Colonocyte keratins stabilize mitochondria and contribute to mitochondrial energy metabolism

Abstract

Keratin intermediate filaments form dynamic filamentous networks, which provide mechanical stability, scaffolding, and protection against stress to epithelial cells. Keratins and other intermediate filaments have been increasingly linked to the regulation of mitochondrial function and homeostasis in different tissues and cell types. While deletion of keratin 8 (K8^-/-) in mouse colon elicits a colitis-like phenotype, epithelial hyperproliferation, and blunted mitochondrial ketogenesis, the role of K8 in colonocyte mitochondrial function and energy metabolism is unknown. We used two K8 knockout mouse models and CRISPR/Cas9 K8^-/- colorectal adenocarcinoma Caco-2 cells to answer this question. The results show that K8^-/- colonocyte mitochondria in vivo are smaller and rounder and that mitochondrial motility is increased in K8^-/- Caco-2 cells. Furthermore, K8^-/- Caco-2 cells displayed diminished mitochondrial respiration and decreased mitochondrial membrane potential compared with controls, whereas glycolysis was not affected. The levels of mitochondrial respiratory chain complex proteins and mitochondrial regulatory proteins mitofusin-2 and prohibitin were decreased both in vitro in K8^-/- Caco-2 cells and in vivo in K8^-/- mouse colonocytes, and reexpression of K8 into K8^-/- Caco-2 cells normalizes the mitofusin-2 levels. Mitochondrial Ca²⁺ is an important regulator of mitochondrial energy metabolism and homeostasis, and Caco-2 cells lacking K8 displayed decreased levels and altered dynamics of mitochondrial matrix and cytoplasmic Ca²⁺. In summary, these novel findings attribute an important role for colonocyte K8 in stabilizing mitochondrial shape and movement and maintaining mitochondrial respiration and Ca²⁺ signaling. Further, how these metabolically compromised colonocytes are capable of hyperproliferating presents an intriguing question for future studies.NEW & NOTEWORTHY In this study, we show that colonocyte intermediate filament protein keratin 8 is important for stabilizing mitochondria and maintaining mitochondrial energy metabolism, as keratin 8-deficient colonocytes display smaller, rounder, and more motile mitochondria, diminished mitochondrial respiration, and altered Ca²⁺ dynamics. Changes in fusion-regulating proteins are rescued with reexpression of keratin 8. These alterations in colonocyte mitochondrial homeostasis contribute to keratin 8-associated colitis pathophysiology.

Keywords: colon; energy metabolism; inflammatory bowel diseases; keratins; mitochondria.

Graphical abstract

Figure 1.

K8^−/− mouse colonocytes display smaller and more round mitochondria. Mitochondrial morphology was assessed by transmission electron microscopy (TEM) imaging of colonocytes in the surface epithelium and the upper regions of crypts in K8^+/+ and K8^−/− mouse proximal colon tissue samples. A: representative TEM images of K8^+/+ and K8^−/− mouse colonocytes. Scale bar = 1 μm. B: the shape of K8^+/+ and K8^−/− colonocyte mitochondria was determined and is displayed as the percentage of long, oval, or round mitochondria of total mitochondria. C and D: analysis of mitochondria area (C) and the number of mitochondria (D) in K8^+/+ and K8^−/− colon tissue TEM images. The results represent the average (n = 6 images/mouse, 2 mice/genotype; 387 K8^+/+ and 330 K8^−/− mitochondria analyzed) ± SD with significant differences shown: **P < 0.01 and ***P < 0.001.

Figure 2.

Mitochondrial motility is increased in Caco-2 cells lacking K8. K8^+/+ and K8^−/− Caco-2 cells were stained with mitochondrial dyes tetramethylrhodamine ethyl ester (TMRE) or MitoTracker Deep Red and tracked by confocal microscopy. A: representative images of total mitochondrial tracks (left: red color; scale bars = 25 μm) and individual mitochondrial tracks (right: various colors; scale bars = 12.5 μm) in TMRE-stained K8^+/+ and K8^−/− Caco-2 cells. The tracks represent the movement of mitochondria over a 2-minute time lapse. B and C: velocities of mitochondria stained with TMRE (B) and MitoTracker Deep Red (C) were quantified. The results represent the average (n = 149–165 mitochondria/genotype for TMRE; n = 145 mitochondria/genotype for MitoTracker Deep Red; small dots represent individual values) ± SD with significant differences shown: ***P < 0.001.

