5248

Abstract. – OBJECTIVE: Diabetic nephropa-thy (DN) is considered as a complication of dia-betes and accounts for about 40%. Reactive ox-ygen species (ROS) level continues to increase in DN. However, whether specific ROS levels can alleviate renal damage by improving mitochon-drial function has not been investigated.

MATERIALS AND METHODS: DN model was established by intraperitoneal injection of strep-tozotocin (STZ). The rats were divided into nor-mal group, STZ model group, and N-acetylcys-teine (NAC) group. Fasting blood glucose was tested to assess the modeling. The renal injury was evaluated by using hematoxylin-eosin (HE) staining and periodate-Schiff staining. Serum creatinine and 24 h urinary protein levels were determined by renal function detection kit. The levels of ROS and malondialdehyde (MDA) were assessed by the kit to evaluate the effects of ox-idative stress and NAC in rats. The mitochondri-al damage marker Cyto C level was detected by Western blot.

RESULTS: Blood glucose, serum creatinine, and urinary protein levels were significantly in-creased in model group compared with the nor-mal group (p<0.05). Blood glucose levels, serum creatinine, and urinary protein levels were mark-edly improved after the ROS and MDA levels re-duced by NAC. Meanwhile, glomerular hypertro-phy, mesangial matrix accumulation, and severe renal injury were observed in the model group. NAC administration markedly improved glomer-ular morphology and mesangial matrix aggre-gation. Cyto C expression in the model group increased significantly compared with normal control group (p<0.05), while NAC effectively in-hibited Cyto C levels.

CONCLUSIONS: Oxidative stress-mediated mitochondrial damage is an important part of the pathogenesis of DN. Inhibition of ROS pro-duction can be a potential target for DN treat-ment.

Key Words:Oxidative stress, Mitochondrial damage, Diabetic

nephropathy.

Introduction

Diabetic nephropathy (DN) is the most common microvascular complications of diabetes that af-fects 20%-40% of type 2 diabetes mellitus (T2DM) patients and may lead to end-stage renal disease1. It seriously affects the morbidity and mortality of diabetes2. The molecular pathological mechanisms leading to DN involve hyperglycemic-induced me-tabolism, hemodynamic and complex interactions between inflammation3,4. These factors alter the function and morphology of the inner walls of blood vessels and interact with neighboring cells, leading to mitochondrial dysfunction and renal en-dothelial dysfunction, thus playing a crucial role in the development of DN5.

Oxidative stress refers to the overproduction of high activity molecules such as reactive oxygen species (ROS) in the body by a variety of harmful stimuli, leading to the oxidation degree excesses oxide removal and the imbalance between oxida-tion and antioxidant system6. The main characte-ristic of cellular oxidative stress is the continuo-us up-regulation of ROS levels. In recent years, it was found that oxidative stress is an important factor to promote the development of DN7,8.

Mitochondria are the main place of cellular oxidative breath. In addition to supplying energy, participating in energy metabolism, and maintai-ning normal life activities of cells, mitochondria are also involved in processes such as cell dif-ferentiation, cell information transmission and apoptosis, and have the functions of regulating cell growth and cell cycle9. Therefore, oxidative stress level is closely related to mitochondrial fun-ction. Cytochrome (Cyto) C locates in the space between outer membrane and endometrium of mi-tochondria. It is an important component of mi-tochondrial oxidative respiratory chain and plays a crucial role in apoptosis. Apoptosis inducer can trigger Cyto C release into the cytoplasm from

European Review for Medical and Pharmacological Sciences 2018; 22: 5248-5254

X. CHEN, M. FANG

Department of Endocrinology, Yiwu Central Hospital, Yiwu, Zhejiang, China

Corresponding Author: Xia Chen, MD; e-mail: zudc8rjmfc71@sina.com

Oxidative stress mediated mitochondrial damage plays roles in pathogenesis of diabetic nephropathy rat

Oxidative stress in diabetic nephropathy

5249

the mitochondria and bind to Apaf-1. The Cyto C/Apaf-1 complex activates caspase-9, which in turn activates caspase-3 and other downstream caspases. The release of Cyto C occurs before the activation of caspases and DNA breaks, thus is a hallmark of apoptosis10. Therefore, we evaluated mitochondrial damage by detecting cytochrome C expression in this study.

Glutathione (GSH) is an important anti-oxidant in the body. N-acetylcysteine (NAC), a precursor of GSH, can directly and specifically inactivate ROS in vivo to inhibit oxidative stress. Up to now, there is still lack of report about whether specific inhibition of ROS levels can mitigate renal dama-ge by improving mitochondrial function.

This study established a rat DN model and adopted ROS inhibitor NAC to explore the role of oxidative stress and its induced mitochondrial damage in DN.

