Impact of sacubitril/valsartan on cardiac and systemic hypoxia in chronic heart failure

Summary In heart failure patients with reduced ejection fraction, Sacubitril/valsartan (S/V) increased proBNP T71 glycosylation, which is regulated negatively by hypoxia via miR-30a in vitro. Using a cohort of 73 HFrEF patients who were transitioned from standard HF medication to S/V, we found that the increase in proBNP T71 glycosylation after S/V was associated with a decrease in cardiac hypoxia. We further found that plasma levels of K709-acteylated HIF1α, HIF-regulated and HIF-independent biomarkers also evolved consistently with a decrease in hypoxia. We further confirmed that biomarker changes were related to hypoxia, in a rat model subjected to isobaric hypoxia. We measured them in rats subjected to isobaric hypoxia. Overall, these data strongly suggest that optimally treated HFrEF patients exhibited subclinical hypoxia that is improved by S/V. The data also posit proBNP T71 glycosylation as a biomarker of cardiac hypoxia.


INTRODUCTION
Sacubitril/valsartan (S/V) is the first-in-class combination of angiotensin receptor blocker and neprilysin inhibitor (ARNi) that has markedly improved HF natural history and outcome in patients with heart failure with reduced ejection fraction (HFrEF). 1 Despite this impressive improvement in outcome, the mode of action of S/V is not fully understood.We previously showed that S/V provoked an increase in proBNP glycosylation at threonine 71 (T71) 2 between before and D30 after the initiation of S/V.ProBNP T71 glycosylation is initiated by the N-acetylglucosaminyltransferase 1/2 (GalNac-T1/2) 3 and is restricted to ventricular cardiomyocytes in the heart in vivo. 4It has recently been shown that GalNac-T1/2 expression is negatively regulated by miRNAs of the miR-30 family, 3 of which miR-30a was shown to be induced by hypoxia in cardiomyocytes in vitro. 5ypoxia is defined as a mismatch between oxygen requirement and delivery to cells.Hypoxia could be attributed to inadequate cardiac output (circulatory hypoxia), inadequate oxygen transport (hypoxemia: decreased oxygen saturation or transport capacity), or inadequate capillary perfusion (microhemodynamic hypoxia).Furthermore, central sleep apnea provokes transient decrease in blood saturation and triggers the expression of hypoxia-regulated factors, 6 and S/V was shown to be beneficial in this context. 7,8Cells modulate their transcriptional program to respond to hypoxia.The transcriptional response to hypoxia is pleiotropic and modulates the expression of genes involved in metabolism, extracellular matrix remodeling, and cell signaling amongst other processes. 9In humans, hypoxia is associated with an increase in renal erythropoietin (EPO) production to regulate erythropoiesis, liver ceruloplasmin production to regulate copper and iron metabolism, 10 and vascular endothelial growth factor (VEGF) and placental growth factor (PlGF) that are involved in angiogenesis.The response to hypoxia is predominantly transcriptional and is driven by two hypoxia-inducible transcriptional factors (HIF) with overlapping and complementary functions. 9HIF transcriptional factors are heterodimers composed of an alpha subunit (HIF1a or HIF2a) that acts as oxygen sensors and a common b-subunit (ARNT: Aryl hydrocarbon Receptor Nuclear Translocator).All subunits are constitutively expressed.HIF1a and HIF2a are hydroxylated and rapidly targeted to the proteasome under normoxic conditions. 11,12Under hypoxic conditions, HIF1a and HIF2a are less hydroxylated, hence less degraded, bind to ARNT and regulate the transcription of their target genes.In addition, the acetylation of HIF1a Lysine 709 (K709ac) increases the expression of HIF1a targets, 13 and cellular hypoxia increases intracellular lactate production which leads to histone lactylation, 14 which also contributes to changes in gene expression.
ARNT can also bind to the Aryl hydrocarbon Receptor (AhR), which is mainly involved in the metabolism of xenobiotic compounds. 15The sharing of ARNT by AhR and HIFa subunit leads to a competition between the two pathways.Amongst potent AhR activators are indoxyl sulphate, para-cresol (p-cresol), and para-cresyl sulphate (p-cresyl-sulphate). 16 Indole and p-cresol are produced by the gut microbiota ll OPEN ACCESS from tryptophan and tyrosine, respectively.They are eliminated by the kidney after sulphoconjugation as indoxyl sulphate and p-cresyl-sulphate; the sulphoconjugation of indole requires a preceding hydroxylation by cytochrome CYP2E1. 17,18In chronic kidney disease, indoxyl sulphate, p-cresol and p-cresyl-sulphate tend to accumulate in bodily fluids resulting in an exacerbated AhR response.The enhanced AhR response outcompete that of hypoxia and AhR activation by indoxyl sulphate reduced the production of EPO. 19Therefore, the response to hypoxia is no longer adapted to the level of hypoxia, contributing to the worsening of kidney disease.
Given the longitudinal nature of the cohort used 2 and the rapid changes observed in proBNP glycosylation at T71 in this cohort, it offered an opportunity to evaluate the possible role of hypoxia as a mechanism underlying the variation of proBNP T71 glycosylation observed in patients treated with S/V.The results obtained prompted us to evaluate whether the improvement in oxygen bioavailability suggested in the heart also applied to other organs.

