013). Dietary mineral imbalances, or lack of exposure to UVB, may lead to metabolic bone disease, which may impair occlusion due to osteodystrophy of the supporting bone and dental tissue malformation (Harcourt-Brown, 1995). Experimentally, minerally imbalanced diets led to cheek tooth elongation and enamel hypoplasia in degus (Octodon degus) (Jekl et al., 2011a,b). The other major dietary factor considered responsible for dental abnormalities are easily digestible (i.e., low-fiber) diets that limit the absolute food intake, because energetic requirements are met by small amounts of such diets, leading to insufficient chewing activity and hence insufficient attrition (Wolf and Kamphues, 1996; Crossley, 2003; Meredith, 2007; Harkness et al., 2010; Lord, 2011). Feeding pet rabbits dried forages as the staple diet item is therefore—amongst other reasons—recommended (Boehmer and Koestlin, 1988; Clauss, 2012). Whether dietary abrasiveness itself, in addition to the effect of food intake and chewing activity, is also important has so far not been investigated. Dental problems are epidemiologically related to the use of low-fiber, energy-dense feeds that are, however, also often minerally imbalanced or leave the animals the choice of selecting minerally imbalanced ingredients (Harcourt-Brown, 1996; Mullan and Main, 2006). Jekl and Redrobe (2013) suggest that a combination of metabolic bone disease and insufficient wear or lack of chewing action is responsible for dental disease in pet rabbits.
Aims of This Study
We aimed at investigating the effect of diet on dental wear in rabbits using a set of four complete, pelleted feeds varying in the amount of internal and external abrasives. The pelleted diets were based on lucerne (Medicago sativa), which naturally contains very low levels of internal abrasives (Wöhlbier, 1983), grass, which contains higher levels of internal abrasives, and grass with the addition of rice hulls, which contain again higher levels (Wöhlbier, 1983). The fourth pelleted diet included grass and rice hulls (internal abrasives), and additionally sand (external abrasives). To avoid differences in the total amount ingested (and hence chewing activity) between these pelleted diets, they were formulated to be isocaloric and isonitrogenic (using an indigestible, non-silicacious filler and soybean meal). The following hypotheses guided our approach:
- Tooth growth compensates for wear; therefore we expect tooth length (TL) to be relatively constant across diets and growth tightly correlated with wear.
- Small, detectable differences in wear between diets reflect diet abrasiveness and/or chewing activity.
- Functional differences between incisors and cheek teeth lead to different wear and growth on different diets, that is,
- incisors are worn more heavily when feeding whole hay that needs more gnawing as compared to pellets;
- cheek teeth, with a chewing action more independent from whether the diet is offered whole or pelleted, are worn more heavily with increasing dietary abrasiveness; external abrasives (sand) lead to a gradient in wear along the maxillary cheek tooth row whereas increased internal abrasives (phytoliths in rice hulls) do not lead to such a gradient (Taylor et al., 2013).
- Abnormal tooth wear will occur more frequently with excessive external abrasives (sand), and (according to 3b) affect the cheek teeth according to their position in the tooth row (anterior ones more affected).
MATERIALS AND METHODS
Animals and Diets
This experiment was approved of by the Cantonal Veterinary Office in Zurich, Switzerland (no. 80/2012). Sixteen female New Zealand White rabbits (mean starting body mass 2.75 ± 0.16 kg; starting age approximately 7 months) were kept individually in hutches (1.00 × 0.75 m2) on wood shavings and with plastic hides without other gnawing opportunities except their diet. Water was provided ad libitum. After a week of acclimatization the rabbits were randomly placed on one of five different diets for a 2-week-period. Diets consisted either of a grass hay fed as whole forage, or of complete pelleted diets formulated to be isocaloric and isonitrogenic, but of increasing abrasiveness from lucerne pellets (L), grass pellets (G), grass and rice hull pellets (GR), and grass and rice hull pellets with an addition of sand (GRS) (Table 1). Pellets were uniform across diets, of approximately 1 cm length and 4 mm diameter, requiring breaking with incisors before they could be chewed. Diets were fed ad libitum. Food intake was measured on a daily basis by weighing food offered and leftover. Rabbits were weighed at the end of each week, and the difference in body mass to the previous measure expressed in %. Food intake rate was measured as a proxy for chewing intensity on the different diets by timing the duration it took each rabbit to ingest 10 g (as fed). After 14 days, a new, again randomly selected diet was introduced to each rabbit with a 2-day-interval of mixing the old and the new diet to prevent digestive problems due to abrupt diet change. At the end of the experiment, all animals had received all diets. To evaluate digestibility of the different diets, total feces from each animal were collected and weighed for three consecutive days in the second week of every 2-week-interval. Representative samples of feeds, leftovers (in the case of hay), and feces were taken, dried at 60°C to constant weight, and prepared for further analyses by grinding. Protein, detergent fibers and acid detergent insoluble acid (ADIA) were measured using standard procedures (Van Soest et al., 1991; VDLUFA, 2007, method 4.1.2, Dumas method; Hummel et al., 2011). A sample of the sand included in GRS was analyzed for mean particle size by wet sieving according to Fritz et al. (2012).
