bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
1 Sex-dependent pathways in hepatocarcinogenesis triggered by 2 deregulated cholesterol synthesis
3 Running title: Hepatocarcinogenesis and deregulated cholesterol synthesis 4 5 Kaja Blagotinšek Cokan1, Žiga Urlep1, Gregor Lorbek1, Madlen Matz-Soja2*, Cene Skubic1, 6 Martina Perše3, Jera Jeruc4, Peter Juvan1, Tadeja Režen1, Damjana Rozman1#
7 1 Centre for Functional Genomics and Bio-Chips, Institute of Biochemistry, Faculty of Medicine, University of Ljubljana, 8 Ljubljana, Slovenia
9 2 Rudol-Schönheimer-Institute of Biochemistry, Divison of General Biochemistry, Faculty of Medicine, University of Leipzig, 10 Leipzig, Germany
11 3 Medical Experimental Centre, Institute of Pathology, Faculty of Medicine, University of Ljubljana, Ljubljana, Slovenia
12 4 Institute of Pathology, Faculty of Medicine, University of Ljubljana, Ljubljana, Slovenia
13 * Current address: Section of Hepatology, Clinic and Polyclinic for Gastroenterology, Hepatology, Infectiology, Pneumology, 14 University Clinic Leipzig, Germany
15 # Correspondence and requests for materials should be addressed to D.R. ([email protected]) 16
17 ABSTRACT
18 We uncover novel pathways of sex-dependent hepatocarcinogenesis due to chronic 19 repression of cholesterol synthesis at the lanosterol 14α-demethylase (CYP51) step. 20 The response to metabolic insult determined by global liver transcriptome, qPCR and 21 sterol metabolite analysis, together with blood parameters, revealed molecular 22 signatures that differ between females and males. The data deduced from the 23 mouse model are highly relevant for humans. The dampened hepatic metabolism 24 presents a hallmark of carcinogenesis, particularly in ageing females, with increased 25 plasma cholesterol and HDL, and a substantial negative enrichment of transcription 26 factors from lipid metabolism, such as NR1B1, LXRα, LRH1, and FXR. Importantly, 27 the carcinogenic signalling pathways (ECM-receptor interaction and PI3K/Akt) are 28 positively enriched, albeit with sex-dependent gene targets. The activated TGF-β, 29 mTOR, Wnt, and estrogen signalling worsen the phenotype, with NFATC1/2 being 30 central to the female phenotype. This collectively leads to activated cell death and 31 diminished basal metabolism. In conclusion, our data underline sex as an important 32 biological variable of hepatocarcinogenesis. We uncover novel cholesterol- 33 dependent transcription factors and signalling pathways as cancer markers in the 34 ageing females.
1 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
35
36 Significance: Chronic repression of the late part of cholesterol synthesis provokes 37 hepatocarcinogenesis with sex-dependent modulation of signalling pathways and 38 transcription factors. Aging females show specific metabolic signatures and a more 39 aggrevated phenotype of metabolism-related HCC.
40
41 Keywords. Cholesterol biosynthesis / hepatocellular carcinoma / lanosterol 14α- 42 demethylase (CYP51A1) / sex dimorphism.
43
44 INTRODUCTION
45 Metabolic abnormalities result in metabolic associated fatty liver disease (MAFLD),
46 previously termed (NAFLD) that affects over 30% of the western population and has
47 no approved drug therapies (1). MAFLD can further develop into hepatocellular
48 carcinoma (HCC) and both diseases are predicted to increase in the following
49 decades (2) not only in men but also in women (3).
50 Among lipid factors associated with progressive liver damage is also cholesterol. The
51 cellular cholesterol concentrations are fine-tuned by a balance of synthesis, uptake,
52 and efflux (4). Genes of cholesterol synthesis differ in their susceptibility to loss of
53 function mutations and can affect cell metabolism in different manners, from
54 promoting cell survival towards manifesting in malignancies (5). Linking cholesterol
55 synthesis and liver disease remains controversial. Cholesterol synthesis was shown
56 to support the growth of HCC lesions after depletion of fatty acid synthesis (6),
57 indicating a cross-talk between fatty acid and cholesterol pathways in carcinogenesis.
58 Overexpression of the squalene epoxidase enzyme of cholesterol synthesis was
59 identified as a driver of NAFLD-induced HCC (7). On the other hand, blocking
60 cholesterol synthesis at the lanosterol 14α-demethylase (CYP51) step leads to the
2 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
61 prepubertal onset of liver injury with ductular reaction and fibrosis and a pronounced
62 male prevalent liver dysfunction before the puberty (8). In young adults, the liver
63 damage was more prominent among female mice due to diminished cholesterol
64 esters and elevated sterol substrates (9).
65 The sex disparity in liver pathologies is far from conclusive. While some studies don’t
66 describe sex as a cofounding factor (6, 7, 10, 11), there is increasing evidence that
67 after the menopause MAFLD occurs at a higher rate in women compared to men
68 (12). Female and male livers are metabolically distinct organs, which was
69 demonstrated also in mathematical models of the liver with unique regulators of sex-
70 specific metabolic outcomes (13). A recent review concluded that clinical and
71 epidemiological studies frequently fail to address sex differences appropriately (12).
72 This study aimed to address the long-term effect of disrupted liver cholesterol
73 synthesis where the knock-out model of ageing mice was applied. Surprisingly, the
74 tumors developed spontaneously mainly in female ageing mice. We identified novel
75 metabolic pathways and transcription factors that can be linked to metabolism
76 associated with female-prone hepatocarcinogenesis and provide mechanistic insights
77 with relevance for humans.
78
79 RESULTS
80 (1) Long-term hepatocyte deletion of Cyp51 results in HCC. We monitored
81 the liver pathology in mice between 12 month and 24 month of age. Moderate to
82 severe ductular reaction bridging the adjacent portal tracts was observed,
83 accompanied by mild inflammation (predominantly mononuclear cells; arrows in Fig.
84 1A), and fibrosis that was more pronounced in females (Sirius red staining (SR); Fig.
3 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
85 1B). Yellow pigment likely representing lipofuscin (arrow in Suppl. Fig. 1) was
86 observed in areas of pronounced ductular reaction, sometimes surrounded by other
87 hepatic cells, evaluated as macrophages (stars in Fig. 1A).
