Effect of chronic intermittent hypoxia (CIH) on neuromuscular junctions and mitochondria in slow- and fast-twitch skeletal muscles of mice—the role of iNOS

Body weightEffect of CIH in WT

In WT mice, 6 weeks of CIH compared to NOX significantly reduced the pre- to post-interventional gain in body weight, i.e., the weight gain between 2 and 3.5 months of age (starting from a 9.1% higher baseline) (Table 1).

Table 1 Weight (g) and weight changes (%) of WT and iNOS−/− mice pre- to post-intervention. Values are given as mean+SEM, n = 6 to 14 animals per group. * p < 0.05, significance between CIH and NOX; ##p < 0.01, ###p < 0.001, significance between WT and iNOS−/−Effect of iNOS w/o CIH

Under NOX conditions, iNOS−/− mice showed a significantly lower weight gain compared to WT (Table 1), notably, starting from by 10.7% (p < 0.001) higher pre-interventional body weight than WT mice. When exposed to CIH, iNOS−/− mice even showed a weight loss and thus differed significantly from NOX conditions compared to NOX (Table 1).

Fiber morphology and fraction in soleus muscleEffect of CIH in WT

The mean CSA covering all fiber types in the soleus muscle did not significantly differ between CIH intervention and NOX in WT (Fig. 1A - B; E, left), though it correlated significantly and positively with the weight gain (r = 0.41; p = 0.047) in WT of both interventional groups. A fiber type-specific analysis showed, that CIH compared to NOX led to a significant decrease in fiber CSA exclusively of type 2a fibers by 13.5% (p < 0.05) (Fig 1E, right). Thereby, CIH as compared to NOX resulted in an increase in type 1 fiber fraction (37% vs. 31%; p < 0.01) in association with a decrease in type 2a fiber fraction (44% vs. 53%; p < 0.001) at unchanged type 2x fiber fraction (Fig. 1A - B; G).

Fig. 1figure 1

Representative images (200-fold magnification, scale bar = 50 μm) of soleus muscle cross sections stained for ATPase after preincubation at pH 4.55 of WT-NOX mouse (A), WT-CIH mouse (B), iNOS−/−-NOX mouse (C), and iNOS−/−-CIH mouse (D). Type 1 fibers are stained dark (white arrow), type 2a fibers are stained light-colored (black star), and type 2x fibers are stained intermediate (black arrowhead). E Overall myofiber CSA [μm2] and F CSA of type 2a myofibers in the soleus muscle of WT and iNOS−/− mice. Fiber-type distribution [%] in the soleus muscle. Percentage of type 1, type 2a, and type 2x fibers of WT mice after 6 weeks of CIH compared with NOX controls (G) and of iNOS−/− mice (H). Values are given as mean + SEM; n = 8 to 14 animals per group. *p < 0.05, **p < 0.01, ***p < 0.001, significance between CIH and NOX

A parallel CSA-specific analysis of fiber distribution showed that CIH compared to NOX significantly increased the fraction of small fibers (CSA < 800μm2) in the soleus muscle (Fig. 2).

Fig. 2figure 2

Histogram of myofiber CSA in the soleus muscle in WT mice. Values are given as mean + SEM; n = 12 to 14 animals per group. *p < 0.05, significance between WT-CIH and WT-NOX

Moreover, there was a significantly higher fraction of centronucleated fibers with CIH compared to NOX (1.5% vs. 0.2%; p = 0.01) in WT soleus muscle, as counted in HE-stained sections (Fig. 3, left). A significant inverse correlation was found between the fraction of centronucleated fibers and the CSA of the fiber population in WT undergoing CIH or NOX (r = − 0.417; p = 0.043).

Fig. 3figure 3

The percentage of centronucleated fibers in the soleus muscle is shown. Values are given as mean+SEM; n = 8 to 14 animals per group. *p < 0.05, significance between CIH and NOX

Effect of iNOS w/o CIH

In both conditions, NOX and CIH, iNOS−/− compared to WT was without any significant effect on fiber CSA of the total fiber population or that of fiber type 1, 2a, or 2x. In iNOS−/− mice, CIH as compared to NOX caused no significant changes in CSA of fibers in total or of type (Fig. 1F), while CIH-related changes in fiber composition partly resembled those in WT (Fig. 1A, C, H): CIH compared to NOX in iNOS−/− mice led to a decrease in type 2a fiber fraction (45% vs. 51%; p < 0.05) and an associated non-significant increase in type 1 fiber fraction (39% vs. 35%) at unchanged type 2x fiber fraction (Fig. 1C, D, H). Concerning the centronucleation, no significant difference was detected between the two genotypes or between CIH and NOX in iNOS−/− mice (Fig. 3).

