Preservation of circadian rhythms by the protein folding chaperone, BiP

Abstract

Dysregulation of collagen synthesis is associated with disease progression in cancer and fibrosis. Collagen synthesis is coordinated with the circadian clock, which in cancer cells is, curiously, deregulated by endoplasmic reticulum (ER) stress. We hypothesized interplay between circadian rhythm, collagen synthesis, and ER stress in normal cells. Here we show that fibroblasts with ER stress lack circadian rhythms in gene expression upon clock-synchronizing time cues. Overexpression of binding immunoglobulin protein (BiP) or treatment with chemical chaperones strengthens the oscillation amplitude of circadian rhythms. The significance of these findings was explored in tendon, where we showed that BiP expression is ramped preemptively prior to a surge in collagen synthesis at night, thereby preventing protein misfolding and ER stress. In turn, this forestalls activation of the unfolded protein response in order for circadian rhythms to be maintained. Thus, targeting ER stress could be used to modulate circadian rhythm and restore collagen homeostasis in disease.-Pickard, A., Chang, J., Alachkar, N., Calverley, B., Garva, R., Arvan, P., Meng, Q.-J., Kadler, K. E. Preservation of circadian rhythms by the protein folding chaperone, BiP.

Keywords: 4PBA; ER stress; Per2::luc; UDCA; collagen.

Figure 1

ER stress dampens circadian rhythm. A) Hspa5 (BiP or GRP78) transcript levels fluctuate over the course of 2 d in mouse tail tendons as assessed by microarray; n = 2 animals/time point. BiP levels rise prior to the increased synthesis of collagen, indicated by procollagen-specific peptides; n = 4 animals/time point. X axis represents hours in dark/dark cycle, as previously described in Yeung et al. (34). B) BiP levels as assessed by Western blot in MEFs over the course of 24 h postsynchronization (PS) with 50% horse serum. Lower traces show the fluctuations in BiP protein levels, assessed by densitometry and normalized to GAPDH, in tail tendons, iTTFs, and MEFs as assessed by CircWave (see Supplemental Fig. S1). C) The effects of ER stress induction on circadian rhythm in ex vivo tail tendons from Per2::luc mice. Tendons were treated with 10 nM thapsigargin in recording medium and then treated with dexamethasone to entrain the circadian rhythm. D) Cell viability after 72 h in response to the indicated doses of thapsigargin (upper) and tunicamycin (lower) as assessed by AlamarBlue (n = 3); error bars represent sd. E) Representative luminescence traces (upper) following 1 h treatment of 10 nM thapsigargin or 100 ng/ml tunicamycin, which dampens the induction of Per2::luc in response to dexamethasone; relative amplitude over 3 d is shown (lower); n = 3 biologic replicates. F) Thapsigargin and tunicamycin are removed from Per2::luc cells after 5 h treatment, then synchronized with dexamethasone and then compared to untreated cells. Relative amplitudes in the first 24 h are shown (lower chart, n = 4 biologic replicates). G) iTTFs expressing congenital goiter–mutated thyroglobulin (ColII-Tg-COG mutant) have reduced inherent and dexamethasone-induced rhythms; n = 2 independently generated cell lines; n = 2 biologic replicates. H) Quantification of amplitudes of Per2::luc traces. I) Levels of spliced XBP1 (XBP1s) and total Xbp1 (XBP1t) indicate that the IRE1-XBP1 arm of the UPR has been activated in ColII-Tg-COG cells; n = 3; error bars represent sd. Dex, dexamethasone; Thaps, thapsigargin; Tunic, tunicamycin; unt, untreated. *P < 0.05, **P < 0.01 (paired Student’s t test).