88 Surprisingly, liver tumors were observed in 12M female and 18M male Cyp51 KO
89 mice (Fig. 2A) defined as macroscopically visible nodules (> 1 mm) of various
90 colours. At 24M the incidence of tumors was 77.8% for KO females and 50% for
91 males. In WTs, a single tumor was found in a 2-year male, histologically described as
92 adenoma (Suppl. Table 2). At 24M, the KOs showed a decreased body weight
93 compared to controls, significantly higher liver weight, and elevated relative liver
94 weights especially in females with the liver tumors (Suppl. Fig. 2 & Suppl. Table 3).
95 Tumors were histologically classified as eosinophilic/clear cell nodules, hepatocellular
96 adenomas, or HCCs (15). Immunohistochemical staining for glutamine synthetase as
97 a recently proposed marker of HCC is shown in Suppl. Fig. 3. For further molecular
98 analyses we selected only tumors histologically classified as HCC, with features such
99 as the absence of portal tracts and hepatic lobules, thickened hepatic plates, solid
100 growth, focal tubular or pseudoglandular structures, focal necrosis, numerous
101 apoptotic cells, and presence of cellular and nuclear polymorphism with brisk mitotic
102 activity (Fig. 2B, C).
103
104 (2) From metabolism associated features of MALD towards HCC in
105 females and males. The long-term ablation of Cyp51 in hepatocytes was studied at
106 the gene and metabolic levels. Deregulation of genes from cholesterol synthesis was
107 more pronounced in the Cyp51 KO females (Fig. 3A & Suppl. Tables 4 - 6).
108 Additionally, expression of Cyp8b1, Cyp7a1 and Cyp27a1 from the bile acid (BA)
109 synthesis pathway was at 24 months also decreased only in female KO mice,
4 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
110 indicating a drop in the classical pathway of BA synthesis (Suppl. Fig. 4 & Suppl.
111 Tables 4 - 6).
112 Plasma cholesterol (CHOL), free fatty acids (FFA), HDL (high-density lipoprotein),
113 and triglycerides (TG) showed opposite trends among females and males during the
114 ageing, with CHOL and HDL increasing in KO females. A significant increase in ALT
115 in female Cyp51 KO mice is an additional indication of chronic hepatic injury (Fig. 3B
116 & Suppl. Table 4).
117 In the livers, we measured CHOL and the intermediate sterols of the cholesterol
118 synthesis (Fig. 3C & Suppl. Table 4). In the KO mice with inactivated CYP51A1
119 enzyme, its substrates lanosterol and dihydrolanosterol were highly elevated (~100-
120 fold and ~500-fold, respectively), proving disrupted cholesterol synthesis.
121 Dihydrolanosterol was higher than lanosterol indicating that 24-dehydrocholesterol
122 reductase (DHCR24) enzyme functions normally. Downstream the pathway,
123 desmosterol was not statistically significantly different between Cyp51 KO and WT
124 mice, however, zymostenol was significantly higher in Cyp51 KO livers. Small or no
125 change between genotypes indicates that sterol intermediates were replaced by
126 transport to the liver or by the upregulated synthesis in other cell types.
127 For liver CHOL, sex represents a statistically significant parameter in the interaction
128 analysis between sex and age. Similarly as for blood CHOL also hepatic CHOL is
129 increased in female mice during the ageing while this was not the case for the males
130 (Fig 3B, C & Suppl. Table 4).
131
132 (3) Transcriptome analysis identified sex differences in gene expression
133 linked to hepatocarcinogenesis. We used Affymetrix microarrays for gene
5 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
134 expression profiling of 24M Cyp51 KO (24M.KO) and controls (24M.WT) mouse liver
135 samples. Data was compared to 19W Cyp51 KO (19W.KO) and controls (19W.WT)
136 (14) (Suppl. Tables 6 – 8) and discussed in the view of the early processes towards
137 carcinogenesis.
138 Hierarchical clustering of differentially expressed (DE) genes showed a clear
139 difference between 24M.WT and 24M.KO mice for both sexes (Fig. 4A) and a clear
140 sex imbalance in the number of DE genes (Fig. 4B). From these, 11.6% DE genes
141 including Trp53, Ezh2, Foxa1, and Rb1 were deregulated only in males, while 63.4%
142 DE genes including Ctnnb1, Gsk3b, Nova1, Lmna, Gstp1, and Mup20 were
143 deregulated only in females (Suppl. Tables 6, 9). Furthermore, a majority of DE
144 genes were positively enriched (56% in females and 68% in males) (Fig. 4B). Using
145 NetworkAnalyst (16), the functionally enriched networks of DE genes showed that
146 positively enriched genes were mostly involved in pathways related to cancer and
147 extracellular matrix (ECM)-receptor interaction pathway, while genes from metabolic
148 pathways were negatively enriched, especially among female KOs (Fig. 4C).
149
150 (4) Enriched KEGG pathways and transcription factors (TFs) as triggers
151 of female prevalent hepatocarcinogenesis. From 323 KEGG pathways, we
152 discovered 209 that were enriched in livers of 19W.KO mice aligned with 19W.WT
153 (Suppl. Tables 7, 10) and 230 in 24M.KO compared to 24M.WT mice (Table 1,
154 Suppl. Table 7), including the activation of HCC, apoptosis, endocytosis, PI3K/Akt
155 signalling, pathways in cancer, and proteoglycans in cancer. ECM-receptor
156 interaction pathway was observed as the most positively enriched in 19W KO
157 females (Suppl. Table 10) and 24M KO females and males (Table 1), which is in line
158 with the histological findings (Fig. 2B). Chronic injury induces numerous molecular
6 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
159 alterations in hepatocytes and causes a gradual accumulation of extracellular matrix
160 components (ECM) that might lead to a complete architectural reconstruction of
161 hepatocytes and tumor development.
162 Interestingly, the TGF-β signalling pathway was significantly elevated only in Cyp51
163 KO female livers (19W: Suppl. Table 10; 24M: Table 1). Additionally, the Wnt
164 signalling and circadian entrainment pathways were activated only in KO females
165 (Table 1). We further explored if these pathways could serve as triggers for female
166 hepatocarcinogenesis in the Cyp51 KO model. The selected protein markers of the
167 TGF-β and Wnt signalling were evaluated by qPCR and IHC. Here TGF-β showed a
168 positive reaction, particularly in female Cyp51 KO mice (Suppl. Fig. 5, 6 & Suppl.