Fiber morphology and fraction in the gastrocnemius muscle

In contrast to the soleus muscle, the gastrocnemius muscle mostly consisting of fiber type 2x revealed no significant effect of CIH vs. NOX and iNOS−/− vs. WT or their combination with regard to CSA (Additional file 2) or percentage of central nuclei (Additional file 3).

NMJ morphology/integrity in soleus muscleEffect of CIH in WT mice

In the soleus muscle of WT mice, BTX staining of NMJ showed that CIH compared to NOX led to a significant 37% smaller postsynaptic NMJ area (Fig. 4A–C, left). When calculating NMJ area relative to fiber CSA, (Fig. 4D, left), the CIH compared to NOX significantly reduced postsynaptic NMJ size. An alternative normalizing of NMJ length to fiber perimeter similarly decreased NMJ size with CIH compared to NOX (Fig. 4E, left). No significant CIH-related change occurred regarding the fraction of fragmented NMJ in WT mice (Fig. 4F, left; G).

Fig. 4figure 4

Representative images (200fold magnification, scale bar = 50 μm) of soleus muscle cross sections stained for BTX of WT-NOX (A) and WT-CIH (B) are shown. Postsynaptic NMJ area in soleus muscle (C). Postsynaptic NMJ area normalized to myofiber CSA (D) and NMJ length relative to myofiber perimeter (E) as well as percentage of fragmented NMJ in soleus muscle (F). Representative images (400-fold magnification) of BTX-stained AChR distribution at muscular NMJ in soleus muscle (G, H). G Postsynaptic BTX-stained NMJ from WT-NOX mouse (white arrow). H Fragmented postsynaptic BTX-stained NMJ of WT-CIH mouse is (black arrow). Values are given as mean + SEM; n = 8 animals per group. *p < 0.05, **p < 0.01, ***p < 0.001, significance between CIH and NOX; #p < 0.05, ##p < 0.01, ###p < 0.001, significance between WT and iNOS−/−

Effect of iNOS−/− w/o CIH

In comparison to WT, iNOS−/− mice demonstrated a significant 55% diminished NMJ area under the condition of NOX (Fig. 4C), with this effect being significant also upon normalization of postsynaptic NMJ area for fiber CSA (Fig. 4D) or, alternatively, of postsynaptic NMJ length for fiber perimeter (Fig. 4E). CIH compared to NOX intervention in iNOS−/− mice led to an additional decrease in NMJ area (in absolute terms) by trend (Fig. 4C, right); however, this effect reached significance when normalizing postsynaptic NMJ area for fiber CSA (Fig. 4D, right) or, alternatively, postsynaptic NMJ length for fiber perimeter (Fig. 4E, right). Notably, a strikingly higher percentage of NMJ fragmentation was observed selectively with iNOS−/− as compared to WT under both conditions (Fig. 4F, right; H).

NMJ morphology/integrity in gastrocnemius and vastus musclesGastrocnemius muscle

In contrast to the soleus muscle, the gastrocnemius muscle in WT mice showed a significantly increased NMJ size with CIH as compared to NOX (72%, p < 0.05, Additional file 4). However, this difference was abolished when normalizing postsynaptic NMR area for fiber CSA. Moreover, unlike the soleus muscle, the gastrocnemius muscle in iNOS−/− mice revealed no significant alterations in postsynaptic NMJ area, and this was also true for both abovementioned normalization of NMJ area or length for fiber CSA or perimeter, respectively.

Vastus muscle

For further evaluation of functional NMJ integrity, double fluorescent staining was used in the vastus muscle to quantify the area of the NMJ presynaptic nerve terminal (anti-vACHhT-antibodies, Fig. 5A, D), of postsynaptic NMJ (BTX, Fig. 5B, E) and that of their coupling (overlay, Fig. 5C, F). In line with the findings in the soleus (but not in gastrocnemius) muscle, there was a significant diminution by 27.9% (p < 0.01) of postsynaptic NMJ in iNOS−/− compared to WT under conditions of NOX (Fig. 5G), whereas CIH effects compared to NOX were absent in WT or iNOS−/− mice. The presynaptic terminal, defined as the vAChT immunoreactive area, remained resistant against CIH- or genotype-related effects (5H). The resulting percentage overlay area, a measure of NMJ integrity, was not significantly affected by CIH intervention or iNOS−/− (Fig. 5I).