|Lucerne meal (%)||60.0||—||—||—||—|
|Grass meal (%)||—||60.0||64.8||64.8||—|
|Rice hulls (%)||—||—||20.0||20.0||—|
|Pure lignocellulose (%)||33.8||27.4||5.0||—||—|
|Soybean meal (%)||—||7.0||5.0||5.0||—|
|Soy oil (%)||1.0||0.4||—||—||—|
|Mineral/vitamin premix (%)||0.2||0.2||0.2||0.2||—|
|Dry matter (% as fed)||91.4||91.9||91.8||92.2||90.8|
|Nutrient composition (g/kg DM)|
|Dry matter digestibility (%)||39.7 ± 9.3||34.3 ± 8.1||41.2 ± 5.7||40.7 ± 11.1||45.1 ± 4.1|
In the dental nomenclature used here, I1 and I1 denote the upper and lower incisor, respectively, maxillary (upper) cheek teeth are denoted by capital letters (P or M for premolars and molars, respectively, i.e., P2-M3), mandibular (lower) cheek teeth by lower case letters (p or m, i.e., p2-m3) (Thenius, 1989). At the beginning of each 2-week-interval, the rabbits' teeth were burred and CT of the head was carried out, and a final CT was taken at the end of the last feeding period. For these procedures the animals were placed under general anesthesia. The animals were sedated with 0.5–1.5 mg/kg midazolam (Dormicum®, Roche AG, Reinach, Switzerland) intramuscularly and anesthesia was induced and maintained with isoflurane administered in oxygen using a facemask 20–30 min later. During anesthesia, punctual marks of the size of approximately 1 mm were burred on the labial side of the I1 and I1 of both sides (total of 4 marks), and on the mesial side of the first lower cheek teeth (p3, cf. Fig. 1a) of either side (total of 2 marks) using a diamond burr (KaVo, EWL, Leutkirch i. A., Germany). Visible tooth crown length, and distances between gingival margin and mark (to assess tooth growth) and between mark and tooth edge (to assess tooth wear) of the I1 and I1 were measured (Wolf and Kamphues, 1995) using an electronic digital caliper (Technocraft® Allchemet AG, Bäretswil, Switzerland, precision 0.01 mm). For the p3, because calipers could not be introduced far enough into the oral cavity, the same measures were estimated using a periodontal probe as a scale. Directly after burring, CT images of the head of each rabbit were obtained. The rabbits' recovery from anesthesia was uneventful in all cases and they usually started eating again two hours later. To monitor wear and growth of incisor teeth, caliper measurements were repeated a week later under manual restraint. Because the visible crowns of the p3 were smaller than those of the incisor teeth and the marks tended to wear off sooner, p3 were monitored more closely under manual restraint every 3–4 days. This was done by digital photography of the mesial side of each p3, taken with a portable endoscopic camera (Envisioner Medical Technologies, Inc., Rockville, Maryland), a LED-battery light source (Karl Storz GmbH, Tuttlingen, Germany) and the use of a rigid telescope (170°, 23 cm × 2.0 mm, Richard Wolf GmbH, Knittlingen, Germany) guided through the metal cone of an otoscope (Fig. 1a). Again, TL and distances from gingival margin to the mark and from the mark to the tip of the tooth were estimated. All manual measurements were taken by the same examiner (J.M.).
Computed Tomography and Evaluation
CT scans (Fig. 1b) were obtained using a 16 slice, spiral CT-scanner (Philips Brilliance 16, Philips Healthcare, Zurich, Switzerland). Images were acquired at 120 KV, 117 mA, a 10 cm FOV with a slice thickness of 1 mm. The rabbit was positioned in ventral recumbency on the CT table to obtain transverse sections. The original data were reconstructed with a soft-tissue and a bone algorithm and was reviewed using a bone window setting (window width = 3,814 HU, window level 594 HU) and a soft tissue window setting (window width = 270 HU, window level = 100 HU). CT images of the heads of all animals were investigated using OsiriX® software (PixmeoSarL, Bernex, Switzerland). Total lengths of all teeth were measured except for the peg teeth and the M3 and m3. Incisor teeth were measured in the sagittal plane, cheek teeth in the coronal plane with the help of an open polygon, a function that made it possible to measure curved structures. For each tooth the total length of the lingual and the buccal side was measured using this function as a curved structure from the base to the apex of the tooth. On CT images special attention was given to the occlusal surface, and it was noted whether spurs existed or if the tooth angle (TA) of the occlusal surface was abnormally sloped. Tooth spurs (TSP) were categorized from no spurs, slight, moderate or severe spurs (Scores 0–3, see Supplementary Material for examples of all scores used in this study), as spurs were too small for exact measurements. It was noted whether the occlusal surface (TSF) of each cheek tooth was either horizontal (Score 0), convex (Score −1) or concave (Score 1). The tooth surface angle (TA) was scored as horizontal (Score 0), sloped to the buccal side (Score −1) or sloped to the lingual side (Score 1). Additionally the dentition was checked for signs of waviness in the sagittal plane (“stepmouth”) and again categorized from no signs, slight, moderate or severe signs of waviness (Scores 0–3). All CT scans were evaluated by the same examiner (J.M.) who was blinded to the diet that the respective scans represented.