169 Table 11). Biosynthesis of unsaturated fatty acids, fatty acid degradation, and
170 peroxisome pathways were negatively enriched in KO females at 19W and 24M as
171 well as males at 24M KO (19W: Suppl. Table 10; 24M: Table 1).
172 Considering the general alteration of DE genes and KEGG pathways in Cyp51 KO
173 mice, the transcription landscape was investigated. The TF enrichment analysis
174 indicated 321 and 142 enriched TFs in 19W.KO and 24M.KO in comparison to WT
175 littermates (Suppl. Table 8). The enriched TFs with specific functions are listed in
176 Suppl. Table 10 (19W), and Table 1 (24M). In both groups, the transcriptional
177 activity of RORC was negatively enriched (at 19W.KO in females only), while SOX9,
178 SP1, AP1, and C-JUN TFs activities were positively enriched. The oncogenic isoform
179 of SOX9 ((Sox9)2) was positively enriched only among females. The expression of
180 the majority of RORC and SOX9 target genes was decreased in Cyp51 KO mice,
181 while Cd36 in Cyp51 KO 19W males and 24M of both sexes was positively enriched
182 (Suppl. Table 12). Additionally, the RORC gene target G6pc, which is involved in the
7 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
183 regulation of the PI3K/Akt signalling pathway, was significantly negatively enriched in
184 female KO mice (Suppl. Table 12).
185 The TF networks were visualized with STRING (17) to determine the interactions.
186 Different colours of nodes represent biological processes in which enriched TFs of
187 Cyp51 KO mice are involved. Surprisingly, SP1 and JUND expressed in both sexes
188 represent nodules with most protein-protein associations linked to
189 hepatocarcinogenesis of Cyp51 KO mice (Suppl. Fig. 7).
190
191 DISCUSSION
192 The hallmark of cancer cells is metabolic reprogramming, causing new cellular
193 demands in selective survival and growth (18). Cholesterol is a precursor of steroid
194 hormones and bile acids and an essential component of cellular membranes. Its
195 circulating levels are under healthy homeostatic conditions regulated by a balance
196 between cellular cholesterol synthesis, dietary intake, and removal of excess
197 cholesterol (19).
198 CYP51A1 is a gene from the late (post-squalene) part of the pathway (20). The
199 disruption of this gene in hepatocytes starts with morphological alterations, such as
200 ductular reaction accompanied by mild inflammation, and can end with malignant
201 liver tumors in ageing mice that are first observed at 12M in females and 18M in
202 males, with the highest incidence in the 24M female Cyp51 KO mice. Also, humans
203 show sex-dependent differences in liver pathologies. Women more commonly
204 present with acute liver failure, benign liver lesions, autoimmune hepatitis, and toxin-
205 mediated hepatotoxicity. Oppositely, malignant liver tumors, and viral hepatitis are
206 less commonly observed in women (21). However, non-alcoholic steatohepatitis
8 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
207 (NASH) as a more serious form of MALD, more commonly progresses to HCC in
208 women (22). Generally, women are less prone to developing HCC than men (due to
209 hormone-dependent resistance), while after the menopause a sharp increase in HCC
210 incidence is witnessed (23).
211 The latest studies indicate that post-squalene intermediates of cholesterol synthesis
212 are also signalling molecules. Inhibiting different enzymes from the pathway results in
213 the accumulation of sterol intermediates with different cellular effects. Sterol products
214 can activate RORC signalling (24), promote oligodendrocyte formation (25), or
215 activate LXRα and inhibit EGFR signalling (26). Inhibitors of lanosterol synthase were
216 identified as potential anticancer drug targets (27) while the accumulation of
217 lanosterol can cause degradation of 3-hydroxy-3-methyl-glutaryl-coenzyme A
218 reductase (HMGCR) (28). Our sterol measurements show significant accumulation of
219 CYP51A1 substrates lanosterol and dihydrolanosterol, with differences between
220 males and females regarding lanosterol. The main difference between the genotypes
221 is likely in the accumulation of sterol substrates rather than a lack of downstream
222 sterols, while the difference between the sexes is shown by the sex-dependent
223 transcriptome response to sterol imbalance. Due to the lack of sterols between
224 lanosterol and zymosterol as proposed RORC natural ligands, we observe at 24M
225 the repression of the RORC signalling pathway in both sexes. The RORC-dependent
226 mechanisms in cancerogenesis were recently described. Oh et al. reported that
227 RORC expression is decreased in basal-like subtype cancers and inversely
228 correlated with histological grade and cancer drivers in breast cancer cohorts (29). In
229 line with our findings, negatively enriched RORC could inversely regulate the TGF-
230 β/EMT signalling in Cyp51 KO mice.
9 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
231 Injured hepatocytes in Cyp51 KO mice produce different ECMs, exemplified by
232 enrichment of laminin and collagen genes (Col5a1, Col4a1, Lamc1, etc.) and strongly
233 positively enriched ECM-receptor interaction pathway in KOs of both sexes. In some
234 way, these alterations could contribute to microenvironmental reorganization and
235 induction of epithelial-mesenchymal transition (EMT) that positively influences tumor
236 progression (30). A higher rate of cancer progression in female Cyp51 KO mice could
237 be explained by the female-specific interplay of positively enriched TGF-β signalling.
238 An activated TGF-β represents a central regulator of chronic liver disease and could
239 contribute to the progression of all disease stages in animal models and humans
240 (31). Interestingly, important crosstalk between TGF-β and Cyp51 gene expression
241 has been shown in the Tgf-βfl/fl; Wnt-Cre mouse model (32).
242 Furthermore, the results of our analysis identified the key transcriptional regulators of
243 DE molecular pathways: positively enriched SOX9, SP1, and negatively enriched
244 RORC. SOX9 is an important candidate for the cancer marker required for the
245 activation of the TGF-β/Smad signalling (33). As SOX9, the tumor progression driver
246 SP1 is also a key factor for the induction of TGF-β during inflammation (34) and its
247 overexpression in Cyp51 KO relates to EMT and migration of pancreatic cancer cells
248 (35). The impact of higher disease progression in female Cyp51 KO mice is reflected
249 also in the interplay of transcription regulator SMAD3 on TGF-β signalling (36).