RESULTS
proBNP glycosylation at T71 is associated with miR-30a in vivo While proBNP T71 glycosylation increased at D30 and D90 after the introduction of S/V 2 , plasma levels of miR-30a decreased (Figure 1A).This decrease in miR-30a plasma levels was strongly inversely correlated with proBNP T71 glycosylation (Figure 1B).T71 proBNP glycosylation was already increased at D7 (Figure S1A), and there was a trend toward a decrease in miR-30a (p = 0.056, Figure S1B).At D14, T71 was markedly increased and miR-30a was markedly decreased when compared to baseline (Figures S1C and S1D), and the strong negative correlation between proBNP T71 glycosylation and miR-30a was also present at D7 and D14 (Figure S1E).

HIF-dependent biomarkers of hypoxia
The increase in proBNP T71 glycosylation and the decrease in miR-30a suggested a decrease in cardiac hypoxia.To evaluate a possible decrease in hypoxia in other tissue and organs, we measured organ-specific (EPO and ceruloplasmin) and systemic (VEGF-A, PlGF, and K709ac HIF1a) biomarkers that are upregulated under hypoxic conditions.At baseline, 94% of patients had EPO >26 mIU/mL (Figure 2A), and 30% had ceruloplasmin concentration >50 mg/dL (Figure 2B).Both EPO and ceruloplasmin concentrations decreased after the introduction of S/V at D30 and D90, although 70% of patients had EPO levels >26 mIU/mL at D90.The ubiquitous biomarkers of hypoxia -i.e., VEGF-A (Figure 2C), PlGF (Figure 2D), and K709ac HIF1a (Figure 2E), markedly decreased at D30 and D90.Of note, EPO, ceruloplasmin, and PlGF were already decreased at D7, while VEGF-A only decreased at D14. K709ac HIF1a did not decrease at D7 and D14 (Figure S2).Although the biomarkers tested can be transcriptionally regulated by HIF1a, there were only moderate correlations between EPO, ceruloplasmin, VEGF-A, and PlGF concentrations and K709ac HIF1a (Figure S3).The decrease in these hypoxia-related biomarkers was not associated with any change in either hemoglobin, hematocrit, or red blood cell count (Figure S4), indicating that the decrease in the various hypoxia-related biomarkers tested did not result from an increase in oxygen transport capacity.Finally, none of the patients exhibited oxygen saturation below 98% at any time point neither did they exhibit Cheyne-Stokes respirations as a symptom of central apnea.