Data were analyzed using mixed-effects linear models (MELMs), accounting for repeated measurements from the same individuals by including Individual as a random effect to ensure correct error terms were being compared. Initially, we compared the general nutritional status of rabbits across diet treatments, including Diet as a main effect, and evaluating several response variables: body mass at the end of the relevant trial period; the relative (as a percentage of initial body mass) change in body mass over the relevant trial period (body masschange(%)); dry matter intake (DMI); acid detergent insoluble ash intake (ADIAI); dry matter digestibility (DMD); and intake rate (speed; measured as the time, in minutes, required to eat 10 g of food). Except for intake rate and DMD (which were only measured in the second week of each diet regime), we also included the Week on a particular diet treatment (1 or 2), and the Previous diet treatment (body mass variables only), as well as interactions between these terms with Diet, to account for autocorrelation and other temporal effects. We then assessed diet effects on TL, growth, and wear using similar MELMs, except that for TL (week 2 only) we also included Tooth as a main effect (as well as its interaction with Diet) to test for differences along the cheek tooth row. We also analyzed wear and growth responses to ADIAI and intake rate independently of the specific diet treatment, by replacing Diet with continuous quantitative variables—that is, wear as an effect on growth, and ADIAI (abrasiveness of diet) and intake rate, respectively, as effects on wear. Finally, to check for diet effects and tooth row position on dental abnormalities (hypothesis 4), we used MELMs with spurs (TSP), tooth surface (TSF), and TA scores, respectively, as well as waviness, as dependent variables, and Diet and Tooth (the latter not in assessments of Waviness) as main and interactive effects. However, for these models we excluded teeth for which abnormality scores were not different from zero, based on one-sample t-tests.
Data for incisors and cheek teeth, and for upper and lower teeth, were analyzed in separate MELMs. In cases where interaction terms were not significant (P > 0.05), these were removed and the relevant models repeated without them. Bonferroni post hoc tests were used for multiple comparisons where necessary. All analyses were carried out in STATISTICA version 8 (Statsoft_Inc., 2007).
General Diet Effects
Rabbit body mass differed across diet treatments, with lower means for rabbits on H than on other diets (F4,119 = 10.178, P < 0.0001 for body mass; and F4,115 = 8.319, P < 0.0001 for body masschange(%)). However, this effect was limited to the first week after a switch to any new diet treatment, as indicated by the significant interaction between Diet and Week for body masschange(%) (F4,115 = 3.646, P < 0.01) (Fig. 2a), and the fact that body mass did not differ between weeks 1 and 2 (F1,119 = 0.592, P = 0.443). An animal's previous diet had no effect on body mass variables (F4,119 = 0.192, P = 0.942 for body mass; F4,115 = 1.286, P = 0.280 for body masschange(%)), further indicating that body mass changes occurred as short-term responses to switches in diet regime, but thereafter stabilized.
The DMI, ADIAI, DMD, and feeding rate of rabbits also differed across diet regimes (F4,123 = 57.467, P < 0.0001; F4,123 = 1814.718, P < 0.0001; F4,59 = 3.719, P < 0.01; and F4,47 = 22.545, P < 0.0001, respectively). DMI was lower for H than all other diets (Bonferonni post hoc P < 0.0001), and low for L relative to pelleted diets, although this latter difference was not always significant (P = 0.089 for G; P = 0.021 for GR; and P = 0.068 for GRS) (Fig. 2b). As expected, a trend of increasing ADIAI was found across diets with increasing levels of abrasiveness, with lowest ADIAI for L, followed by G, then GR, and finally highest for GRS (P < 0.0001 in all cases) (Fig. 2c). On GRS, approximately 70% of ADIAI was due to external abrasives (sand). On H, rabbits experienced similar ADIAI levels as for GR (P = 0.999). These patterns for DMI and ADIAI were consistent across both weeks of feeding trials (week effect F1,123 = 2.158 and 1.424, P = 0.144 and 0.235, respectively). The effect of diet on DMD (Table 1) probably occurred because of higher digestibility of H than G (P < 0.01). Rabbits required more time to ingest H than all other diets (P < 0.0001 in all cases) (Fig. 2d).
Effects on Tooth Length, Wear, and Growth
Diet had a significant influence on TL (F4,60 = 9.203, P < 0.0001 for I1; F4,360 = 65.391, P < 0.0001 for upper cheek teeth; F4,297 = 38.834, P < 0.0001 for lower cheek teeth) (Fig. 3a,c,e). Only in the I1 (F4,60 = 0.944, P = 0.445) did diet have no influence on TL (Fig. 3a). In I1 TL was lower for H than for other diets (Bonferroni post hoc P < 0.0001 to <0.001), whereas in cheek teeth TL for GRS was lower compared with other diets (P < 0.0001). TL was also lower for H and GR compared with L and G diets in upper and lower cheek teeth (P < 0.0001–0.032), indicating an effect of an increased level of internal abrasives.