250 Increased plasma CHOL and HDL, and decreased plasma TG and FFA, are
251 observed only in female KO mice. This is a possible effect of induced sex hormone-
252 binding globulin (SHBG) by specific targets of an estrogen signalling pathway (37).
253 SHBG is positively correlated with HDL and negatively with TGs and has been
254 already associated with different cancers in ageing humans (38, 39). Supporting
255 other hepatocarcinogenesis studies (40, 41), the outcome of altered cholesterol and
10 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
256 BA homeostasis was a chronic liver injury in our Cyp51 KO mouse model with
257 widespread fibrosis, inflammation, and malignant transformation due to toxic sterol
258 accumulation. Additionally, sterol accumulation, in our case the consequence of
259 inhibition of CYP51A1 (14), could be a cause of the observed widespread increase of
260 hepatic progenitors (also known as a ductular reaction) that may also differentiate
261 into mature hepatocyte-like cells with oncogenic potential (42).
262 Based on our data we propose that the repression of CYP51A1 enzyme can
263 contribute to HCC development. In humans, whole-body deletion of CYP51A1 is not
264 reported, since this gene is essential and is embryonically lethal in the mouse (43).
265 Only two studies are reporting the CYP51A1 deleterious mutations, which resulted in
266 infant liver failure and death (44, 45). Numerous polymorphisms in CYP51A1 have
267 been associated with different phenotypes in humans (46), but the consequences of
268 CYP51A1 heterozygosity in human hepatocarcinogenesis need yet to be established.
269 CYP51A1 has a low mutation rate, lower than other cholesterol synthesis genes (47).
270 The CYP51A1 rs37417517 variant was found associated with the female-specific
271 breast cancer and variant rs229188 with the women ageing genetics (48, 49).
272 Additionally, the COSMIC database revealed the presence of somatic CYP51A1
273 point mutations in tumor samples, some also with predicted deleterious effect on
274 activity (Suppl. Tables 13, 14).
275
11 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
276 CONCLUSIONS
277 The hallmark of cancer cells is metabolic reprogramming, causing new cellular
278 demands in selective survival and growth. Innactivation of CYP51A1 from cholesterol
279 synthesis provokes global transcriptome changes. Multiple signalling pathways and
280 transcription factors are deregulated in a sex-dependent manner (Suppl. Fig. 8)
281 which leads towards metabolic-related HCC. Female-specific cholesterol-related
282 mechanism of hepatocarcinogenesis was aligned with human data (Fig. 5) and is in
283 line with the fact that the liver is a sex-dimorphic metabolic organ and a major organ
284 of cholesterol homeostasis. A novel finding of our work is that cholesterol synthesis
285 deregulation due to repression of Cyp51 initially results in NASH-like features that
286 progress towards hepatocellular carcinoma in a sex-specific manner, with more
287 aggrevated phenotype and specific signalling pathways in the aging females.
288
12 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
289 MATERIALS AND METHODS
290 Detailed materials and methods are described in Supplementary Materials and
291 methods. The experimental design of the study is shown in the Suppl. Fig. M1.
292 Animals. The homozygous hepatocyte-specific Cyp51 knock-out (Cyp51 KO) and
293 wild-type (Cyp51 WT) mice of both sexes were used following relevant guidelines
294 and regulations. The mice were sacrificed with cervical dislocation at 12, 18, and 24
295 months (M). Organs were removed, patho-morphology examined, freshly frozen, and
296 stored at – 80 °C.
297 Ethics statement. Mouse experiments were approved by the Veterinary
298 Administration of the Republic of Slovenia (license numbers 34401-31/2011/4 and
299 34401-52/2012/3) and were conducted in agreement with the National Institute of
300 Health guidelines for work with laboratory animals, national legislation and Directive
301 2010/63/EU on the protection of animals used for scientific purposes.
302 Histology. Hematoxylin and eosin (H&E) staining, Sirius red (SR) staining and
303 immunohistochemical reaction for glutamine synthetase, TGF-β and β-catenin were
304 carried out on 4-5 µm formalin-fixed and paraffin-embedded liver sections.
305 Analysis of plasma parameters. The concentration of lipids (total and HDL
306 cholesterol, FFA, TG) and activity of liver enzymes (ALT and AST) were measured in
307 plasma by commercial kits.
308 Liver sterol analysis. Sterol intermediates were isolated as described (50) and
309 identified and quantified by LC-MS. After normalization based on internal standard,
310 the concentration of individual sterols was calculated corresponding to the standard
311 sterols from Avanti Lipids.
13 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
312 Gene expression profiling of liver samples of 24M old Cyp51 WT and KO mice of
313 both sexes was performed using Affymetrix GeneChip™ Mouse Gene 2.0 ST arrays
314 and validated by qPCR. Samples of 20 mice were obtained from normal livers as
315 controls (wild-type, WT) and macroscopically visible liver tumors with surrounding
316 tissue as KO. The data were analysed using R/limma controlling false discovery rate
317 at α=0.05 and deposited to GEO under accession number GSE127772. KEGG
318 PATHWAY and TRANSFAC databases were used for functional enrichment studies.
319 Gene expression data from 19 weeks (W) (14) WT and KO mice of both sexes were
320 used to determine processes that contribute to tumor progression from 19W towards
321 24M. Network diagrams were created using NetworkAnalyst and STRING. The
322 proposed mechanistic scheme was designed from data obtained from mice
323 experiments and aligned to human literature data to increase their relevance.
324
14 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
325 DECLARATIONS
326 Ethical Approval and Consent to participate
327 None.
328 Consent for publication
329 None.
330 Availability of supporting data
331 The datasets generated during the current study are available in the Gene
332 Expression Omnibus database (https://www.ncbi.nlm.nih.gov/geo/) under accession
333 number GSE127772 and GSE58271.
334 Competing interest
335 The authors declare no competing financial interests.
336 Funding
337 This work was supported by the Slovenian Research Agency (ARRS) program grants
338 P1-0104, P1-0390, J1-9176. K. Blagotinšek Cokan, G. Lorbek, and Ž. Urlep is/were
339 supported by the graduate fellowship of ARRS.
