Epineurial adipocytes are dispensable for Schwann cell myelination


Address correspondence and reprint requests to Roman Chrast, Department of Medical Genetics, University of Lausanne, 1005 Lausanne, Switzerland.

E-mail: roman.chrast@unil.ch


Previous clinical observations and data from mouse models with defects in lipid metabolism suggested that epineurial adipocytes may play a role in peripheral nervous system myelination. We have used adipocyte-specific Lpin1 knockout mice to characterize the consequences of the presence of impaired epineurial adipocytes on the myelinating peripheral nerve. Our data revealed that the capacity of Schwann cells to establish myelin, and the functional properties of peripheral nerves, were not affected by compromised epineurial adipocytes in adipocyte-specific Lpin1 knockout mice. To evaluate the possibility that Lpin1-negative adipocytes are still able to support endoneurial Schwann cells, we also characterized sciatic nerves from mice carrying epiblast-specific deletion of peroxisome proliferator-activated receptor gamma, which develop general lipoatrophy. Interestingly, even the complete loss of adipocytes in the epineurium of peroxisome proliferator-activated receptor gamma knockout mice did not lead to detectable defects in Schwann cell myelination. However, probably as a consequence of their hyperglycemia, these mice have reduced nerve conduction velocity, thus mimicking the phenotype observed under diabetic condition. Together, our data indicate that while adipocytes, as regulators of lipid and glucose homeostasis, play a role in nerve function, their presence in epineurium is not essential for establishment or maintenance of proper myelin.

Abbreviations used

motor nerve conduction velocity


phosphate-buffered saline


peroxisome proliferator-activated receptors

Peripheral nerves are composed of three distinct tissue compartments: the epineurium, perineurium, and endoneurium. A substantial amount of studies have focused on the endoneurial compartment of the peripheral nerve, harboring Schwann cells surrounding axons, which communicate information between the central neural system and the periphery. The role of both perineurium and epineurium remains less clear. The perineurium, which is mostly composed of fibroblasts, is thought to play a role as a diffusion barrier between the endoneurium and the epineurium (Parmantier et al. 1999; Pina-Oviedo and Ortiz-Hidalgo 2008). The adipocyte-rich epineurium was previously suggested to play a role in the mechanical protection of the nerve from compression damage or to serve as a source of fatty acids for the endoneurium (Barkmeier and Luschei 2000; Verheijen et al. 2003; Moayeri and Groen 2009). Epineurial adipose tissue may also function as an endocrine organ (Deng and Scherer 2010). Inactivation of either of the above mentioned epineurial adipocyte functions may potentially lead to defects in PNS function.

Similar to the situation in adipocytes, strictly controlled lipid metabolism plays a crucial role in glial cell ability to produce and maintain myelin membrane (Chrast et al. 2011). Interestingly, a previously described mutation in the Lpin1 gene present in ‘fatty liver dystrophy’ (Lpin1fld/fld) mice affects both glial cells and adipocytes, leading to a demyelinating peripheral nerve neuropathy and epineurial lipoatrophy (Langner et al. 1991; Peterfy et al. 2001; Verheijen et al. 2003). We previously demonstrated that Schwann cell-specific Lpin1 inactivation contributes to the development of pronounced peripheral neuropathy (Nadra et al. 2008). Here, we use aP2Cre/+/LpfEx2-3/fEx2-3 mice, in which Lpin1 is deleted in adipocytes, to address whether the compromised function of epineurial adipocytes also contributes to the neuropathy observed in Lpin1fld/fld mice. In addition, we have used recently generated Sox2Cre/+/PPARγemL–/L– mice, which completely lack adipocytes as a consequence of epiblastic deletion of peroxisome proliferator-activated receptor gamma-γ (Pparg) (Nadra et al. 2010), to evaluate the consequence of lipoatrophy present in this model on myelinated Schwann cells. Our characterization of these two models shows that proper function or presence of epineurial adipocytes is not critical for establishment or maintenance of Schwann cell myelin.



BALB/cByJ-Lpinfld/+ mice were obtained from the Jackson Laboratory. Generation of Sox2Cre/+/PPARgemL–/L– mice have been previously described (Nadra et al. 2010). To generate mice with adipocyte-specific Lpin1 inactivation, the Lpin1fEx2-3/fEx2-3 mice (Nadra et al. 2008) were crossed with aP2Cre transgenic mice (He et al. 2003). The details of the characterization of these mice are described elsewhere (Nadra et al. 2012, under review). Experiments were performed in accordance with the legal requirements of the University of Lausanne and of the Canton of Vaud (Switzerland).