HIF-independent biomarkers of hypoxia
To further ascertain a decrease in hypoxia rather than an inadequate decrease in the response to hypoxia, we measured markers that are positively regulated by hypoxia in a HIF-independent fashion. 20We found that plasma levels of osteopontin decreased (Figure 3A) while the oxygen-dependent ferroxidase activity of ceruloplasmin (EC 1.16.3.1)markedly increased at D30 and D90 (Figure 3B).Finally, we (B) Relationship between plasma miR-30a plasma levels and proBNP T71 glycosylation.Variables in A were analyzed using repeated-measure ANOVA on logtransformed data followed by the paired Student's t test corrected for multiple comparisons (holm).Correlations between variables were estimated with Spearman rank correlation and expressed by the correlation coefficient (r).The number of subjects is indicated in each panel.
measured plasma levels of lactylated histone, which is a marker of intracellular lactate accumulation, in a limited subset of patients and found it was strongly positively correlated to plasma levels of K709ac HIF1a (Figure 3C).Another way to evaluate a genuine decrease in hypoxia is to compare the production of p-cresyl sulphate and indoxyl sulphate, the latter requiring molecular oxygen to be produced in hepatocytes (Figure S5).The plasma levels of p-cresol (Figure 4A) and its sulphoconjugate p-cresyl sulphate (Figure 4B) remained unchanged throughout the study.In marked contrast, the plasma levels of indoxyl sulphate increased by 2.9 [0.6-10.8]folds between baseline and D30, and by 3.0 [1.1-12.3]folds between D30 and D90 (Figure 4C), while its precursor, tryptophan, remained unchanged throughout the study (Figure 4D).The increase in indoxyl sulphate was not associated with the worsening of kidney function as eGFR remained unchanged between baseline and D30 and marginally declined between D30 and D90 (Figure 4E).Finally, there was a strong negative correlation between the plasma levels of indoxyl sulphate and K709 HIF1a (Figure 4F).(C) Relationship between HIF1a acetylated on lysine 709 (HIF1a K709ac) and lactylated histones in a subset of patients.Variables in A and B were analyzed using repeated-measure ANOVA on log-transformed data followed by the paired Student's t test corrected for multiple comparisons (holm).Correlations between variables were estimated with Spearman rank correlation and expressed by the correlation coefficient (r).The number of subjects is indicated in each panel.

Specificity of biomarker changes
To test the specificity of S/V over the up-titration of Valsartan, we first estimated a Valsartan dose-equivalence for all patients treated with ARB and ACEi at baseline (10 mg Ramipril daily = 320 mg Valsartan daily), and for patients treated with S/V at D30 (97 mg/103 mg S/V bid = 320 mg Valsartan daily). 21Patients were split into two groups based on whether their Valsartan dose-equivalence was up-titrated between baseline and D30 or not.The variations of all the biomarkers tested were similar between the two groups, suggesting that most of the effects observed are specific to S/V rather than the up-titration of Valsartan alone (Table 1).

Biomarker validation in a hypoxic rat model
To fully validate the relationship between the biomarkers tested and hypoxia, we subjected wild-type male rats to isobaric hypoxia (15% oxygen), and measured those biomarkers before, after a week under isobaric hypoxic atmosphere, and after a week back under normal atmosphere.As expected, K709ac HIF1a increased with hypoxia (Figure 5A), which was accompanied by an increase in miR-30a (Figure 5B), EPO (Figure S6A), ceruloplasmin (Figure S6B), and VEGF-A (Figure S6C) plasma concentrations.HIF1a-independent markers lactylated histone (Figure 5C) and osteopontin (Figure S6D) also increased, while ceruloplasmin activity decreased (Figure S6E).Indoxyl sulphate decreased upon hypoxia (Figure 5D), unlike its precursor tryptophan (Figure S6F) and p-cresol (Figure S6G) and p-cresyl sulphate (Figure S6H) which remained unchanged throughout.After placing the rats back in a normal atmosphere for a week, all the markers trended toward baseline levels, although not reaching them.As observed in the cohort, there were strong positive correlations between K709ac HIF1a and lactylated histones (Figure 5E) or indoxyl sulphate (Figure 5F) plasma concentrations.