Figure 3.

Mitochondrial respiration and respiratory capacity, but not glycolysis, are diminished in K8^−/− Caco-2 cells. A: Seahorse XF Cell Mito Stress Test Kit was used to measure the oxygen consumption rates of K8^+/+ and K8^−/− Caco-2 cells, and the results were normalized to total protein quantity. B–G: Seahorse XF Mito Stress Test Report Generator was used to calculate basal respiration (B), ATP production (C), maximal respiration (D), spare respiratory capacity (E), proton leak (F), and nonmitochondrial oxygen consumption (G). H: Seahorse XF Glycolytic Rate Assay Kit was used to measure the glycolytic proton efflux rate of K8^+/+ and K8^−/− Caco-2 cells, and the results were normalized to total protein quantity. Agilent Seahorse Analytics was used to calculate basal glycolysis. The results represent the average oxygen consumption rate/glycolytic proton efflux rate (n = 4/genotype, with 5–6 technical replicates each) ± SD with significant differences shown: *P < 0.05 and **P < 0.01.

Figure 4.

Loss of K8 leads to decreased mitochondrial membrane potential (ΔΨm) in Caco-2 cells. A and B: ΔΨm was assessed in K8^+/+ and K8^−/− Caco-2 cells using the ΔΨm-sensitive probe tetramethylrhodamine ethyl ester (TMRE). FCCP, which causes mitochondrial depolarization and decreases ΔΨm, was used as a positive control. A: TMRE fluorescence intensity levels analyzed by flow cytometry. The measured median fluorescence intensity (MFI) values were normalized to the MFI of unstained cells. The results represent the average relative MFI (n = 5 replicate assays/genotype, with 50,000–100,000 cells/assay) ± SD with significant differences shown: *P < 0.05, **P < 0.01, and ***P < 0.001. B: TMRE fluorescence intensity levels analyzed by plate reader assay. The measured fluorescence intensity values were corrected for the background signal in unstained controls. The results represent the average (n = 4, with 5–29 technical replicates each per genotype and condition/assay; small dots represent individual values) fold change in TMRE fluorescence intensity relative to K8^+/+ ± SD with significant differences shown: ***P < 0.001.

Figure 5.

K8^−/− Caco-2 cells exhibit decreased protein levels of oxidative phosphorylation complexes and mitochondrial regulatory proteins, and reintroduction of K8 rescues the downregulation of mitofusin-2. A–H: K8^+/+ and K8^−/− Caco-2 cells were analyzed by immunoblotting (A) for oxidative phosphorylation protein complex V (CV) subunit ATP5A (B); complex III (CIII) subunit UQCRC2 (C); complex II (CII) subunit SDHB, complex IV (CIV) subunit COX II (D); and complex I (CI) subunit NDUFB8 (E) and mitochondrial regulatory proteins prohibitin (F), mitofusin-2 (G), and trichoplein (H). The K8 genotype is shown by the K8 blot (notably, the same K8 blot is presented in Fig. 7G, as the same sample set was used here). The immunoblots were quantified and normalized against β-actin. The results represent the average (n = 3/genotype) fold change relative to K8^+/+ ± SD, with significant differences shown: *P < 0.05. I–K: K8^−/− Caco-2 cells transfected with pcDNA or K8 and K18 DNA plasmids and K8^+/+ cells were analyzed by immunoblotting (I) for mitofusin-2 (J) and K8 (K). The immunoblots were quantified and normalized against β-actin. The results represent the average (n = 3/genotype) fold change relative to K8^+/+ ± SD, with significant differences shown for mitofusin-2: *P < 0.05 and **P < 0.01. MW (kDa) indicates the closest molecular mass marker on the Western blots.

Figure 6.

