Paracrine cellular senescence exacerbates biliary injury and impairs regeneration

Abstract

Cellular senescence is a mechanism that provides an irreversible barrier to cell cycle progression to prevent undesired proliferation. However, under pathological circumstances, senescence can adversely affect organ function, viability and regeneration. We have developed a mouse model of biliary senescence, based on the conditional deletion of Mdm2 in bile ducts under the control of the Krt19 promoter, that exhibits features of biliary disease. Here we report that senescent cholangiocytes induce profound alterations in the cellular and signalling microenvironment, with recruitment of myofibroblasts and macrophages causing collagen deposition, TGFβ production and induction of senescence in surrounding cholangiocytes and hepatocytes. Finally, we study how inhibition of TGFβ-signalling disrupts the transmission of senescence and restores liver function. We identify cellular senescence as a detrimental mechanism in the development of biliary injury. Our results identify TGFβ as a potential therapeutic target to limit senescence-dependent aggravation in human cholangiopathies.

Fig. 1

Cellular senescence is an intrinsic characteristic of PBC/PSC. Explanted human livers diagnosed as Normal, PSC and PBC respectively (series of N = 7 per group). a Immunohistochemical detection of p21 in bile ducts. Far right, digital magnification of p21-positive hepatocytes (top) and cholangiocytes (bottom). b PSC and PBC show p16 expression (red) in cholangiocytes (green). Far right, quantification of p16 in K19-positive cholangiocytes. Median age of each of the groups is also included in this figure. c PSC and PBC show DCR2 expression (green) in cholangiocytes (red). Far right, quantification. d PSC and PBC show γH2A.X expression (green) in cholangiocytes (red). Far right, quantification. e Immunohistochemical detection of p27 increases in PSC and PBC compared to Normal Liver. ** denotes p < 0.01, *** denotes p < 0.001, **** denotes p < 0.0001 (Mean ± SEM). ANOVA, Sidak’s multiple comparisons test. Scale bars = 50 µm

Fig. 2

Cholangiocyte-specific deletion of Mdm2 results in p21-driven cellular senescence. a Schematic explanation of the model: K19CreER^T promoter is used to flox out Mdm2 in cholangiocytes. Upon tamoxifen administration, Cre recombinase is activated and Mdm2 is cleaved. The STOP sequence upstream tdTom reporter sequence is also floxed out making the identification of primary senescent events traceable through tdTom expression. b Increased levels of p21 in bile ducts 2 days after tamoxifen administration. Mouse data is presented at day 2 after final tamoxifen administration (N = 6 per group). (Left) K19Cre-positive mice without tamoxifen injection. (Center) K19Cre-negative mice after tamoxifen injection. (Right) K19Cre-positive mice after tamoxifen injection. c Cell-cycle-arrested cells do not proliferate. 0.1% (Mean ± SEM) of cholangiocytes co-express p53 (red) and Ki67 (green) 2 days after tamoxifen administration. d From top to bottom: 53BP1, γH2A.X and DCR2 (red) in K19-positive cholangiocytes (green). Notice the presence of 53BP1-positive, γH2A.X-positive and DCR2-positive hepatocytes in Cre+ +TM group. Bottom panel: Cre− +TM shows some DCR-positive hepatocytes (white arrowheads). e Recombined K19-positive cholangiocytes (green) can be traced by means of tdTom expression (red). f tdTom (red) co-localizes with p53 (green) in bile ducts. g Increased intensity of αSMA-positive cells (green) contiguous to tdTom-positive senescent-cholangiocytes (red) compared with tdTom-negative bile ducts at day 21 after induction N = 6. For an explanation about how this result was acquired, refer to Supplementary Fig. 5. h Increased total percentage of αSMA-positive cells in the Cre+ +TM group at day 21. i Increased deposition of collagen measured by PicroSirius Red (PiSR) at 21 days after tamoxifen induction. Right, quantification. ** denotes p < 0.01 (Mean ± SEM), Student’s ttest. Scale bars = 50 µm