TL was not influenced by the previous diet (F4,119 = 2.429, P = 0.052 for I1; F4,104 = 0.575, P = 0.681 for I1; F4,72 = 2.111, P = 0.088 for p3), and although there were significant diet × previous diet interactions in some cases (F15,104 = 2.669, P < 0.01 for I1; F15,72 = 2.827, P < 0.01 for p3), these seldom revealed different effects within diets. Tooth position also had an influence on TL of cheek teeth, but only in the upper jaw (F4,360 = 920.588, P < 0.0001; lower jaw: F3,297 = 1.473, P = 0.222), with P4 having highest TL, followed by M1, P3, M2, and then P2 (P < 0.0001 in all cases). Actually, a significant diet × tooth interaction for upper cheek tooth TL (F16,360 = 2.031, P = 0.011) was found, indicating that tooth position influenced not only TL, but also mediated effects of diet on TL: for the GRS diet, TL was not significantly different from GR or H in P2 and M1 (P = 0.092–0.999), from GR in P3 (P = 0.353), from H in P4 (P = 0.931), and from G, GR, and H in M2 (P = 0.451–0.999).
Despite the fact that TL was not consistent across diets, measured tooth growth was positively and significantly related to tooth wear in I1 (b = 0.564, t111 = 8.796, P < 0.0001; Fig. 4b), I1 (b = 0.603, t104 = 5.371, P < 0.0001; Fig. 4c), and p3 (b = 0.515, t57 = 8.833, P < 0.0001; Fig. 4d). The 95% confidence intervals for the slopes of these relationships, even accounting for repeated measures across individuals in the models, were always less than 1.0. This indicates that compensatory growth was not exact (hypothesis 1), and hence diet effects on TL were evident (hypothesis 2). Actually, both growth (I1: F4,98 = 27.407, P < 0.0001; p3: F4,59 = 15.870, P < 0.0001) and wear (I1: F4,103 = 67.109, P < 0.0001; p3: F4,50 = 53.322, P < 0.0001) were influenced by diet in similar—but inverse—ways as compared with TL. That is, in I1, growth and wear were highest for H, followed by GRS, and lowest for L, G, and GR (Bonferroni post hoc P < 0.0001 to <0.01), whereas in p3, GRS was associated with higher values than all other diets (P < 0.0001 in all cases) (Fig. 4a). Differences in diet effects between incisors (most wear occurred on H diets) and p3 (most wear occurred on GRS diets) reflect predicted effects of gnawing (hypothesis 3a) versus chewing (hypothesis 3b) on the distribution of abrasive pressures throughout the mouth. Diet also had a significant effect on growth and wear of I1 (F4,110 = 9.824, P < 0.0001; and F4,93 = 9.730, P < 0.0001, respectively), with H associated with fastest growth (P < 0.0001–0.010) and L and G diets associated with the least wear (P < 0.0001–0.038); wear on GR was significantly higher than that on L (P = 0.011). Interestingly, the time exposed to a particular diet influenced growth and wear in different ways, at least for p3: growth was faster in week 1 than week 2 on a particular diet (F1,59 = 14.293, P < 0.001), whereas wear was greater in week 2 (F1,50 = 8.920, P = 0.004).
Further evidence for differences in diet effects on wear in incisors versus cheek teeth was revealed by comparing relationships between wear and ADIAI (abrasiveness) or intake rate across teeth. For I1, both ADIAI and intake rate had significant positive effects on wear (b = 0.024, t44 = 2.495, P = 0.016; and b = 0.024, t44 = 6.195, P < 0.0001, respectively), whereas only ADIAI had a significant effect on p3 wear (b = 0.110, t13 = 4.309, P < 0.001; intake rate: b = 0.005, t13 = 0.291, P = 0.776). In I1, however, relationships between wear with ADIAI and intake rate were never significant (b = −0.006–0.009; P = 0.171–0.983).
Differences in diet effects across tooth positions (diet × tooth interaction described above) indicate that a wear gradient may be evident along the cheek tooth row (hypothesis 3b). Analysis of TL data presented as a percentage of the value recorded for diet L (Fig. 3b,d,f) revealed a significant diet × tooth interaction (upper cheek teeth only: F12,285 = 1.902, P = 0.034; tooth effect for lower jaw: F3,234 = 0.195, P = 0.900) that is useful for understanding the hypothesized wear gradient. This interaction occurred because only on the most abrasive diet, GRS, did this relative TL measure differ significantly between teeth (P = 0.195–0.999 across teeth for all other diets), with P3 having lower values than M1 and M2 (P < 0.001–0.028).