DISCUSSION
In this study, we show that HFrEF patients under optimal medication exhibit a moderate subclinical cellular hypoxia in numerous tissues and organs that improved after the introduction of S/V.Importantly, the improvement in hypoxia preceded clinical improvement as measured by the NYHA classification.Furthermore, our data support proBNP T71 glycosylation as a biomarker of cardiac hypoxia.
We found that most of the HFrEF patients included in this study had subclinical hypoxia affecting multiple tissues or organs, even under optimal treatment with HF therapies.The introduction of S/V led to consistent improvements in all biomarkers of hypoxia by day 30, which (F) Relationship between HIF1a acetylated on lysine 709 (HIF1a K709ac) and indoxyl sulphate.Variables in A-E were analyzed using repeated-measure ANOVA on log-transformed data followed by the paired Student's t test corrected for multiple comparisons (holm).Correlations between variables were estimated with Spearman rank correlation and expressed by the correlation coefficient (r).The number of subjects is indicated in each panel.
appears to be specific of the combination of sacubitril and valsartan.However, S/V also markedly increased indoxyl sulphate plasma levels, which could inappropriately blunt the response to hypoxia via AhR activation without any change in oxygenation.However, several lines of evidence suggest that this scenario is unlikely: (1) the increase in indoxyl sulphate likely reflects an increase in oxygen-dependent hydroxylation of indole 17,18 as indoxyl sulphate also increased after placing rats back in normoxic conditions after a week under hypoxic conditions; (2) indoxyl sulphate unlikely accumulated due to reduced renal clearance as p-cresyl sulphate, which is cleared by the kidney $3 times less effectively by the kidney than indoxyl sulphate, 22 remained unchanged throughout the study despite a marginal decline in kidney function at D90  .Evolution of hypoxia biomarkers in rats subjected to isobaric hypoxia Plasma levels of HIF1a acetylated on lysine 709 (HIF1a K709ac, (A), miR-30a (B), lactylated histones (C) in rats under normoxic conditions (21% O 2 ), after one week in isobaric hypoxia (15% O 2 ), and one week after returning to normoxia (21% O 2 ).(E) Relationship between HIF1a K709ac and lactylated histones.(D) Plasma level of indoxyl sulphate in rats under normoxic conditions (21% O2), after one week in isobaric hypoxia (15% O2), and one week after returning to normoxia (21% O2).(F) Relationship between HIF1a K709ac and indoxyl sulphate.Variables in A-C and E were analyzed using repeated-measure ANOVA on log-transformed data followed by the paired Student's t test corrected for multiple comparisons (holm).Correlations between variables were estimated with Spearman rank correlation and expressed by the correlation coefficient (r).There were 6 rats in the experiment that were followed longitudinally throughout the study.
only; (3) several HIF-independent yet oxygen-dependent processes (histone lactylation, ceruloplasmin oxidase activity) improved both in humans and rats; and (4) the correlation between lactylated histone and K709ac HIF1a indicate that the decrease in the latter likely results from increased cellular oxygenation.In light of these results, the strong negative correlation between indoxyl sulphate and K709ac HIF1a likely reflects their inverse dependency on oxygen rather than a competition between the two processes, and may also reflect the contribution of AhR in the proper down-regulation of the response to hypoxia when oxygenation has improved. 23Also, the indoxyl sulphate/p-cresyl sulphate ratio appears as a biomarker of intracellular oxygen availability in hepatocytes when their clearance is not reduced.Of note, some hypoxia-related biomarkers improved at day 7 and day 14 after the introduction of S/V -i.e., before the improvement of symptoms, suggesting that the improvement of tissue perfusion is rapid.However, there was only a trend toward a decrease in plasma mir-30a levels while proBNP T71 glycosylation already decreased.This might reflect the fact that proBNP is secreted while miRNAs are released via exosomes, therefore not necessarily directly reflecting the short-term dynamic of miRNA change inside the cell.The fact that all the biomarkers tested continued to improve at day 90 only reflects the slow shutdown of the response to hypoxia, as observed in rats.Taken together, these results strongly suggest that changes in biomarkers resulted from a true improvement in hypoxia and that S/V rapidly improves tissue perfusion.
Several non-exclusive mechanisms could account for the improvement in oxygen availability observed.Results from the PARADIGM-HF, 1 PROVE-HF 24 and EVALUATE 25 trials suggested that cardiac function and central hemodynamic improved, although EVALUATE did not find any impact of S/V on aortic stiffness. 25Alternatively, S/V may improve oxygen availability independently from any hemodynamic effect as it has been observed for the attenuation of renal injury in a rat model of obesity, 26 or by reducing central apnea. 7,8Finally, it is unlikely that the improvement of hypoxia results from an increase in oxygen transport capacity as neither hemoglobin, hematocrit nor red blood cell count changed throughout the study.Furthermore, patients exhibited normal oxygen saturation and did not show evidence of central apnea, although we did not evaluate apnea during sleep.Therefore, the mechanism whereby S/V reduced tissue hypoxia remains to be characterized but is likely to involve macro-and micro-hemodynamic parameters, as well as possible hemodynamic-unrelated effects.This improvement of hypoxia could also explain changes observed after the introduction of S/V in other biomarkers such as troponin and GLP-1. 2,27In particular, long-term exposure of intestinal enteroendocrine L cells to indole limits the secretion of GLP-1 28 and the anticipated decrease in indole via its increase in sulphoconjugation after S/V initiation is expected to increase plasma GLP-1 in synergy with its direct protection from NEP degradation by sacubitril. 2 As observed for ANP, 29 the effect of S/V on GLP-1 appears to be direct and indirect, and could participate in the metabolic benefit of S/V in HFrEF. 2,30The impact of S/V on hypoxia also echoes the benefit observed in a rat model of pulmonary arterial hypertension, 31 altogether being encouraging results for the undergoing trials testing S/V in this disease.Taken together, these data strongly suggest that the improvement of tissue perfusion participates in the mode of action of S/V.
Finally, we confirmed in vivo the negative relationship between miR-30a and proBNP T71 glycosylation, strongly suggesting that proBNP T71 glycosylation is a biomarker of ventricular cardiomyocyte hypoxia. 4Importantly, patients with acute dyspnea of cardiac (ADHF) and noncardiac (non-ADHF) origins had similar levels of proBNP T71 glycosylation, albeit radically different natriuretic peptide levels. 32In non-ADHF patients, lower proBNP T71 glycosylation levels were achieved at lower natriuretic peptide plasma levels than in ADHF patients, 32 suggesting that ADHF and non-ADHF patients suffer from similar levels of cardiac hypoxia, and that in that instance, natriuretic peptide levels do not reflect the level of hypoxia the heart is subjected to.Taken together, these data strongly suggest that proBNP glycosylation at T71 is a biomarker of cardiac hypoxia in HFrEF and acute conditions.
In conclusion, HFrEF exhibit subclinical tissue/organ hypoxia that is improved by S/V, and which may account for S/V clinical benefit.The cause of this improvement in tissue/organ hypoxia is unknown and could rely on hemodynamic-and non-hemodynamic-related mechanisms that are yet to be identified.More generally, subclinical hypoxia has long been overlooked and could be clinically relevant in the management of chronically and acutely ill patients, although further studies are warranted.