Fig. 3

Paracrine senescence in the model is TGFβ-dependent. Mouse data is presented at day 2 after final tamoxifen administration (N = 6 per group). (Left) K19Cre-positive mice without tamoxifen injection. (Center) K19Cre-negative mice after tamoxifen injection. (Right) K19Cre-positive mice after tamoxifen injection. a p53 (red, indicative of primary senescence events in the cholangiocytes) and p21 (green) co-localize in the bile ducts. However, p21-positive p53-negative cholangiocytes were observed. b p21 (green) co-localizes with HNF4α-positive hepatocytes (red). c p27 (green) colocalizes with p53-positive cholangiocytes (red). p27 is expressed in cholangiocytes and hepatocytes in a paracrine manner. d p27 (green) co-localizes with HNF4α-positive hepatocytes (red). e Markers of senescence (such as 53BP1, H2A.X and DCR2, in red) are expressed in cholangiocytes and liver parenchyma, while expression of p53 (green) is restricted to the cholangiocytes. Single channels for this figure are provided in Supplementary Fig. 8a. Scale bars = 50 µm. f qRT-PCR analysis of Tgfb1 in the isolated bile ducts of Cre− +TM (N = 5) and induced Cre+ +TM (N = 6) K19-Mdm2^flox/floxtdTom^LSL mice day 2 after induction. *** denotes p < 0.005 (Mean ± SEM). Mann−Whitney test. g Analysis of common SASP’s factors by qRT-PCR (from left to right, Tgfbr1, Tgfbr2, Nfkb, Il1a and Il6). * denotes p < 0.05, ** p < 0.01 (Mean ± SEM). Mann−Whitney test. h Western blot of phospho-NF-κB-p65 and pSmad2/3. Each band represents the bile ducts isolated from one mouse. Blots are companied by at least one marker position (molecular weights MW, in kDa). Far right, western blot densitometry quantification (normalized to bActin control) show increased expression of NF-κBp-65 and pSmad2/3 after induction of the model. * denotes p < 0.05 (Mean ± SEM). Mann−Whitney test. i qRT-PCR of Nfkb in hepatocytes isolated from Cre− and Cre+ K19-Mdm2^flox/floxtdTom^LSL mice at day 2 after last tamoxifen administration (N = 5-6 per group). ** denotes p < 0.01 (Mean ± SEM). Mann−Whitney test

Fig. 4

Cellular senescence in cholangiocytes aggravates biliary injury. a Schematic representation of the experiment. Two days after the induction of the model the DDC diet was administered for 1 week. Experimental groups include: DDC diet control group (DDC, where mice are administered with oil, N = 6) and Senescence + DDC diet (S + DDC, where animals receive tamoxifen, N = 8). b Serum analysis shows increased liver damage in the presence of senescent cholangiocytes. c Ki67-proliferating (green) cholangiocytes (red) decrease in the S + DDC group. d Increased number of p21-positive cholangiocytes in the S + DDC group. e Total percentage of proliferating cells decrease in the S + DDC group. f Increased number of p21-positive HNF4α-positive hepatocytes in the S + DDC group. g Increase of total number of p27 cells in the S + DDC group. h Increased presence of αSMA-positive cells (green) in the proximity of tdTom-positive senescent cholangiocytes (red). i Increased collagen deposition in the S + DDC group. j Increased number of F4/80-positive macrophages (red) in the S + DDC group. * denotes p < 0.05, ** p < 0.01, *** p < 0.001. (Mean ± SEM), Student’s ttest. Scale bars = 50 µm

Fig. 5

Cellular senescence in cholangiocytes impairs the regenerative response of liver parenchyma. a Schematic representation of the experiment. After induction of the model, we waited 2 days for display of senescence and performed 70% PH. Livers were recovered at day 2 after the surgery (peak of DNA replication in hepatocytes) and day 7 (when the liver mass is restored). Two groups were included in this experiment; PH (N = 8) and Senescence + PH (S + PH, N = 7). Sham animals are included as controls of the PH technique (N = 4). b Whole livers collected at 48 h post PH. From left to right: Sham (no manipulation of the liver), PH (which shows normal regenerative process by compensatory hypertrophy) and S + PH. Notice the absence of compensatory hypertrophy in the S + PH liver. c Decrease of Ki67 (green) in the S + PH group in comparison with PH group and Sham at 48 h post PH. Far right, quantification of total proliferation in the liver. *p < 0.05, ****p < 0.0001, (Mean ± SEM). ANOVA, Sidak’s multiple comparisons test. Scale bars = 50 µm. d Increase of p21-positive hepatocytes in S + PH group at 48 h post-PH. **p < 0.01, ***p < 0.001 (Mean ± SEM). ANOVA, Sidak’s multiple comparisons test. e Increase of total p27 in S + PH group at 48 h post-PH. ****p < 0.0001 (Mean ± SEM). ANOVA, Sidak’s multiple comparisons test. f Survival curve for PH vs. S + PH at 1 week show a significant decrease of survival for S + PH mice (PH, N = 4; S + PH, N = 6). Log-Rank (Mantel−Cox test)