Motor nerve conduction velocity

All animals were anesthetized with a mixture of 10 mL/g Ketanarkon 100 (1 mg/mL; Streuli, Uznach, Switzerland) and 0.1% Rompun (Bayer, Zurich, Switzerland) in phosphate-buffered saline (PBS). The left and right sciatic nerves were stimulated at the sciatic notch and distally at the ankle via bipolar electrodes with supramaximal square-wave pulses (5 V) of 0.05 ms. The latencies of the compound muscle action potentials were recorded as previously described (de Preux Charles et al. 2010). Results were expressed as the mean and SD.

Quantitative PCR

Sciatic nerves were harvested at 3 months of age and placed in ice-cold PBS (pH 7.4). The perineurium and epineurium were gently dissected away from the endoneurium along the whole length of the nerve as previously described (Verheijen et al. 2003). Total RNA isolation, cDNA synthesis, and relative quantitative RT-PCR were performed as previously described (de Preux et al. 2007). Results were normalized using the reference gene Ubiquitin. Information concerning sequences of oligonucleotides used for quantitative PCR is available upon request.

Toluidine blue staining

Mice were perfused with 1% Paraformaldehyde and 2% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.3) for 5 min. Sciatic nerves were dissected and post-fixed by immersion in the fixative solution for 2 h at 4°C, washed in 0.1 M cacodylate buffer, and osmicated for 4 h in 1% OsO4 (Sigma-Aldrich Chemie, Buchs, Switzerland). Nerves were rinsed in water, dehydrated, and embedded in epon 812 resin (Fluka). One-micrometer sections were stained with 1% toluidine blue and examined by light microscopy.

Oil-red-O and Nile red stainings

A fresh working solution of Oil-red-O (Sigma-Aldrich Chemie, Buchs, Switzerland) was prepared by dilution of the Oil-red-O stock solution (5 g/L in 98% isopropanol) in distilled water at a ratio of 3 : 2. The working solution was allowed to stand for 10 min after mixing and was filtered with a 0.45-μm pore-sized filter. Subsequently, sciatic nerves were shortly washed in PBS, stained with the filtered working solution of Oil-red-O for 10 min, and washed for 10 min in demineralized water.

Sciatic nerve cross-sections were mounted in Nile red (Sigma) solution (0.5 mg/mL in acetone) diluted 1000x in 75% glycerol and visualized using a Zeiss Axioplan 2 microscope with an AxioCam MRc camera and AxioVision release 4.5 software (Carl Zeiss, Feldbach, Switzerland).


Tail vein blood glucose was determined using a glucometer Ascencia Contour (Bayer).


PNS myelination is not affected by compromised adipocyte function

Our previous data suggested that the defects in epineurial adipocytes in Lpin1fld/fld mice potentially contribute to their peripheral neuropathy phenotype (Verheijen et al. 2003). To evaluate this hypothesis, we crossed previously generated Lpin1fEx2-3/fEx2-3 mice (Nadra et al. 2008) with aP2Cre transgenic mice (He et al. 2003). aP2Cre-mediated deletion of exons 2 and 3 results in loss of lipin 1 function in adipose tissue after the formation of fat depots, thus not affecting early steps of adipogenesis (detailed characterization of these mice is available in Nadra et al. 2012, under review). The Oil-Red-O staining of sciatic nerves from 3-month-old animals revealed that the amount of adipocytes present in the epineurium was not affected in aP2Cre/+/LpfEx2-3/fEx2-3 mice as compared to control mice (Fig. 1a). We therefore analyzed the expression of selected adipocyte markers in peri/epineurium of sciatic nerves from aP2Cre/+/LpfEx2-3/fEx2-3 and control mice. As expected, we observed a substantial reduction of Lpin1 expression in epineurial adipocytes (Fig. 1b). Importantly, the change in Lpin1 expression leads to molecular changes in epineurial adipocytes as indicated by decreased expression of fatty acid-binding protein 4 (Fabp4/aP2), which is a marker of adipocyte differentiation (Rosen et al. 2000), and diacylglycerol acyltransferase 1 (Dgat1), which catalyzes the final step of triacylglycerol synthesis (Yen et al. 2008) (Fig. 1b). To evaluate the possible consequence of these changes on endoneurial Schwann cells, we first performed a structural characterization of the PNS. Our histological analysis revealed that myelination was not disturbed in aP2Cre/+/LpfEx2-3/fEx2-3 sciatic nerves (Fig. 2a). In line with these observations, the electrophysiological evaluation of sciatic nerve function measured at different developmental stages did not reveal a reduction in motor nerve conduction velocity (MNCV) in aP2Cre/+/LpfEx2-3/fEx2-3 mice as compared to control mice (Fig. 2b). Interestingly, qPCR analysis performed using endoneurium isolated from 3-month-old aP2Cre/+/LpfEx2-3/fEx2-3 mice revealed that while mRNA levels of Lpin1 or of Krox20 and Krox24, transcription factors involved in regulation of myelination, did not change, the expression of myelin genes Mpz and Pmp22 was slightly increased (Fig. 2c). These observations indicated that while the compromised function of epineurial adipocytes induced by loss of lipin 1 function does not affect the capacity of Schwann cells to produce myelin, it may have an effect on Schwann cell myelin gene expression.