Limitations of the study
Our study has some limitations.First, the study population is rather small and not randomized which may limit the statistical performance.However, the changes observed in the various parameters tested are consistent across the population.Second, S/V up-titration was left at the discretion of the cardiologist according to the tolerance and patient clinical status, and therefore it is not possible to precisely assess the relationship between dose and time.Third, we did not measure any direct parameters of tissue oxygenation nor did we study respiratory patterns and sleep apnea.However, the methods available can only measure tissue oxygen in the skin non-invasively and skeletal muscle oxygenation invasively.Finally, the open design and the lack of a control arm may have affected the assessment of the clinical response even though the study is longitudinal and each patient is under his/her control.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:

Study population
This study was performed according to the current revision of the Helsinki Declaration, approved by the Institutional Review Board (IRB 00006477; agreement: 16-035), and registered at clinicaltrials.govunder the NCT01374880 identifier.The study population was previously described 2,29 and its main characteristics are reported in Table S1.Briefly, it consisted in 73 patients with HFrEF who were transitioned from an Angiotensin-converting enzyme inhibitor/Angiotensin receptor blocker to sacubitril/valsartan as per ESC guidelines. 33Blood samples were collected before the first administration of S/V (D0), at one (D30) and three months (D90) after the initiation of S/V.For some patients, blood samples were also collected one (D7) and two (D14) weeks after the initiation of S/V.Of note, the number of measurements may vary due to plasma exhaustion.

Plasma sampling
Venous blood samples were collected in tubes containing EDTA.Blood samples were immediately centrifuged at 3,500 rpm for 15 min at 4 C, frozen and stored at À80 C until further use.