Fig. 5figure 5

Confocal projection images of representative NMJ from the vastus muscle stained for postsynaptic AChR (BTX, A and D, red) and for nerve terminal vAChT (B and E, green). The merge is showing the colocalization of BTX and vAChT (C and F). Example is given of an innervated muscular NMJ where the BTX-stained area is largely covered by the vAChT staining (AC). In contrast, DF are showing a NMJ where the AChR-stained areas are only partially covered by vAChT staining, representing denervation. Scale bar in panel F is 10 μm. Mean postsynaptic (G) and presynaptic (H) NMJ areas and overlay areas of BTX and vAChT (I) in the vastus muscle are shown. Values are given as mean + SEM; n = 4 to 6 animals per group. ##p < 0.01, significance between WT and iNOS−/−

Inflammatory markers (IL1β, CD68) in soleus and gastrocnemius muscles

In WT soleus muscle, a significantly almost 9% reduced density in IL 1ß-positive cells was found under CIH compared to NOX (p < 0.05) (Fig. 6E) and iNOS−/− compared to WT, as well as iNOS−/− CIH vs. NOX was without any significant effect (Fig. 6E). In the gastrocnemius, neither CIH- nor iNOS−/−-related changes were detected.

Fig. 6figure 6

Representative images (200fold magnification, scale bar = 50 μm) of soleus muscle cross sections stained for CD68 of WT-NOX (A), WT-CIH (B), iNOS−/−-NOX (C), and iNOS−/−-CIH (D) are shown. CD68-positive cells are marked with arrows. Density of IL 1β- (E) and CD68- (F) positive cells in the soleus muscle is shown. Mean number of macrophages per NMJ (G) and mean number of macrophages per denervated NMJ (H). Merge of confocal projection images of representative NMJ from the soleus muscle stained for nerve terminal vAChT (green) and CD68 (red) showing the colocalization of NMJ and macrophages (H). Values are given as mean + SEM, n = 8 animals per group. * p < 0.05, significance between CIH and NOX, #p < 0.05, significance between WT and iNOS−/−

Analyzing the status of interstitial macrophages (CD68-positive cells), sections of WT mice soleus displayed a significant 36% reduction of macrophage density in CIH vs. NOX (Fig. 6A, B, F) while no such differences were found in the gastrocnemius muscle. In comparison to WT, iNOS−/− mice demonstrated a significant 78% increased CD68 density under the condition of CIH compared to NOX (Fig. 6C, D, F). Furthermore, a significantly diminished density was observed with iNOS−/− as compared to WT under NOX (p < 0.05) (Fig. 6A, C, F).

Given that macrophages are reported to be present at the NMJ after nerve injury, we studied the amount of macrophages next to the intact as well as damaged NMJ in the soleus muscle using double fluorescent staining (Fig. 6I). NMJ were considered denervated when the overlay of BTX and vAChT staining was less than 10%.

The mean number of macrophages covering all NMJ in soleus muscle did not significantly differ between CIH intervention and NOX in WT or iNOS−/− (Fig. 6G). A specific analysis of denervated NMJ showed, that CIH compared to NOX led to a significant decrease of macrophages next to the endplate (0.30 vs. 0.81; p < 0.02) (Fig. 6H).

Furthermore the macrophage density in the soleus muscle correlated significantly and positively with the overlay area of BTX and vAChT (r = 0.55; p = 0.004) as well as negatively with the amount of denervated NMJ (r = − 0.59; p = 0.001).

Mitochondrial ultrastructure in soleus muscle

The following ultrastructural mitochondrial abnormalities were quantified in soleus muscle of WT exposed to NOX (Fig. 7A) or CIH (Fig. 7B) as well as in iNOS−/− mice in NOX (Fig. 7C) or CIH (Fig. 7D): Swollen matrix (Fig. 7H), disruption of the outer mitochondrial membrane (Fig. 7I), a complete loss of internal architecture (Fig. 7J) and mitochondria with multi-lamellar bodies (Fig. 7K).