Figure 1.

Adipocyte-specific inactivation of Lpin1 affects adipocyte function in peripheral nerve peri-epineurium of aP2Cre/+/LpfEx2-3/fEx2-3 mice. (a) Oil-red-O staining of sciatic nerves from control (aP2+/+/LpfEx2-3/fEx2-3) and knockout (aP2Cre/+/LpfEx2-3/fEx2-3) mice. The black arrows indicate fat depots. (b) Quantitative PCR evaluation of Lpin1, Fabp4, and Dgat1 expression in the peri-epineurial compartment of sciatic nerve from 3-month-old control (aP2+/+/LpfEx2-3/fEx2-3) and knockout (aP2Cre/+/LpfEx2-3/fEx2-3) mice (= 3; *< 0.05).

Figure 2.

Adipocyte-specific Lpin1 inactivation does not alter myelination. (a) Nile Red (i, iii) and toluidine blue (ii, iv) stainings of crosssections of sciatic nerves from 3-month-old control (aP2+/+/LpfEx2-3/fEx2-3; i–ii) and knockout (aP2Cre/+/LpfEx2-3/fEx2-3; iii–iv) mice. (b) Motor nerve conduction velocity (MNCV) measurements in control and knockout mice at depicted age (n =6). (c) Quantitative PCR measurements of Lpin1, Krox24, Krox20, Mpz, and Pmp22 expression in the endoneurial compartment of sciatic nerve from 3-month-old control and knockout mice (n = 3; *p < 0.05).

Complete absence of epineurial adipocytes does not affect PNS myelination

The observed change in the expression of myelin genes in the endoneurium of aP2Cre/+/LpfEx2-3/fEx2-3 mice left open the possibility that Lpin1–/– adipocytes present in their epineurium are still able, at least on the transcriptional level, to influence endoneurial Schwann cells. We therefore analyzed PNS function in Sox2Cre/PPARγemL–/L– mice, which completely lack both white and brown adipocytes (Nadra et al. 2010). Similar to Lpin1fld/fld mice (Verheijen et al. 2003), 3-month-old Sox2Cre/+/PPARγemL–/L– animals also completely lack epineurial fat (Fig. 3a and b). However, contrary to Lpin1fld/fld mice, the structure of myelinated axons in Sox2Cre/+/PPARγemL–/L–did not reveal any significant changes (Fig. 4a). Also, the expression of myelin genes (Mpz and Pmp22) was not affected in the endoneurium of Sox2Cre/+/PPARγemL–/L– mice (Fig. 4b). Interestingly, despite the absence of detectable structural changes in their myelin, Sox2Cre/+/PPARγemL–/L– mice did show a reduction in their MNCV (Fig. 4c). This functional disability is likely the consequence of the presence of strong hyperglycemia (Fig. 4d) in Sox2Cre/+/PPARγemL–/L– mice, similar to the situation observed during diabetic peripheral neuropathy (de Preux Charles et al. 2010; Zenker et al. 2012).

Figure 3.

Absence of epineurial adipocytes in Sox2Cre/+/PPARγemL-/L- mice. (a) Oil-red-O staining of sciatic nerves from 3-month-old Sox2+/+/PPARγ+/+ (control, i), Sox2Cre/+/PPARγemL-/L- (mutant, ii), and Lpin1fld/fld mice (iii). The black arrows indicate fat depots. (b) Quantitative PCR measurement of PPARγ expression in the endoneurial (Endo) and peri-epineurial (P/E) compartments of sciatic nerve from control (Sox2+/+/PPARγ+/+) and conditional knockout (Sox2Cre/+/PPARγemL-/L-) mice (= 3; < 0.05).

Figure 4.

Morphological and functional characterization of peripheral nerve endoneurium in Sox2Cre/+/PPARgemL-/L- mice. (a) Nile Red (i–iii) and toluidine blue (iv–vi) stainings of cross-sections of sciatic nerves from Sox2+/+/PPARγ+/+ (control; i and iv), Sox2Cre/+/PPARγemL-/L- (mutant; ii and v), and Lpin1fld/fld mice (iii and vi). (b) Quantitative PCR measurements of Mpz and Pmp22 expression in the endoneurial compartment of sciatic nerve from Sox2+/+/PPARγ+/+ and Sox2Cre/+/PPARγemL-/L- mice (= 3). (c) Motor nerve conduction velocity (MNCV) measurements in Sox2+/+/PPARγ+/+ and Sox2Cre/+/PPARγemL-/L- mice (= 3; *< 0.001). (d) Measurements of glucose in the blood of Sox2+/+/PPARγ+/+ and Sox2Cre/+/PPARγemL-/L- mice (= 3). All analyses were done in 3-month-old mice.