TSP and TSF scores were significantly higher than zero for most teeth (t79 = 2.237–18.158, P < 0.0001–0.028) (Supplementary Materials). Only P2 and p3 had TSP scores not different from zero (t79 = 1.000, P = 0.320 and t79 = 1.929, P = 0.057, respectively). For TSF scores, only P2 did not differ from zero (t79 = −1.000, P = 0.320), whereas p4 never showed abnormalities in TSF and so the data had zero variance. These four cases were thus omitted from further analyses of dental abnormalities. For TA, however, the only teeth to present scores different from zero were P2 (<0, t79 = −2.587, P = 0.012) and p3 (>0 in the lower jaw, t79 = 2.867, P < 0.01) (Supplementary Materials). TA scores were not different from zero in P3 (t79 = −1.348, P = 0.181), P4 and M2 (t79 = −1.000, P = 0.320), and p4 (t79 = 1.000, P = 0.320), while M1, m1, and m2 never presented abnormalities. Hence, for TA scores, only data for P2 and p3 were retained in further analyses, and no diet × tooth interaction could be tested statistically for this measure.
Whereas TSP scores in the upper cheek teeth were only relevant on GRS, the lower cheek teeth had TSP scores on all diets (Supplementary Materials). Diet effects were significant for TSP scores (upper jaw F4,297 = 13.759, P < 0.0001, with GRS diets differing from all others, Bonferroni post hoc P < 0.0001; lower jaw F4,218 = 3.981, P < 0.01, with GR being higher than H, Bonferroni post hoc P < 0.01). There was no significant effect of tooth in the upper jaw (F3,297 = 1.264, P = 0.287), but in the lower jaw TSP scores were higher for m2 than p4 and m1 (F2,218 = 3.981, P < 0.01; Bonferroni post hoc P < 0.0001). The diet × tooth interaction term was never significant for TSP scores (upper jaw F32,285 = 1.248, P = 0.293; lower jaw F8,210 = 1.417, P = 0.191).
In contrast with TSP, TSF scores (Supplementary Materials) were not significantly affected by diet (upper jaw F4,285 = 2.202, P = 0.069; lower jaw F4,210 = 1.175, P = 0.323), but TSF scores did differ across teeth (upper jaw F3,285 = 77.764, P < 0.0001; lower jaw F2,210 = 5.711, P < 0.01) with both M2 and m2 having the highest scores (P < 0.0001–0.032). However, the tooth effect was not consistent across diets (diet × tooth interaction for upper jaw F12,285 = 2.824, P < 0.01; and lower jaw F8,210 = 2.098, P = 0.037). This occurred in the upper jaw because the difference between M1 with M2 was only significant for diets L, GR, and H (P < 0.0001–0.050), and on GRS there were in fact no differences across teeth (P = 0.137–0.999). Thus, diet did seem to influence TSF scores to some extent; whereas different teeth generally had different scores, the most abrasive diet (GRS) made all teeth similar. In the lower jaw, the difference between m2 with m1 and p3 was only significant for diet H (P < 0.01)—for all pelleted diets, teeth did not differ in TSF scores (P = 0.305–0.999).
TA scores (for which only P2 and p3 had scores different from zero, Supplementary Materials) differed across diets in both the upper (F4,60 = 9.615, P < 0.0001) and lower jaw (F4,60 = 5.625, P < 0.001), in that GRS diets were typically different from all other diets (P < 0.0001 to <0.01). TA scores for GRS were only not different from one other diet—GR, and then only in the lower jaw (P = 0.252).
Waviness—a proxy for overall abnormality in dentition—differed significantly across diets (F4,60 = 5.863, P < 0.001), with scores for GRS being higher than for all other diets (Bonferroni post hoc P < 0.001–0.012) (cf. Fig. 1b and Supplementary Materials).
This study confirms relevant effects of internal (dietary) and external abrasives on tooth wear, with important differences between the two functional tooth groups (incisors and cheek teeth), but also within the cheek tooth row. The observed patterns allow detailed interpretations on the process of tooth wear, including differences between maxillary and mandibular teeth. In our model species, tooth wear and growth were positively correlated, but correlation was not completely compensatory, which led to treatment-specific differences in TL. The growth rate for rabbit cheek teeth measured in this study was higher than previously reported in the literature.
There was an obvious change in body weight when feeding rabbits grass hay only. Most rabbits initially lost weight during this feeding period, although hay was provided ad libitum. In the second week of hay feeding, however, rabbits on average maintained their (lower) body mass (Fig. 1a). When fed pelleted diets, animals mostly gained weight. A dislike of the lucerne diet (L) was evident, which was already reported previously (Wolf et al., 1999). In order to achieve isocaloric and isonitrogenous feeds, L had to comprise the highest proportion of lignocellulose (Table 1), which may have reduced their acceptance. Surprisingly, the inclusion of sand did not affect the acceptance of diet GRS.
Tooth growth compensates for wear; therefore, we expect tooth length to be relatively constant across diets and growth tightly correlated with wear.