340 Author contributions
341 Study design and supervision: D.R.; Experiment conduct: K.B.C., G.L., M.P., C.S.;
342 Data analysis: P.J.; Data evaluation and interpretation: K.B.C., M.P., J.J., C.S, D.R.;
343 Material support: M.M.S.; Writing of the manuscript: K.B.C., M.P., J.J., T.R., P.J.,
344 D.R. All the authors contributed to the final version of the manuscript.
345
15 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
346 Acknowledgements
347 We thank K. Kodra, H. Klavžar, M. Kužnik, and D. Mahn for their help with
348 experiments. We acknowledge S. Horvat and R. Keber for their previous contribution
349 to the development of the Cyp51 knock-out mouse model. Thanks also to N. Nadižar
350 for help in statistical data evaluation and figures preparation and P. Ivanuša for help
351 with sterol isolation. Thanks to collaborators I. Vovk and M. Križman for LC-MS
352 method development.
353
354 REFERENCES
355 1. Eslam M, Sanyal AJ, George J, an international consensus p. MAFLD: A consensus-driven 356 proposed nomenclature for metabolic associated fatty liver disease. Gastroenterology. 2020 Feb 7. 357 PubMed PMID: 32044314. 358 2. Nakagawa H, Hayata Y, Kawamura S, Yamada T, Fujiwara N, Koike K. Lipid Metabolic 359 Reprogramming in Hepatocellular Carcinoma. Cancers (Basel). 2018 Nov 15;10(11). PubMed PMID: 360 30445800. Pubmed Central PMCID: 6265967. 361 3. Burra P, De Martin E, Gitto S, Villa E. Influence of age and gender before and after liver 362 transplantation. Liver transplantation : official publication of the American Association for the Study 363 of Liver Diseases and the International Liver Transplantation Society. 2013 Feb;19(2):122-34. PubMed 364 PMID: 23172830. 365 4. Chua NK, Coates HW, Brown AJ. Cholesterol, cancer, and rebooting a treatment for athlete's 366 foot. Science translational medicine. 2018 Apr 18;10(437). PubMed PMID: 29669852. 367 5. Horvat S, McWhir J, Rozman D. Defects in cholesterol synthesis genes in mouse and in 368 humans: lessons for drug development and safer treatments. Drug metabolism reviews. 2011 369 Feb;43(1):69-90. PubMed PMID: 21247357. 370 6. Che L, Chi W, Qiao Y, Zhang J, Song X, Liu Y, et al. Cholesterol biosynthesis supports the 371 growth of hepatocarcinoma lesions depleted of fatty acid synthase in mice and humans. Gut. 2020 372 Jan;69(1):177-86. PubMed PMID: 30954949. Pubmed Central PMCID: 6943247. 373 7. Liu D, Wong CC, Fu L, Chen H, Zhao L, Li C, et al. Squalene epoxidase drives NAFLD-induced 374 hepatocellular carcinoma and is a pharmaceutical target. Science translational medicine. 2018 Apr 375 18;10(437). PubMed PMID: 29669855. 376 8. Urlep Z, Lorbek G, Perse M, Jeruc J, Juvan P, Matz-Soja M, et al. Disrupting Hepatocyte Cyp51 377 from Cholesterol Synthesis Leads to Progressive Liver Injury in the Developing Mouse and Decreases 378 RORC Signalling. Scientific reports. 2017 Jan 18;7:40775. PubMed PMID: 28098217. Pubmed Central 379 PMCID: 5241696. 380 9. Lorbek G, Urlep Z, Rozman D. Pharmacogenomic and personalized approaches to tackle 381 nonalcoholic fatty liver disease. Pharmacogenomics. 2016 Jul;17(11):1273-88. PubMed PMID: 382 27377717. 383 10. Wilson MA, Buetow KH. Novel Mechanisms of Cancer Emerge When Accounting for Sex as a 384 Biological Variable. Cancer research. 2020 Jan 1;80(1):27-9. PubMed PMID: 31722998. Pubmed 385 Central PMCID: 6942614.
16 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
386 11. Li CH, Haider S, Shiah YJ, Thai K, Boutros PC. Sex Differences in Cancer Driver Genes and 387 Biomarkers. Cancer research. 2018 Oct 1;78(19):5527-37. PubMed PMID: 30275052. 388 12. Lonardo A, Nascimbeni F, Ballestri S, Fairweather D, Win S, Than TA, et al. Sex Differences in 389 Nonalcoholic Fatty Liver Disease: State of the Art and Identification of Research Gaps. Hepatology. 390 2019 Oct;70(4):1457-69. PubMed PMID: 30924946. Pubmed Central PMCID: 6766425. 391 13. Cvitanovic Tomas T, Urlep Z, Moskon M, Mraz M, Rozman D. LiverSex Computational Model: 392 Sexual Aspects in Hepatic Metabolism and Abnormalities. Frontiers in physiology. 2018;9:360. 393 PubMed PMID: 29706895. Pubmed Central PMCID: 5907313. 394 14. Lorbek G, Perse M, Jeruc J, Juvan P, Gutierrez-Mariscal FM, Lewinska M, et al. Lessons from 395 hepatocyte-specific Cyp51 knockout mice: impaired cholesterol synthesis leads to oval cell-driven 396 liver injury. Scientific reports. 2015 Mar 5;5:8777. PubMed PMID: 25739789. Pubmed Central PMCID: 397 4350092. 398 15. Deschl U, Cattley RC, Harada T, Kuttler K, Hailey JR, Harting F, et al. Liver, Gallbladder, and 399 Exocrine Pancreas. In: Mohr U, editor. International Classification of Rodent Tumors: The Mouse. 400 Verlag Berlin Heidelberg new York: Springer; 2001. p. p.59-86. 401 16. Xia J, Benner MJ, Hancock RE. NetworkAnalyst--integrative approaches for protein-protein 402 interaction network analysis and visual exploration. Nucleic acids research. 2014 Jul;42(Web Server 403 issue):W167-74. PubMed PMID: 24861621. Pubmed Central PMCID: 4086107. 