K8^−/− mouse colonocytes display decreased protein levels of oxidative phosphorylation complexes and mitochondrial regulatory proteins. A–H: isolated colonocytes from K8^+/+, K8^−/− and K8^+/− mice were analyzed by immunoblotting (A) for oxidative phosphorylation protein complex V (CV) subunit ATP5A (B); complex III (CIII) subunit UQCRC2 (C); complex II (CII) subunit SDHB, complex IV (CIV) subunit COX II (D); and complex I (CI) subunit NDUFB8 (E) and mitochondrial regulatory proteins prohibitin (F), mitofusin-2 (G), and trichoplein (H). The K8 genotype is shown by the K8 blot (notably, the same K8 blot is presented in Fig. 7J, as the same sample set was used here). The immunoblots were quantified and normalized against Hsc70. The results represent the average (n = 3 male mice/genotype) fold change relative to K8^+/+ ± SD, with significant differences shown: *P < 0.05 and **P < 0.01. MW (kDa) indicates the closest molecular mass marker on the Western blots.

Figure 7.

The mitochondrial Ca²⁺ response is attenuated in Caco-2 K8^−/− cells and colonic epithelium of mice lacking K8. A and D: K8^+/+ and K8^−/− Caco-2 cells were transfected with aequorin targeted to the mitochondrial matrix (mito-Aeq; A) or untagged aequorin (D; no target, cytoplasmic localization), and Ca²⁺ responses were measured following stimulation with ATP. Representative mitochondrial (A) and cytoplasmic (D) Ca²⁺ responses are shown. B, C, E, and F: changes in mitochondrial and cytoplasmic Ca²⁺ levels were quantified and displayed as Δ (maximum – minimum) Ca²⁺ concentration ([Ca²⁺]) (B and E) and area under the curve (AUC) [Ca²⁺] (C and F). The results represent the average Ca²⁺ responses (n = 24/genotype for mitochondrial [Ca²⁺] and n = 35–36/genotype for cytoplasmic [Ca²⁺] from 3 experiments; small dots represent individual values) ± SD with significant differences shown: ***P < 0.001. G–L: K8^+/+ and K8^−/− Caco-2 cells (G–I) and isolated colonocytes from male K8^+/+, K8^−/−, and K8^+/− mice (J–L) were analyzed by immunoblotting for the mitochondrial Ca²⁺ transporters voltage-dependent anion channel (VDAC) and mitochondrial calcium uniporter (MCU). The K8 genotype is shown by the K8 blots (notably, the same K8 blot is shown in Figs. 5A and 7G, as both used the same Caco-2 sample set, and, similarly, the same K8 blot is shown in Figs. 6A and 7J, as both used the same mouse colonocyte sample set). The immunoblots were quantified and normalized against β-actin or Hsc70. The results represent the average (n = 3/genotype) fold change relative to K8^+/+ ± SD, with significant differences shown: *P < 0.05 and **P < 0.01. M–Q: VDAC, MCU, prohibitin, and mitofusin-2 protein levels were determined in male and female K8^flox/flox and K8^flox/flox;Villin-Cre (endogenous K8 knockdown) mouse colonocytes by immunoblotting (a representative Western blot is shown). The immunoblots were quantified and normalized against Hsc70. The results represent the average fold change relative to K8^+/+ ± SD (n = 6 mice/genotype for VDAC and MCU; n = 3 mice/genotype for prohibitin and mitofusin-2) with significant differences shown: *P < 0.05, **P < 0.01, and ***P < 0.001. MW (kDa) indicates the closest molecular mass marker on the Western blots.

Figure 8.

Graphical summary. K8 stabilizes mitochondria by restricting mitochondrial movement and contributes to maintaining mitochondrial size, shape, energy metabolism, and Ca²⁺ signaling in colonocytes. In the absence of K8, colonocyte mitochondria are smaller, rounder, and more motile. K8^−/− mitochondria also display diminished respiration and mitochondrial membrane potential (ΔΨm) and decreased oxidative phosphorylation complex, prohibitin, and mitofusin-2 protein levels. Furthermore, Ca²⁺ levels are decreased and Ca²⁺ dynamics are altered in the mitochondrial matrix and cytoplasm of K8^−/− Caco-2 cells, while the mitochondrial Ca²⁺ transporters voltage-dependent anion channel (VDAC) and mitochondrial calcium uniporter (MCU) are downregulated in K8^−/− mouse colonocytes. Taken together, the above-mentioned alterations in mitochondrial dynamics and function are likely contributing factors to the diminished mitochondrial energy metabolism in K8^−/− colonocytes and may contribute to K8-associated colitis pathophysiology. [Ca²⁺], Ca²⁺ concentration; IFs, intermediate filaments.