Materials and Methods

Main Materials and Reagentsβ-actin antibody was purchased from Kan-

gcheng Biological Company (Shanghai, China). Col IV and fibronectin (FN) antibodies were pur-chased from Abcam Biotechnology (Cambridge, MA, USA). Rabbit Anti-Mouse IgG (H + L) and Rabbit Anti-Mouse IgG (H + L) were purchased from Proteintech Co., Ltd. (Wuhan, China). Rat reactive oxygen species (ROS) enzyme-linked immunosorbent assay (ELISA) kit and MDA de-tection kit were purchased from AMAX (Wuhan, China). HE staining kit and periodate-Schiff kit were purchased from ZSBio Tech. (Beijing, Chi-na). NAC was purchased from Sigma-Aldrich (St. Louis, MO, USA). Other reagents were purchased from Sigma-Aldrich (St. Louis, MO, USA).

Main InstrumentsGel imaging system UVP Multispectral Ima-

ging System (Sacramento, CA, USA). PS-9 se-mi-dry transfer electrophoresis instrument was purchased from Jinmix Ltd. (Dalian, China). Thermo-354 microplate reader was purchased from Thermo Fisher Scientific Company (New York, NY, USA).

Experimental AnimalsFemale Wistar rats aged 6-8 weeks old

and weighed 180-220 g were purchased from Zhejiang University Experimental Animal Cen-ter (Zhejiang, China). The rats were raised in

a clean animal room with temperature at 24°C and the relative humidity at 60%. The day/ni-ght cycle was 12 h. The rats were free to eat and drink, and the padding was changed daily to avoid infection.

Rats were used for all experiments, and all procedures were approved by the Animal Ethics Committee of Yiwu Central Hospital (Zhejiang, China).

Modeling and AdministrationThe rats were divided into normal group, mo-

del group, and NAC group. After adaptive bree-ding, the rat was intraperitoneally injected with STZ (45 mg/kg) dissolved in freshly prepared citrate buffer. In the NAC group, NAC (5 mg/kg) was injected through the tail vein at 2 h befo-re STZ injection to inhibit the increase of ROS. Two weeks later, the rats were sacrificed and the model was confirmed by HE staining and blood glucose level detection.

Hematoxylin-Eosin (HE) StainingAfter the rats were sacrificed, the kidneys were

fixed with 4% paraformaldehyde for 24 h and then dehydrated to in 30% sucrose concentration at 4°C. After optimal cutting temperature com-pound (OCT) embedding, they were frozen in a microtome for continuous freezing. The frozen sections were dried and subjected to conventional HE staining. After gradient ethanol dehydration, the sections were hyalinized in xylene. At last, the sections were observed under an optical micro-scope and photographed.

Periodate-Schiff StainingAfter conventional frozen, the section was tre-

ated with 0.5% periodic acid oxidation for 10 min and washed by distilled water. Then, the section was incubated in Schiff reagent at 37°C for 30 min and redyed in Mayer hematoxylin for 1 min. After gradient ethanol dehydration, the section was sealed by neutral balsam.

Fasting Blood Glucose and Serum Creatinine Detection

The rat was fasted for one day at 36 h after mo-deling. The blood sample was taken from the tail vein and blood glucose was measured by gluco-meter. The whole blood was centrifuged to obtain serum. Serum creatinine was detected on automa-tic biochemical analyzer by basic picric acid colo-rimetric method according to the manual.

X. Chen, M. Fang

5250

24 h Urine Protein DetectionFung’s metabolic cage was used to collect 24

h urine on the day before scarification. The 24 h urine protein was measured on biochemical auto-matic biochemical analyzer.

Total Protein ExtractionA total of 100 mg rat ovary samples was stored

at -80°C. The sample was lysed by 1 ml radio-immunoprecipitation assay (RIPA) potent lysate (Beyotime Biotechnology, Shanghai, China) for 15 min and centrifuged at 12000 ×g and 4°C for 10 min. The supernatant was tested by bicincho-ninic acid (BCA) method to determine protein concentration. At last, the sample was boiled for 5 min to denature the protein.

Western BlotTotal protein was separated by 10% sodium do-

decyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene difluoride (PVDF) membrane at 300 mA for 1 h. Next, the membrane was incubated in primary an-tibody (ColIV, FN, cyto C, 1:1000) at 4°C over-night. After washed by Tween/Tris-buffered salt solution (TTBS) solution, the membrane was in-cubated in secondary antibody (1:1000) at 37°C for 2 h and tested by enhanced chemiluminescen-ce (ECL).

ROS and MDA Levels DetectionROS level was determined by ELISA accor-

ding to the manual. Specially, the serum was ad-ded to the 96-well plate and incubated in HRP conjugated secondary antibody at 37°C for 1 h. After washed for 5 times, the plate was reacted with 3,3’,5,5’-tetramethylbenzidine (TMB) sub-strate avoid of light for 20 min. At last, the plate was read on microplate reader at 490 nm to obtain the absorbance value.