404 17. Szklarczyk D, Morris JH, Cook H, Kuhn M, Wyder S, Simonovic M, et al. The STRING database 405 in 2017: quality-controlled protein-protein association networks, made broadly accessible. Nucleic 406 acids research. 2017 Jan 4;45(D1):D362-D8. PubMed PMID: 27924014. Pubmed Central PMCID: 407 5210637. 408 18. Infantino V, Dituri F, Convertini P, Santarsiero A, Palmieri F, Todisco S, et al. Epigenetic 409 upregulation and functional role of the mitochondrial aspartate/glutamate carrier isoform 1 in 410 hepatocellular carcinoma. Biochimica et biophysica acta Molecular basis of disease. 2019 411 Jan;1865(1):38-47. PubMed PMID: 30321589. 412 19. Cruz PM, Mo H, McConathy WJ, Sabnis N, Lacko AG. The role of cholesterol metabolism and 413 cholesterol transport in carcinogenesis: a review of scientific findings, relevant to future cancer 414 therapeutics. Frontiers in pharmacology. 2013 Sep 25;4:119. PubMed PMID: 24093019. Pubmed 415 Central PMCID: 3782849. 416 20. Acimovic J, Rozman D. Steroidal triterpenes of cholesterol synthesis. Molecules. 2013 Apr 417 4;18(4):4002-17. PubMed PMID: 23558541. Pubmed Central PMCID: 6269928. 418 21. Guy J, Peters MG. Liver disease in women: the influence of gender on epidemiology, natural 419 history, and patient outcomes. Gastroenterol Hepatol (N Y). 2013 Oct;9(10):633-9. PubMed PMID: 420 24764777. Pubmed Central PMCID: 3992057. 421 22. Kohi MP. Gender-Related Differences in Hepatocellular Carcinoma: Does Sex Matter? Journal 422 of vascular and interventional radiology : JVIR. 2016 Sep;27(9):1338-41. PubMed PMID: 27566425. 423 23. Hartwell HJ, Petrosky KY, Fox JG, Horseman ND, Rogers AB. Prolactin prevents hepatocellular 424 carcinoma by restricting innate immune activation of c-Myc in mice. Proceedings of the National 425 Academy of Sciences of the United States of America. 2014 Aug 5;111(31):11455-60. PubMed PMID: 426 25049387. Pubmed Central PMCID: 4128121. 427 24. Santori FR, Huang P, van de Pavert SA, Douglass EF, Jr., Leaver DJ, Haubrich BA, et al. 428 Identification of natural RORgamma ligands that regulate the development of lymphoid cells. Cell 429 metabolism. 2015 Feb 3;21(2):286-98. PubMed PMID: 25651181. Pubmed Central PMCID: 4317570. 430 25. Hubler Z, Allimuthu D, Bederman I, Elitt MS, Madhavan M, Allan KC, et al. Accumulation of 431 8,9-unsaturated sterols drives oligodendrocyte formation and remyelination. Nature. 2018 432 Aug;560(7718):372-6. PubMed PMID: 30046109. 433 26. Gabitova L, Restifo D, Gorin A, Manocha K, Handorf E, Yang DH, et al. Endogenous Sterol 434 Metabolites Regulate Growth of EGFR/KRAS-Dependent Tumors via LXR. Cell reports. 2015 Sep 435 22;12(11):1927-38. PubMed PMID: 26344763. Pubmed Central PMCID: 4581991.
17 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
436 27. Phillips RE, Yang Y, Smith RC, Thompson BM, Yamasaki T, Soto-Feliciano YM, et al. Target 437 identification reveals lanosterol synthase as a vulnerability in glioma. Proceedings of the National 438 Academy of Sciences of the United States of America. 2019 Mar 28. PubMed PMID: 30923116. 439 28. Chen L, Ma MY, Sun M, Jiang LY, Zhao XT, Fang XX, et al. Endogenous sterol intermediates of 440 the mevalonate pathway regulate HMGCR degradation and SREBP-2 processing. Journal of lipid 441 research. 2019 Oct;60(10):1765-75. PubMed PMID: 31455613. Pubmed Central PMCID: 6795084. 442 29. Oh TG, Wang SM, Acharya BR, Goode JM, Graham JD, Clarke CL, et al. The Nuclear Receptor, 443 RORgamma, Regulates Pathways Necessary for Breast Cancer Metastasis. EBioMedicine. 2016 444 Apr;6:59-72. PubMed PMID: 27211549. Pubmed Central PMCID: 4856749. 445 30. Yang JD, Nakamura I, Roberts LR. The tumor microenvironment in hepatocellular carcinoma: 446 current status and therapeutic targets. Seminars in cancer biology. 2011 Feb;21(1):35-43. PubMed 447 PMID: 20946957. Pubmed Central PMCID: 3050428. 448 31. Dooley S, ten Dijke P. TGF-beta in progression of liver disease. Cell and tissue research. 2012 449 Jan;347(1):245-56. PubMed PMID: 22006249. Pubmed Central PMCID: 3250614. 450 32. Pelikan RC, Iwata J, Suzuki A, Chai Y, Hacia JG. Identification of candidate downstream targets 451 of TGFbeta signaling during palate development by genome-wide transcript profiling. Journal of 452 cellular biochemistry. 2013 Apr;114(4):796-807. PubMed PMID: 23060211. Pubmed Central PMCID: 453 3777336. 454 33. Matheu A, Collado M, Wise C, Manterola L, Cekaite L, Tye AJ, et al. Oncogenicity of the 455 developmental transcription factor Sox9. Cancer research. 2012 Mar 1;72(5):1301-15. PubMed PMID: 456 22246670. Pubmed Central PMCID: 3378515. 457 34. Vizcaino C, Mansilla S, Portugal J. Sp1 transcription factor: A long-standing target in cancer 458 chemotherapy. Pharmacology & therapeutics. 2015 Aug;152:111-24. PubMed PMID: 25960131. 459 35. Jungert K, Buck A, von Wichert G, Adler G, Konig A, Buchholz M, et al. Sp1 is required for 460 transforming growth factor-beta-induced mesenchymal transition and migration in pancreatic cancer 461 cells. Cancer research. 