Plasma enzymatic activity
Ceruloplasmin oxidase activity was measured, as previously described 36 : briefly, plasma samples containing 2 mg protein were added in duplicate to plastic cuvettes containing 0.6 mL of 0.1M acetate buffer (pH 6.0), 0.3 mL 0.25% p-phenylenediamine (PPD) in 0.1M acetate buffer, which had been equilibrated at 37 C for 5 min; blank samples containing 2 mg plasma protein, 0.3 mL 0.1M acetate buffer, 0.3 mL 0.1% NaN 3 in acetate buffer, and 0.3 mL 0.25% PPD in acetate buffer were prepared for each test sample; the reaction mixtures were incubated at 37 C, and the absorbance read at 530 nm after 10 and 40 min; ceruloplasmin oxidase activity was calculated in terms of oxidase units, where one oxidase unit = (Abs40min -Abs10min) 3 1000.

Quantification of ProBNP glycosylation by mass spectrometry
ProBNP glycosylation at threonine 71 32 and histone lactylation 14  ).The digests were then desalted using SepPak-C18 SPE cartridges, dried and resuspended in 80% Acetonitrile, 0.1% TFA.Peptides were separated using a Dionex Ultimate 3000 HPLC system equipped with a HILIC (hydrophilic interaction liquid chromatography) column, using a similar protocol to the hSAX method described previously. 37The peptide fractions were desalted using SepPak-C18 SPE plates and then resuspended in 5% formic acid for LC-MS/MS analysis.For Trypsin+Lys-C double digests, peptide chromatography was performed using a Dionex RSLCnano HPLC.Peptides were injected directly onto a 75 mm 3 50 cm PepMap-C18 column using the following mobile phases: 2% acetonitrile +0.1% formic acid (Solvent A) and 80% acetonitrile +0.1% formic acid (Solvent B).The linear gradient began with 5% A to 35% B over 220 min with a constant flow rate of 200 nL/min.The peptide eluent flowed into a nanoelectrospray emitter at the front end of a Q-Exactive (quadrupole Orbitrap) mass spectrometer (Thermo Fisher).A typical ''Top10'' acquisition method was used and an in-house developed software inspired by 38 was used to evaluate peptide identifications and abundance ratios.
miRNA quantification by qPCR miR-30a was measured by qPCR.Total RNA was extracted from plasma samples using the mirVana PARIS kit (Ambion, Thermo Fisher Scientific, St-Quentin en Yvelines, France).Hemolyzed samples were avoided since hemolysis was reported to alter miR plasma content. 39Spikedin synthetic C. elegans miRNA controls (cel-miR39, 54 and 238; Qiagen, Courtaboeuf, France) were added to samples for correction of extraction efficiency.After DNAse treatment, RNAs were reverse transcribed with the miScript reverse transcription kit (Qiagen).cDNA was diluted 10-fold before quantitative PCR with the miScript SYBR Green PCR kit (Qiagen).A control without reverse transcriptase and a control without RNA was added to each PCR plate to ensure the absence of contaminating DNA and to check for non-specific amplification, respectively.Primers and probes for miR-30a were purchased from Applied Biosystems.Expression values were normalized using the mean Ct of the spiked-in controls and calculated with the 2 ÀDCt formula.

Quantification of metabolites by HPLC and mass spectrometry
Plasma tryptophan was measured by HPLC with coulometric detection as previously described. 40p-cresyl sulfate, and indoxyl sulfate were measured using LC-MS/MS with stable isotopic dilution as previously described. 41The determination of unconjugated p-cresol by LC-MS/MS sample preparation included derivatization with dansyl chloride according to. 42

QUANTIFICATION AND STATISTICAL ANALYSIS
Statistical analyses were performed using R version 4.0.3. 43Normality was assessed using the Shapiro-Wilk test and continuous variables are expressed as median and interquartile range.Longitudinal data were analyzed using repeated measures ANOVA on log-transformed values using a linear mixed model generated with the lme function of the nlme package in R. 44 Post-hoc analysis was performed using the paired Student's t test corrected for multiple comparisons (holm).Comparisons between baseline and D7 or D14 were performed using the Wilcoxon signed-rank tests.Relationships between continuous variables were assessed using Spearman rank correlation (r).A p-value < 0.05 was considered statistically significant.

ADDITIONAL RESOURCES
The human study was registered at clinicaltrials.govunder the NCT01374880 identifier.