Our data, generated through analysis of the aP2Cre/+/LpfEx2-3/fEx2-3 and Sox2Cre/+/PPARγemL–/L– knockout mice, present the first insight into the consequences of either disturbed epineurial adipocyte function or of complete adipocyte absence on peripheral nerve myelination.

Our analysis of PNS morphology and function in aP2Cre/+/LpfEx2-3/fEx2-3 knockout mice indicated that the functional integrity of epineurial adipocytes does not play a major role in myelination. This conclusion is further strengthened by the data from the analysis of Sox2Cre/+/PPARγemL–/L– mice. In accordance with the known requirement of PPARγ for adipocyte differentiation (Rosen et al. 2000), the PPARγ null mice have no BAT or WAT, thus providing a unique model to study the physiological consequences of a complete congenital lipodystrophy on peripheral nerve myelination. Contrary to previously described PNS deficits in Lpin1 complete knockout mice (Lpin1fld/fld), which also develops severe lipodystrophy (Langner et al. 1991; Verheijen et al. 2003), the Sox2Cre/+/PPARγemL–/L– mice show no detectable structural myelin abnormalities despite complete absence of epineurial fat.

It is possible that lipids originating from non-glial endoneurial cells and/or from the circulation compensate through horizontal lipid transfer for lipid-compromised epineurial adipocytes, as previously suggested in the situation of glia-specific lipid mutants (Saher et al. 2005; Verheijen et al. 2009). Alternatively, Schwann cells may increase their lipid synthesis to respond to reduced and/or abolished epineurial lipid input. Both of these possibilities will require further exploration.

The observed slight increase (~50%) in the level of expression of Mpz and Pmp22 in aP2Cre/+/LpfEx2-3/fEx2-3 may potentially have pathological consequences as, in humans, gene dosage variation affecting genes encoding myelin proteins leads to peripheral neuropathy (Lupski et al. 1991; Maeda et al. 2012). However, previous data from mice suggest that for Pmp22, approximately 100% over-expression and for MPZ approximately 80% over-expression is necessary to reproduce this phenotype (Magyar et al. 1996; Huxley et al. 1998; Wrabetz et al. 2000). The absence of the PNS phenotype in aP2Cre/+/LpfEx2-3/fEx2-3 mice, where the level of Pmp22 and Mpz mRNA over-expression is 50–60%, therefore, indicates that the level of Pmp22 and Mpz over-expression is not reaching the threshold necessary to induce a detectable PNS phenotype and/or that this over-expression is absent during the process of myelination which, in mice, is mostly completed during the first post-natal month. In addition, the fact that we observed increased Pmp22 and Mpz mRNA expression level in the absence of increased Krox20 levels provides supplementary evidence for the previously suggested role of Krox20-independent myelin gene activation (Parkinson et al. 2003).

Interestingly, we have detected a reduced MNCV in Sox2Cre/+/PPARγemL–/L– animals. A reduced nerve conduction velocity was previously observed in both type I and type II diabetic situations even before the presence of PNS structural changes (de Preux Charles et al. 2010; Zenker et al. 2012). It is therefore possible that the reduced MNCV, which is present in Sox2Cre/+/PPARγemL–/L– mice and not accompanied by any detectable nerve structural changes, is a consequence of their substantial hyperglycemia.

Although we cannot formally exclude that defective or missing epineurial adipocytes in aP2Cre/+/LpfEx2-3/fEx2-3, and Sox2Cre/+/PPARγemL–/L– mice, respectively, do not lead to subtle changes in lipid composition of peripheral nerve myelin, the comparison between the data presented in this study and previously observed strong myelin defects in peripheral nerves of animals with Schwann cell-specific lipid defects (Nadra et al. 2008; Verheijen et al. 2009) further strengthen the concept that myelinating glial cells in the PNS predominantly rely on local, cell autonomous, lipid metabolism.


We thank Jean-Christophe Stehle and Jean Daraspe for technical assistance. We thank Patrick Gouait for his help in animal maintenance. This work was supported by grants from the Swiss National Science Foundation to R.C. (grants PP00P3_124833 and 31003A_135735/1) and to B.D. (grant 31003A_135583/1). The authors have no conflicting financial interests.