We found a positive and significant relation between tooth growth and wear in upper incisors, lower incisors and p3, but compensatory growth was not exact, therefore resulting in diet-related different total TLs. In this study, effects of diet on teeth were monitored over constant periods of 2 weeks. We cannot address the question whether, over longer periods of time, growth would have compensated completely for wear, so that differences between diets would have disappeared (assuming an absolute length-set point). Alternatively, the detected differences in TL between diets in this study could have been an effect of an even shorter mismatch between growth and wear within a few initial days, after which the length measured in our study remains constant (assuming a relative growth/wear-setpoint). The finding that p3 growth was faster in week 1 than week 2, whereas wear was greater in week 2, could suggest that a stable equilibrium between wear and growth was not yet reached for the cheek teeth in the course of our experiment; however, this result could also reflect numerical differences in food intake between the weeks (Fig. 2b), with the second week representing a more or less stable equilibrium. The fact that for incisors, no difference in wear and growth between the weeks was evident, could represent evidence for a relative growth/wear-setpoint, indicating that absolute TL should vary (within boundaries) among rabbits on different diets. Fluctuations in incisor growth and wear over time in animals fed a constant diet documented by Wolf and Kamphues (1996), however, suggest that the balance between wear and growth represents a dynamic equilibrium.
Until now most authors assumed that growth and wear of cheek teeth are noticeably lower than that of incisors, and suggested a growth rate of 0.5–0.7 mm/week in cheek teeth (Table 2). The only exception were von Koenigswald and Golenishev (1979), who measured growth rates of cheek teeth by staining the growth-zone of the tooth with tetracycline and measured a growth rate in rabbit cheek teeth of 1.1–1.3 mm/week. In our study, growth rates of cheek teeth varied between 1.37–3.23 mm/week. These findings underline that growth of rabbit teeth is flexible, and that variation in response to wear should be expected.
|Study||Method||Diet||Upper incisor||Lower incisor||Lower cheek teeth|
|Shadle (1936)||Tooth mark||nm||2.0||2.4|
|Spannbrucker et al. (1977)||nm||nm||2.1–2.3||2.1–2.3|
|von Koenigswald and Golenishev (1979)||Enamel staining||nm||2.5||2.7||1.1–1.3|
|Lobprise and Wiggs (1991)||nm||nm||2||2.4|
|Wolf and Kamphues (1995)||Tooth mark||Carrots||1.68||1.61||1.64||1.45|
|Jekl and Redrobe (2013)||nm||nm||2–4||2–4||2–4||2–4||0.7–0.93||0.7–0.93|
|This study||Tooth marka||L||1.54||1.27||1.72||1.57||1.47||1.23|
Small, detectable differences in wear between diets reflect diet abrasiveness and/or chewing activity.
Differences in TL, wear and growth can be linked to diet abrasiveness in this study for incisors and cheek teeth. We could not only confirm an effect of external abrasives, with TL mostly shorter/wear higher on diet GRS (containing sand) than on other diets, but also an effect of internal abrasives, where diets GR and H resulted in shorter cheek teeth than diets L and G, and wear on diet GR was higher in lower incisors than diet L. In the course of the debate whether internal abrasives (phytoliths) can actually cause relevant tooth wear and, correspondingly, drive evolutionary adaptations to this wear (see Introduction), our study thus provides evidence that internal abrasives do have a measurable effect on dental tissue. Differences in 3D surface texture between diets of different levels of internal abrasives (Schulz et al., 2013) thus will also translate into differences in wear. Whether the effects of internal abrasives can, in nature, ever act separately from the effect of external abrasives, remains debatable to date in particular in the context of grazing, with grasses containing high levels of phytoliths but also presumed to be particularly prone to grit contamination (Damuth and Janis, 2011). In ungulates, the finding that the interspecific correlation of hypsodonty with a habitat (i.e., external abrasives) proxy (precipitation) is not improved if mesowear (i.e., an internal abrasives proxy) is included in the analysis (Kaiser et al., 2013), could indicate that hypsodonty is mainly a reaction to external abrasives, the effects of internal abrasives notwithstanding. In primates, the finding that internal abrasives are related to enamel thickness (Rabenold and Pearson, 2011) might indicate that internal abrasives can also exert a selective pressure. For conclusive results, internal and external abrasives would have to be measured in the same forage samples to judge whether situations of high internal but low external abrasives (as suggested by Walker et al., 1978) occur frequently.
Functional differences between incisors and cheek teeth lead to different wear and growth on different diets.
- Incisors are worn more heavily when feeding whole hay that needs more gnawing as compared to pellets.
- Cheek teeth, with a chewing action more independent from whether the diet is offered whole or pelleted, are worn more heavily with increasing dietary abrasiveness; external abrasives (sand) lead to a gradient in wear along the maxillary cheek tooth row whereas increased internal abrasives (phytoliths in rice hulls) do not lead to such a gradient (Taylor et al., 2013).