Fig. 6

CM-SASP induces senescence in a paracrine manner. a Schematic representation of the experiment. Biliary cells are cultured with SASP-conditioned media (CM-SASP) and Control-conditioned media (CM-Ctrol). After 1 week, different markers of senescence were evaluated. b Number of total Ctrol-treated biliary cells increase (suggesting that they continue proliferating). Total number of CM-SASP-treated biliary cells plateau, indicative of an impaired proliferative response. **** denotes p < 0.0001 (Mean ± SEM). Mann−Whitney test. N = 7–8 biological replicates. c qRT-PCR shows a significant increased expression of cdkn1a (p21), cdkn1b (p27) and cdkn2a (p16) in the CM-SASP-treated biliary cells at day 7. * denotes p < 0.05 (Mean ± SEM). Mann−Whitney test. N = 4 biological replicates. d Morphological changes observed in the cell cultures. CM-SASP-treated biliary cells become flatter, larger and their content is vacuolized. e Increased SA-βGal (red) expression in CM-SASP-treated biliary cells with decreased BrdU incorporation (green). Far right, total percentage of SA-βGal or BrdU per DAPI-positive nuclei per field. f Increased expression of different senescence markers in CM-SASP-treated biliary cells compared with the CM-Ctrol-treated biliary cells. Below, total percentage of those markers per DAPI-positive nuclei. * denotes p < 0.05 (Mean ± SEM). Mann−Whitney test. N = 4 biological replicates. Scale bars = 50 µm

Fig. 7

Cellular senescence can be transmitted in vitro. Inhibition of TGFβ decreases paracrine senescence. a Experimental scheme; biliary cells were treated with RAS-conditional media (CM-RAS) and CONTROL-conditional media (CM-Ctrol) for a week. Media is eliminated and cells washed with PBS five times. YFP-positive biliary cells (YFP-biliary cells) are then added and after 1 week different markers of senescence in the YFP-biliary cells population were assessed. b Representative images of the assessed markers of senescence and proliferation (BrdU) in the co-culture of senescent biliary cells (YFP-negative) and YFP-positive cells. Data is presented as biliary cells treated with CM-STOP and CM-RAS at final time point (14 days). Far right, percentage of each marker in the YFP-biliary cells population. * denotes p < 0.05 (Mean ± SEM). Mann−Whitney test. N = 4 biological replicates. Scale bars = 50 µm. c Experimental scheme for the use of TGFβ inhibitor (LY-2157299); YFP-biliary cells and different concentrations of LY-2157299 are added to the senescent biliary cells. After 1 week different markers of senescence in the YFP-biliary cells population were assessed. d Decrease of senescence factors and increase in proliferation in YFP-biliary cells with increasing concentrations of LY-2157299. * denotes p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 (Mean ± SEM). ANOVA, Sidak’s multiple comparisons test. N = 4 biological replicates

Fig. 8

Use of TGFβ inhibitors in vivo impairs transmission of paracrine senescence and improves liver function. a Pattern of administration of LY2157299 by oral gavage after induction of senescence in the K19-Mdm2^flox/floxtdTom^LSL model. b Representative images of p21 (red) in cholangiocytes (green) in mice treated with vehicle, 10 or 20 mg/kg of LY2157299. c Percentage of p21-positive cholangiocytes diminishes with the administration of LY2157299. d Total percentage of p21 diminishes with LY2157299 administration. e Decrease of p27 total number of cells with LY2157299 administration. f Percentage of tdTom-positive cholangiocytes is not altered. Statistic analysis for c–f * denotes p < 0.05, ** denotes p < 0.01, *** denotes p < 0.001, **** denotes p < 0.0001 (Mean ± SEM). ANOVA, Sidak’s multiple comparisons test (N = 5 per group). g After induction of senescence in cholangiocytes, LY2157299 was administered to the mice by oral gavage. Then DDC diet was administered to the mice for 1 week. Experimental groups for this experiment include: Senescence + Vehicle + DDC (Veh, N = 5) and Senescence + Inhibitor + DDC (Inh, N = 5). h Decrease in serum transaminase levels. i Increased levels of Ki67-positive cholangiocytes with the use of LY2157299. Far right, quantification. j Increased levels of total Ki67 per field with LY2157299. k Trend to decrease (p = 0.0952) of total p27 levels. l Left, decreased expression of p21-positive hepatocytes. Right, decreased expression of p16-positive hepatocytes. m Decreased collagen deposition with the use of LY2157299. n Trend to decrease (p = 0.1508) of total number of F4/80 macrophages per field. * denotes p < 0.05, ** p < 0.01 (Mean ± SEM). Student’s ttest. Scale bars = 50 µm