2007 Feb 15;67(4):1563-70. PubMed PMID: 17308095. 462 36. Vervoort SJ, Lourenco AR, Tufegdzic Vidakovic A, Mocholi E, Sandoval JL, Rueda OM, et al. 463 SOX4 can redirect TGF-beta-mediated SMAD3-transcriptional output in a context-dependent manner 464 to promote tumorigenesis. Nucleic acids research. 2018 Oct 12;46(18):9578-90. PubMed PMID: 465 30137431. Pubmed Central PMCID: 6182182. 466 37. Caldwell JD, Jirikowski GF. Sex hormone binding globulin and aging. Hormone and metabolic 467 research = Hormon- und Stoffwechselforschung = Hormones et metabolisme. 2009 Mar;41(3):173- 468 82. PubMed PMID: 18956301. 469 38. Hautanen A. Synthesis and regulation of sex hormone-binding globulin in obesity. 470 International journal of obesity and related metabolic disorders : journal of the International 471 Association for the Study of Obesity. 2000 Jun;24 Suppl 2:S64-70. PubMed PMID: 10997612. 472 39. Vaidya D. Sex hormones and circulating lipoprotein levels. Future Lipidology. 2008;3(6):603- 473 6. 474 40. Cichoz-Lach H, Michalak A. Oxidative stress as a crucial factor in liver diseases. World journal 475 of gastroenterology. 2014 Jul 7;20(25):8082-91. PubMed PMID: 25009380. Pubmed Central PMCID: 476 4081679. 477 41. Wu XZ, Chen D. Origin of hepatocellular carcinoma: role of stem cells. Journal of 478 gastroenterology and hepatology. 2006 Jul;21(7):1093-8. PubMed PMID: 16824058. 479 42. Fang CH, Gong JQ, Zhang W. Function of oval cells in hepatocellular carcinoma in rats. World 480 journal of gastroenterology. 2004 Sep 1;10(17):2482-7. PubMed PMID: 15300889. Pubmed Central 481 PMCID: 4572146. 482 43. Keber R, Motaln H, Wagner KD, Debeljak N, Rassoulzadegan M, Acimovic J, et al. Mouse 483 knockout of the cholesterogenic cytochrome P450 lanosterol 14alpha-demethylase (Cyp51) 484 resembles Antley-Bixler syndrome. The Journal of biological chemistry. 2011 Aug 19;286(33):29086- 485 97. PubMed PMID: 21705796. Pubmed Central PMCID: 3190716.
18 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
486 44. Gillespie RL, Urquhart J, Anderson B, Williams S, Waller S, Ashworth J, et al. Next-generation 487 Sequencing in the Diagnosis of Metabolic Disease Marked by Pediatric Cataract. Ophthalmology. 488 2016 Jan;123(1):217-20. PubMed PMID: 26233629. 489 45. Patel N, Anand D, Monies D, Maddirevula S, Khan AO, Algoufi T, et al. Novel phenotypes and 490 loci identified through clinical genomics approaches to pediatric cataract. Human genetics. 2017 491 Feb;136(2):205-25. PubMed PMID: 27878435. Pubmed Central PMCID: 5783298. 492 46. Rezen T, Ogris I, Sever M, Merzel F, Golic Grdadolnik S, Rozman D. Evaluation of Selected 493 CYP51A1 Polymorphisms in View of Interactions with Substrate and Redox Partner. Frontiers in 494 pharmacology. 2017;8:417. PubMed PMID: 28713270. Pubmed Central PMCID: 5492350. 495 47. Lewinska M, Zelenko U, Merzel F, Golic Grdadolnik S, Murray JC, Rozman D. Polymorphisms 496 of CYP51A1 from cholesterol synthesis: associations with birth weight and maternal lipid levels and 497 impact on CYP51 protein structure. PloS one. 2013;8(12):e82554. PubMed PMID: 24358204. Pubmed 498 Central PMCID: 3866192. 499 48. Michailidou K, Lindstrom S, Dennis J, Beesley J, Hui S, Kar S, et al. Association analysis 500 identifies 65 new breast cancer risk loci. Nature. 2017 Nov 2;551(7678):92-4. PubMed PMID: 501 29059683. Pubmed Central PMCID: 5798588. 502 49. Yashin AI, Wu D, Arbeeva LS, Arbeev KG, Kulminski AM, Akushevich I, et al. Genetics of aging, 503 health, and survival: dynamic regulation of human longevity related traits. Frontiers in genetics. 504 2015;6:122. PubMed PMID: 25918517. Pubmed Central PMCID: 4394697. 505 50. Acimovic J, Lovgren-Sandblom A, Monostory K, Rozman D, Golicnik M, Lutjohann D, et al. 506 Combined gas chromatographic/mass spectrometric analysis of cholesterol precursors and plant 507 sterols in cultured cells. Journal of chromatography B, Analytical technologies in the biomedical and 508 life sciences. 2009 Jul 15;877(22):2081-6. PubMed PMID: 19525158. 509 510
19 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
511 FIGURE LEGENDS
512 Figure 1. Blocking cholesterol synthesis in KOs resulted in histological
513 alterations of the liver, particularly among females. (A) Histological examination
514 revealed ductular reaction (*) and inflammation (→) (H&E, original magnification
515 x200), N=5-10. (B) Sirius red stain (SR, original magnification x200) showed more
516 pronounced fibrosis in female transgenic mice, N=3-5. M, month; KO, Cyp51 KO;
517 WT, Cyp51 WT.
518
519 Figure 2. Blocking cholesterol synthesis in KOs resulted in
520 hepatocarcinogenesis, the bigger tumor histologically consistent with HCCs.
521 (A) Representative gross image of liver tumors in 24M KO mice. (B) Representative
522 histologic image of HCC showing different architectural abnormalities. (C) Higher
523 magnification showing cellular and nuclear polymorphism with brisk mitotic activity.