Fig. 7figure 7

Representative TEM images of WT (A, B) and iNOS−/− (C, D) soleus muscle under CIH (B, D) vs. NOX (A, C) (scale bars = 100 nm). Percentage of damaged (E) and swollen mitochondria (F) are shown. Mitochondria presented various morphological abnormalities such as outer membrane rupture (I; black arrow), loss of cristae (J), or multi-lammellar bodies (K; black arrowhead). G and H correspond to intact mitochondria completely (G) or partially (H) filled with cristae. Values are given as mean + SEM, n = 7 to 9 animals per group. ***p < 0.001, significance between CIH and NOX, ###p < 0.001, significance between WT and iNOS−/−

Effect of CIH in WT

The percentage of damaged mitochondria, classified as <50% filled with cristae, was significantly 1.8-fold higher in CIH vs. NOX (Fig. 7E, left), while an increase in the percentage of swollen mitochondria with CIH did not reach significance (Fig. 7F, left). The difference between CIH vs. NOX regarding the percentage of mitochondria containing multi-lamellar bodies did not reach significance (12.8% vs. 6.6%; p > 0.05).

Effect of iNOS−/− w/o CIH

Somewhat reminiscent of CIH vs. NOX effects in WT, iNOS−/− compared to WT mice at NOX conditions revealed a significant 2.1-fold increase of damaged mitochondria (Fig. 7E). Moreover, the percentage of swollen mitochondria in iNOS−/− vs. WT mice was significantly increased under both conditions (NOX > 6-fold; CIH > 5-fold) (Fig. 7F).

Mitochondrial ultrastructure in the gastrocnemius muscle

In the gastrocnemius muscle, the abovementioned mitochondrial alterations were observed neither with CIH vs. NOX in WT nor with iNOS−/− of either condition (Additional files 5 and 6).

Correlations between NMJ fiber and mitochondrial morphology

In the soleus muscle, the NMJ area correlated significantly with the percentage of damaged mitochondria (r = 0.584, p = 0.002) as well as with the ratio between type 1 and type 2a fibers (r = 0.397, p = 0.05) (Table 2).

Table 2 Correlations of NMJ area and fragmented NMJ, percentage of damaged mitochondria, myofiber CSA, and fiber type ratio (1/2a) in soleus muscles, n = 5 to 8 animals per groupTranscripts of iNOS, SOCS3, IL6, SOD2, and pro-/anti-apoptotic markers

Notably, mRNA expression of iNOS was undetectable in soleus (as well as in gastrocnemius) muscle of WT mice undergoing NOX or CIH intervention (Table 3). However, importantly, an iNOS expression was well detectable in the liver of WT mice, where it decreased significantly and massively by factor 0.12 (Table 3) after CIH-treatment. The absence of iNOS expression was proven in the liver and soleus (and gastrocnemius) muscle of iNOS−/− mice.

Table 3 Relative gene expression in soleus muscle and liver of WT and iNOS−/− mice

Importantly, in the soleus muscle of WT mice, SOCS3 expression was found to be >10-fold increased with CIH compared to NOX (Table 3). Similarly, iNOS−/− prompted a SOCS3 upregulation that was >10-fold in NOX and >4-fold in CIH as compared to WT at NOX (Table 3). Notably, mRNA expression of SOCS3 was undetectable in the gastrocnemius of WT mice. Comparing iNOS−/− CIH and NOX, SOCS3 increased by factor 2.8.

IL6 as an upstream factor of SOCS3 showed neither CIH- nor iNOS−/−-related changes in soleus muscle (Table 3). In WT mice gastrocnemius muscle, CIH compared to NOX led to a 50% decrease of IL6 mRNA expression while in iNOS−/− no relevant difference between CIH and NOX was found.

Moreover, in WT mice, CIH compared to NOX led to an almost 50% decrease in mtSOD mRNA expression in soleus muscle. Similarly, iNOS−/− resulted in a 33% and a 45% decrease in SOD2 expression in soleus with NOX and CIH, respectively (Table 3).

Furthermore, the screening for apoptotic markers (BAX, BCL2, caspase 3) showed neither CIH- nor iNOS−/−-related changes in soleus muscle (Table 3).

留言 (0)

沒有登入
gif