Figure 2

Impact of ER stress on collagen production. A) Western blot analysis on conditioned medium (CM) collected from iTTFs after 48 h of treatment with thapsigargin, tunicamycin, brefeldin, and monensin suggests that secreted collagen-I is suppressed in treated cells (n = 2). B) Diagram of the secretory pathway marking the position of action for the inhibitors brefeldin A and monensin. C) The effects of brefeldin A and monensin, at various concentrations, on the survival of fibroblasts following 72 h treatment. n = 3; error bars represent sd. D) The effects of brefeldin A and monensin, at various concentrations, on the survival of fibroblasts following 72-h treatment. E) Levels of CHOP/DDIT3 indicate that the unfolded protein response has been activated following 5-h treatment with brefeldin A and monensin, similar to the effects of thapsigargin and tunicamycin. F, G) Using in-cell Western blot, the effects of thapsigargin and tunicamycin on the production of extracellular collagen (F) and intracellular collagen (G) are shown, with the effects of pretreatment with thapsigargin or tunicamycin for different times; error bars represent sd of triplicate measurements. H) Per2::luc signals in lung fibroblasts isolated from wild-type control and ClockΔ19 mice. Cells were recorded for 24 h before treatment with 100 nM dexamethasone at the indicated time. Representative traces are shown; n = 3. I) Representative images of collagen fibers formed by WT and ClockΔ19 fibroblasts 12 h apart (am and pm, d 2) after synchronization with dexamethasone. Scale bar, 10 μm. J) Scores of the number of collagen fibers per cell were assessed over 2 d; peak numbers of collagen fibers are observed 11 h after the synchronization event; data on each day are normalized to this time point: 11 h postsynchronization or 35 h postsynchronization (11 h postsynchronization + 24 h). On average, 2000 cells per time point were scored. Δ19, ClockΔ19; Dex, dexamethasone; Mon, monensin; n.s., not significant; Thaps, thapsigargin; Tunic, tunicamycin; unt, untreated; WT, wild type. *P < 0.05; **P < 0.01 (paired Student’s t test).

Figure 3

BiP retains collagen in the ER. A) iTTFs were transduced with pCMMP-BiP-IRES-mRFP; red fluorescent protein (RFP)–positive cells were sorted to form a BiP-overexpressing cell line (iTTF + BiP). B, C) Levels of BiP mRNA (B) and protein (C) were assessed in sorted populations. D) Immunofluorescence detection of collagen fibers in control and BiP-overexpressing iTTFs; cells were synchronized with dexamethasone and collected at different time points. Scale bar, 10 μm. E) Scores of the numbers of collagen fibers counted per cell at the indicated times; n = 2 biologic repeats, with 3–4 regions scored per sample; ∼200 cells per region were scored. *P < 0.05, **P < 0.01 (paired Student’s t test). F) The effects of BiP overexpression on the transcript levels of collagen-I (Col1a1 and Col1a2); n = 2 biologic repeats. G) Colocalization of collagen-I and the ER marker PDI in control and BiP-overexpressing iTTFs. OE, overexpressing; PS24 and PS30, 24 and 30 h postsynchronization of the circadian clock by dexamethasone. Scale bar, 10 μm.

Figure 4

BiP overexpression strengthens circadian rhythm. A, B) Representative traces of control and BiP-overexpressing iTTFs; BiP-expressing cells have a more robust circadian rhythm in unsynchronized populations (A) and following dexamethasone-induced synchronization (B). C) Amplitude of Per2::luc signals from control and BiP-overexpressing iTTFs, unsynchronized and synchronized by dexamethasone, in the first 48 h, quantified from n = 3 biologic replicates; error bars represent sd. D) The activation of the UPR in BiP-overexpressing cells was assessed by monitoring the expression of ATF4/CHOP, Xbp1 splicing (XBP1s), and ATF6 transcripts; n = 2, each analyzed in triplicate, error bars represent sd. Dex, dexamethasone; OE, overexpressing. *P < 0.05 (paired Student’s t test).

Figure 5

Chemical chaperones strengthen circadian rhythm and enhance collagen secretion. A) Effects of 4PBA treatment on the survival of iTTFs as assessed by AlamarBlue; triplicate analysis, error bars represent sd. B) The effects of 100 µM 4PBA on circadian rhythm in unsynchronized populations (upper panel) and following dexamethasone treatment (lower panel); representative traces are shown. C) Amplitudes of Per2::luc signals following 4PBA treatment; n = 3–4 biologic replicates. D) Effects of UDCA treatment on the survival of iTTFs as assessed by AlamarBlue; triplicate analysis, error bars represent sd. E) The effects of 100 µM UDCA on circadian rhythm in unsynchronized populations (upper panel) and following dexamethasone (lower panel). F) Amplitudes of Per2::luc signals following UDCA treatment. n = 3–4 biologic replicates. G) Effects of 4PBA on the expression of components of the UPR [ATF4, CHOP, spliced XBP1 (XBP1s), total XBP1 (tXBP1), ATF6] after 72 h treatment; n = 2 each analyzed in triplicate; error bars represent sem. H, I) The effects of 4PBA treatment on the production of collagen fibers following 72-h treatment (H) and quantified (I); n = 2 with 5 regions scored in each sample; ∼200 cells/region were scored; error bars represent sd. Scale bar, 10 μm. J) The impact of 4PBA treatment on collagen-I transcripts (Col1a1 and Col1a2); n = 2 each analyzed in triplicate; error bars represent sd. Dex, dexamethasone. *P < 0.05, **P < 0.01 (paired Student’s t test).