524 Hematoxylin and Eosin (H&E), original magnification A, B x100, C x400. KO, Cyp51
525 KO; T, tumor.
526
527 Figure 3. Age profiles of the compensatory changes in cholesterol
528 homeostasis in KOs and WTs. (A) Expression of hepatic genes from cholesterol
529 synthesis N = 5 mice per sex/genotype/age group. (B) Plasma biochemical
530 parameters, N=5-10 mice per sex/genotype/age group. (C) Levels of selected hepatic
531 sterol intermediates and cholesterol, N=4-6 mice per sex/genotype/age group. Light
532 color bands - 95% confidence interval. Dots - individual measurements. KO, Cyp51
533 KO; WT, Cyp51 WT, F = female; M = male.
534
20 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
535 Figure 4. Target genes associated with blocking cholesterol synthesis. (A)
536 Heat-map of top 50 DE transcripts that showed significant differences in at least one
537 comparison: hierarchical clustering distinguished between the genotypes and sex of
538 24M.KO~WT mouse livers. (B) Stacked column plots with numbers of differentially
539 up- and down-regulated sex-specific genes of 24M Cyp51 KO mice. (C) The
540 functional enriched networks of up- and down-regulated DE genes in female and
541 male 24M.KO~WT mice using NetworkAnalyst. The size of nodes corresponds to the
542 number of DE genes within a particular molecular process. All pathway nodules are
543 differentially expressed (p<0.05), the gradation rising from red to yellow (most
544 enriched) in the network of positively (left) and negatively (right) enriched molecular
545 processes. Circles representing cancer pathways and ECM-receptor interaction are
546 surrounded by green lines.
547
548 Figure 5. Sex-dependent metabolic and transcriptional changes after disrupted
549 Cyp51 from cholesterol synthesis align with hepatocarcinogenesis. CYP51A1
550 enzyme was repressed in hepatocytes of both sexes. However, chronic metabolic
551 insult resulted in a variety of sex-specific metabolic changes. In the liver of both
552 sexes at 24 months, we observed accumulation of hepatic lanosterol and
553 dihydrolanosterol (metabolites of CYP51A1 enzyme), while hepatic cholesterol is
554 increased in females and decreased in males, both WTs and KOs. There is a
555 substantial negatively enrichment of TFs related to metabolism in female Cyp51 KO
556 mice, such as the NR1B1:RXR , LXR :RXR, LRH1, and FXR. The carcinogenesis
557 related signalling pathways (ECM-receptor interaction and PI3K/Akt) are positively
558 enriched, with sex-specific gene targets. The activated TGF-β and estrogen
559 signalling pathways worsen the phenotype, particularly among females. Increased
21 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
560 plasma CHOL and HDL, and decreased plasma TGs and FFA, are observed only in
561 female KO mice.
562 Arrows indicate active connections between enriched genes, KEGG pathways, and
563 TFs. Blocked arrows represent repression. Hatched arrows represent possible
564 transcription regulation of the selected pathway or connection between selected
565 pathways in KO mice. Metabolites, DE genes, KEGG pathways, and TFs are violet if
566 positively enriched and green if negatively enriched. Icons were used from the
567 Reactome icon library. All selected mouse signalling and metabolic pathways and
568 TFs were aligned with human literature data.
569
570 TABLE LEGEND
571 Table 1. The most significantly enriched KEGG pathways and their regulation
572 by DE TFs in 24M KOs. Parameters shown in bold are statistically significant. Dark
573 green logFC - the most negative enriched pathway/TF, dark violet logFC – the most
574 positive enriched pathway/TF. KEGG, Kyoto Encyclopaedia of Genes and Genomes;
575 * p<0.05, ** p<0.01, *** p<0.001.
576
22 bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. bioRxiv preprint doi: https://doi.org/10.1101/2020.09.03.280974; this version posted September 4, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
Table 1. The most significantly enriched KEGG pathways and their regulation
by DE TFs in 24M KOs. Parameters shown in bold are statistically significant. Dark
green logFC - the most negative enriched pathway/TF, dark violet logFC – the most
positive enriched pathway/TF. KEGG, Kyoto Encyclopaedia of Genes and Genomes;
* p<0.05, ** p<0.01, *** p<0.001.
24M.KO~WT 24M.KO~WT FEMALES MALES TRANSCRIPTION FEMALES MALES KEGG PATHWAY adj. adj. FACTOR adj. adj. LogFC LogFC LogFC LogFC P-value P-value P-value P-value <0.001 <0.001 <0.001 <0.001 Cell cycle 1.60 2.13 SP1 1.47 1.49 *** *** *** *** <0.001 <0.001 <0.001 <0.001 Pathways in cancer 1.89 1.84 C-JUN 1.42 1.50 *** *** *** *** ECM-receptor <0.001 <0.001 <0.001 <0.001 1.82 1.85 JUND 1.05 0.91 interaction *** *** *** *** Proteoglycans in <0.001 <0.001 <0.001 <0.001 1.63 1.61 C-FOS 1.47 1.33 cancer *** *** *** *** <0.001 <0.001 <0.001 <0.001 Apoptosis 1.65 1.49 AP1 1.17 0.88 *** *** *** *** <0.001 <0.001 <0.001 0.004 p53 signalling pathway 1.26 1.33 C/EBPβ 0.91 0.87 *** *** *** ** <0.001 <0.001 <0.001 <0.001 MicroRNAs in cancer 0.87 0.95 Egr1 0.95 1.03 *** *** *** *** Hepatocellular <0.001 0.006 <0.001 0.001 0.95 0.72 SOX9 0.90 0.63 carcinoma *** ** *** ** 0.020 0.005 Circadian entrainment 0.42 0.27 0.197 (SOX9)2 0.67 0.50 0.062 * ** 0.007 0.008 0.004 TGF-β pathway 0.45 0.14 0.427 HIF1α 0.43 0.58 ** ** ** 0.035 Wnt signalling pathway 0.32 0.018 * 0.24 0.118 NFATC1 0.36 0.29 0.156 * PI3K/Akt signalling <0.001 <0.001 0.031 0.009 1.61 1.51 CLOCK -0.26 -0.41 pathway *** *** * ** 0.001 0.009 0.043 Fatty acid degradation -1.18 -1.12 LXRα:RXRα -0.56 -0.19 0.602 ** ** * 0.009 <0.001 0.019 Metabolic pathways -1.80 -0.69 0.367 α -1.60 -1.17 ** HNF4 *** * <0.001 0.018 Peroxisome -1.82 -1.20 RORα *** * Primary bile acid 0.002 0.004 <0.001 <0.001 -1.29 -1.40 RORC -1.68 -2.10 biosynthesis ** ** *** *** Biosynthesis of 0.018 <0.001 <0.001 -0.87 -0.46 0.261 NR1B1 -0.80 -0.59 unsaturated fatty acid * *** ***