This hypothesis was supported, as food intake rate (a proxy for chewing activity) was a significant factor for upper incisor wear, but not for cheek teeth or lower incisors. In incisors, differences in the chewing movement for whole forage and other feeds have been documented (Fig. 4 in Weijs and Dantuma, 1981), with a prolonged posterior–anterior movement of the lower against the upper incisor on hay, which should increase both attrition and abrasion. In contrast, abrasives intake alone was correlated to wear of cheek teeth, indicating that chewing action of cheek teeth is more independent of whether the diet is offered whole or pelleted. This latter finding contradicts a hypothesis of Crossley (2003, Fig. 6), who suggested that mandibular grinding chewing movements are reduced in their extent on pelleted feeds as compared to natural vegetation (based on Weijs et al., 1989, who did not report the composition of the pelleted diet they used), which should lead to less and more uneven wear of cheek teeth on pelleted diets. Weijs and Dantuma (1981) found that chewing muscle activity when chewing pellets did not differ significantly from that of chewing hay (but was lower for chewing carrots).
Corresponding to the different influence factors for the wear of incisors and cheek teeth, growth rates for incisors and cheek teeth were independent from each other, which requires a tooth-specific mechanism for adjusting growth to the tooth-specific wear. A gradient in the cheek tooth row developed particularly on GRS (external abrasives); a similar numerical effect of GR (internal abrasives only) (Fig. 3d) was not significant. Taylor et al. (2013) speculated that external abrasives might affect the front cheek teeth more, because they should be present on the outside of ingested plant material but become mixed into the ingesta bolus when passing further along the oral cavity towards the hind cheek teeth. This explanation appears inapplicable in our experimental setup, because the external abrasives (sand) were homogenously mixed into the pelleted diet GRS. The alternative explanation of Taylor et al. (2013) is that due to the geometry of mandibular movement, grinding movements should be more pronounced at tooth positions distant from the mandibular joint, that is, in the front cheek teeth, which might therefore experience higher wear. This explanation could also apply to our findings, although the fact that the P3, not the P2 was most heavily affected, remains unexplained. One could speculate on general differences in dental hardness, or on an effect of the diastema, which allows ingesta to evade the action of the P2 more by slipping off into the diastema than that in the case of the P3 where the preceding premolar prevents such “slipping.”
The finding that for both incisors and cheek teeth, the most pronounced wear effect (triggered by chewing or abrasives, respectively) affects the maxillary teeth more than the mandibular ones might appear counter-intuitive, because the relative movement of teeth against each other (attrition) should not affect any position disproportionately. There is in fact ample evidence for ungulates that lower molars have more rounded cusps (more abraded attrition facets) than upper molars (Fortelius and Solounias, 2000; Franz-Odendaal and Kaiser, 2003; Kaiser and Fortelius, 2003), suggesting relatively more abrasion. Kaiser and Fortelius (2003) hypothesized that this difference might be simply due to the effect of gravity, but that explanation clearly does not fit the present results in rabbits. An ad hoc explanation of the relatively more pronounced abrasion of the upper teeth that we found in the rabbit experiment could be the following: ingesta rests, due to gravity, on the lower teeth, with which it is then moved against the upper dentition. The naturally occurring tooth spurs on the rabbits' lower cheek teeth (Supplementary Material) will contribute to holding the ingesta in place. While remaining more stationary in relation to the lower teeth, the ingesta is then moved across the upper occlusal surface for grinding, resulting in relatively more ingesta movement along the upper than the lower occlusal surface during the chewing movement (Weijs and Dantuma, 1981), which leads in turn to more abrasion affecting the maxillary teeth. In other words, we suggest, for the rabbit, a special case of the “inverted pestle-and-mortar” system (see Lucas, 1979 for usage of this terminology), where the mandibular teeth represent a special kind of “pestle” both holding the food in place and grinding it along the “mortar” of the maxillary teeth. If this interpretation were correct, it would have to apply to the incisors, too, assuming that the relative movement of ingesta during their cutting action is also more pronounced in relation to the upper than to the lower teeth. Actually, anatomical drawings (Weijs and Dantuma, 1981) and photographs of rabbit front teeth in Crossley (2003) indicate a concave occlusal surface of the maxiallary incisor, suggesting that a similar “inverted pestle-and-mortar” mechanism may be present in the incisors, too.
On the other hand, the observed wear patterns could be related to systemic differences between the upper and lower teeth themselves, such as a structural difference in the histology of dental tissues. A general explanation for why lower teeth should be more resistant to wear than upper teeth was in fact proposed by Fortelius (1985, p. 58f, Figs. 3 and 4) for ungulates: Because of the geometry of the occlusal facets, the direction of the main cutting edges of lower teeth is transversely to the enamel prisms, whereas on the upper teeth, their direction is along the prism axes, probably allowing relatively more tissue loss. Whether this explanation can be expanded to rabbits remains to be demonstrated. In rabbit cheek teeth, two forms of enamel have been described, a basal radial (more wear-resistant) and an outer irregular (less wear-resistant) enamel (von Koenigswald et al., 2010). Their respective distribution on upper and lower cheek teeth remains to be quantified, but in combination with the mentioned “inverted pestle-and-mortar” mechanism, the less-wear resistant irregular enamel might be particularly worn down on the mandibular cheek teeth. In the incisors, where the single-layered enamel structure is uniform between the upper and lower position (von Koenigswald, 1996), differences in wear appear more likely to be due to other causes than enamel structure. More detailed knowledge on the relative movement of ingesta along dental surfaces, and about the histological directional anatomy of dental surfaces is required to test these hypotheses.