Figure 1 .
Figure1.proBNP glycosylation at T71 and miR-30a in vivo (A) Plasma levels of miR-30a at baseline (D0), and D30 and D90 after the introduction of S/V.(B) Relationship between plasma miR-30a plasma levels and proBNP T71 glycosylation.Variables in A were analyzed using repeated-measure ANOVA on logtransformed data followed by the paired Student's t test corrected for multiple comparisons (holm).Correlations between variables were estimated with Spearman rank correlation and expressed by the correlation coefficient (r).The number of subjects is indicated in each panel.

Figure 2 .Figure 3 .
Figure 2. Evolution of HIF1a-dependent biomarkersPlasma concentrations of erythropoietin (EPO, A), ceruloplasmin (B), vascular endothelial growth factor A (VEGF-A, C), placental growth factor (PlGF, D), and HIF1a acetylated on lysine 709 (HIF1a K709ac, E) at baseline (D0), and D30 and D90 after the introduction of S/V.Variables were analyzed using repeatedmeasure ANOVA on log-transformed data followed by the paired Student's t test corrected for multiple comparisons (holm).The number of subjects is indicated in each panel.The dotted line in A and B indicate the upper limit of the normal range.

FFigure 4 .
Figure 4. Evolution of para-cresol and indoxyl sulphatePlasma levels of para-cresol (p-cresol, A), and its sulphoconjugate para-cresyl sulphate (B) at baseline (D0), and D30 and D90 after the introduction of S/V.Plasma levels of indoxyl sulphate (C) and its precursor tryptophan (D) at baseline (D0), and D30 and D90 after the introduction of S/V.(E) Estimated glomerular filtration rate (eGFR) calculated according to the 2021 revision of the CKD-EPI formula at baseline (D0), and D30 and D90 after the introduction of S/V.(F) Relationship between HIF1a acetylated on lysine 709 (HIF1a K709ac) and indoxyl sulphate.Variables in A-E were analyzed using repeated-measure ANOVA on log-transformed data followed by the paired Student's t test corrected for multiple comparisons (holm).Correlations between variables were estimated with Spearman rank correlation and expressed by the correlation coefficient (r).The number of subjects is indicated in each panel.

FFigure 5
Figure 5. Evolution of hypoxia biomarkers in rats subjected to isobaric hypoxia Plasma levels of HIF1a acetylated on lysine 709 (HIF1a K709ac, (A), miR-30a (B), lactylated histones (C) in rats under normoxic conditions (21% O 2 ), after one week in isobaric hypoxia (15% O 2 ), and one week after returning to normoxia (21% O 2 ).(E) Relationship between HIF1a K709ac and lactylated histones.(D) Plasma level of indoxyl sulphate in rats under normoxic conditions (21% O2), after one week in isobaric hypoxia (15% O2), and one week after returning to normoxia (21% O2).(F) Relationship between HIF1a K709ac and indoxyl sulphate.Variables in A-C and E were analyzed using repeated-measure ANOVA on log-transformed data followed by the paired Student's t test corrected for multiple comparisons (holm).Correlations between variables were estimated with Spearman rank correlation and expressed by the correlation coefficient (r).There were 6 rats in the experiment that were followed longitudinally throughout the study.

Table 1 .
Biomarker relative variations (%) in patients whose Valsartan dose equivalence was not up-titrated or up-titrated between baseline and D30 Data are represented as median and interquartile range.The number of patients for each comparison is indicated between brackets.K709ac HIF1a: HIF1a acetylated on lysine 709, EPO: erythropoietin, VEGF-A: vascular endothelial growth factor A, PlGF: placental growth factor.

TABLE
d RESOURCE AVAILABILITY B Lead contact B Materials availability B Data and code availability d EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS were measured by mass spectrometry (LC-MS/MS) as previously described.HIF-1a K709ac was measured by MS/MS.Proteins were reduced with 25 mM Tris(2-carboxyethyl)phosphine hydrochloride (TCEP) in denaturing urea buffer, alkylated with iodoacetamide, and diluted with digest buffer (100 mM Tris pH 8.0 + 1 mM CaCl 2 ) to reach 4 M urea, before being digested with 1:50 Lys-C (Custom Biotech, ref. 11058533103).The digestion was diluted with digest buffer to reach 0.8 M urea and digested with trypsin (Promega; 1:50