Abnormal tooth wear will occur more frequently with excessive external abrasives (sand), and (according to 3b) affect the cheek teeth according to their position in the tooth row (anterior ones more affected).
We found significantly more dental abnormalities, including general tooth spurs (TSP), changes in the occlusal TSF, abnormal TAs and stepmouth formation, when feeding GRS pellets. This suggests that GRS not only increased general tooth wear but also the frequency of localized insults. However, in contrast to our prediction, the effects were not clearly more distinct in the anterior cheek teeth: whereas the anterior cheek teeth were affected for TA, this was not the case for TSP. Additionally, TSF abnormalities also affected the posterior cheek teeth. These findings give further evidence that abrasives (whether internal or external) did not change their position in the ingesta bolus in this study (the first hypothesis from Taylor et al., 2013 as explained above), and also do not correspond to presumed differences in tooth hardness, but suggest that differences in a tooth wear gradient due to external abrasives are rather linked to effects of differences in chewing movements between tooth positions (the second hypothesis from Taylor et al., 2013 as explained above).
Rabbits usually use their cheek teeth in a lateral figure-of-8 type pattern, which brings multiple cheek teeth in occlusion. It has been suggested than when feeding on pellets or grains, the vertical phase of the chewing action is more pronounced and the lateral one reduced, altering the pattern of tooth contact (Crossley, 2003; Lord, 2011). This could lead to dental abnormalities, because rabbits with abnormal chewing pattern do not efficiently wear down the whole occlusal surface. However, the results of this study do not support this hypothesis. Tooth spurs were not generally more frequent when feeding pellets than when feeding hay, and the wear of cheek teeth was not related to diet structure (see Hypothesis 2 above). While comprehensive tests for grain or pelleted feeds based on grain products remain to be performed (e.g., unpublished observations in Prebble and Meredith, 2014), our findings suggest that high-fiber pelleted feeds based on forages do not lead to more dental abnormalities in the short term than whole hay.
The ingestion of soil has been linked to gross dental abnormalities, such as stepmouth, in captive wild ruminants (Martin Jurado et al., 2008). Whether external abrasives cause such abnormalities or not will most likely depend on their particle size. At 0.233 mm (Table 1), the mean particle size of the sand used in this study was much larger than the dimensions usually given for phytoliths of 0.005–0.250 mm (reviewed in Strömberg, 2004). However, differences in the hypsodonty index and mesowear score in ungulates led Kaiser et al. (2013) to suggest that the grain size of the majority of external abrasives that affect herbivores must be at or even below the size of phytoliths. The effects of such external abrasives on teeth remain to be determined experimentally. It appears likely that such abrasives will not lead to dental abnormalities but only increase wear.
Our results underline that irrespective of the undoubted effects of external abrasives, internal abrasives have an effect on tooth wear and could therefore act as evolutionary drivers of dental adaptations; whether they do act in this manner will depend on ecological scenarios where external abrasives do not represent an overriding signal. In veterinary practice, many clinicians believe in a constant tooth growth that needs to be worn down to prevent tooth elongation and malocclusion. We found that growth and wear of teeth vary, depending on intake of internal and external abrasives, and that wear and growth are tightly correlated. It is unclear how regulation between growth and wear takes place, and elucidating the associated feedback mechanisms represents a promising area of future research. Based on these findings, it is our opinion that diet alone may be less likely to cause dental problems in pet rabbits, due to flexible growth that reacts to wear. Other causes such as mineral imbalances or genetics should be considered when diagnosing dental disease, and a minerally balanced diet and breeding hygiene (not allowing affected animals to reproduce) may be the most promising prophylactic approaches. In cases where incisor overgrowth has occurred repeatedly, offering whole forages rather than pelleted diets appears the more promising strategy, but irrespective of differences in the absolute wear of the upper incisors, the use of forage-based pellets did not cause more abnormal wear than feeding whole hay in this study. Whether extreme diets, such as energy-dense concentrates of which small amounts meet energy requirements and that need not be chewed intensively (e.g., grain-based pellets) can have more deleterious effects due to an absence of wear than the diets of this study remains to be investigated in healthy animals.
We thank Anja Tschudin and Dietmar Ranz for support in pelleted diet formulation, Urs Müller and Luciano Schmid for support in the construction of the husbandry facilities, veterinary assistants Sandra Mosimann and Monique Lauer for helping with anesthesia, endoscopy, and husbandry, Mike Patthey for running the CT scans, Jennifer Kunz, Irene, Samuel, Emilia and Daniel Clauss, and Fred and Manu Elsemann for support in animal husbandry, and special thanks to Prisca Müller and Nicole Imhof for helping in feces collection. We thank Daniela Kalthoff and an anonymous reviewer for comments on the manuscript.