MDA was detected by the kit upon the principle of its reaction with thiobarbituric acid to produce

red color. The sample was measured at 530 nm to calculate the MDA concentration.

Statistical AnalysisAll data were analyzed on SPSS 18.0 softwa-

re (SPSS Inc., Released 2009. PASW Statistics for Windows, Chicago, IL, USA). The Student’s t-test was used to compare the differences betwe-en two groups. Tukey’s post-hoc test was used to validate the ANOVA for comparing measurement data between groups. p<0.05 was considered as statistical significance.

Results

HE StainingHE staining is the basic staining method to ob-

serve cell morphology changes [11]. Compared with the normal group, glomerular volume signi-ficantly increased in diabetic rat. NAC treatment led the renal enlargement unobvious. As shown in Figure 1, glomerular hypertrophy was evident in the diabetic model group, and NAC treatment effectively prevented such histological change.

Periodate-Schiff StainingPeriodate-Schiff staining is a method for de-

tecting polysaccharides, glycolipids, and mu-cins in glycogen and tissues [11]. Kidney tissue showed blue cell nucleus and bright purple cyto-plasm after periodic acid-Schiff staining. In the diabetic model group, mesangial matrix aggrega-tion was more evident, whereas NAC treatment markedly reduced the phenomenon (Figure 2).

Fasting Blood Glucose, Serum Creatinine, and 24 h Urine Protein Content Changes

Fasting blood glucose is an important in-dicator to assess the severity of diabetes12. As shown in Figure 3A, the fasting blood glucose

Figure 1. HE staining of diabetic rat kidney.

Oxidative stress in diabetic nephropathy

5251

level in the diabetic group was significantly in-creased compared with the normal control group (p<0.05). NAC intervention declined the blood glucose level.

24 h urine protein is the earliest and most common use of clinical indicator of DN, which is an independent factor of the cardiovascular risk in T2DM patients13,14. The 24 h urinary pro-tein results were shown in Figure 3B. Compared with the normal control group, the 24 h urinary protein level was significantly elevated in the diabetic model group (p<0.05). NAC treatment apparently decreased the level of 24 h urine pro-tein (p<0.05).

Creatinine is a small molecule substance that can pass through the glomerular filtration and is rarely absorbed in the renal tubules. The crea-tinine produced daily is almost always excreted from the urine and generally not affected by uri-ne volume. Clinical serum creatinine is one of the main ways to evaluate renal function. Serum creatinine elevation refers to impaired renal fun-ction15. It was found that serum creatinine level was markedly up-regulated (p<0.05) in diabetic model group and significantly decreased after NAC treatment (p<0.05) (Figure 3C).

ColIV and Fibronectin Protein Expressions in Renal Tissue

ColIV and FN in renal tissues were also used to detect the degree of renal injury. As shown in Figure 4, and the expressions of ColIV and FN in diabetic model group were significantly in-creased. NAC treatment markedly inhibited the expressions of ColIV and FN.

ROS and MDA Levels in Renal TissueOxidative stress is a key factor that leads to

mitochondrial damage. Oxidative stress presents as a significant increase of ROS and MDA levels. ROS and MDA levels were significantly increa-sed in the diabetic model group (p<0.05), while NAC markedly reduced ROS and MDA levels (p<0.05) (Figure 5).

Cytochrome C Expression in Renal TissueCyto C is a crucial electron mediator during

biooxidation. Cyto C is released from the mito-chondria into the cytoplasm during mitochondria damage, which can induce apoptosis through Ca-spase and other pathways. Cytoplasmic cyto C expression was markedly enhanced in the diabe-tic model renal tissue compared with the normal

Figure 2. Periodate-Schiff staining of diabetic rat kidney.

Figure 3. Fasting blood glucose, serum creatinine, and 24 h urine protein content changes. *p<0.05, compared with NC; #p<0.05, compared with DM.

X. Chen, M. Fang

5252

control group. Cytoplasmic Cyto C levels were apparently decreased after NAC administration (Figure 6).

Discussion

DN is the most important complication of dia-betes with upward trend in China. It has become the second cause of end-stage renal disease, only after glomerulonephritis. Because of its complex metabolic disorders, end-stage renal disease is often more difficult than the other kidney diseases in the treatment. Therefore, timely prevention and treatment is of great significance for the delay of diabetic nephropathy16. In this study, the role of oxidative stress-mediated mitochondrial damage in diabetic nephropathy was explored.

Oxidative stress refers to the degree of oxida-tion beyond removal when the body subjected to a variety of harmful stimuli, leading to oxidation and anti-oxidant system imbalance and tissue da-mage. In recent years, some theories connected high ROS production with various metabolic-re-lated chronic diseases, such as atherosclerosis and diabetes17. Previous studies suggested that oxidati-ve stress is an important cause of diabetic nephro-pathy, and some of the alleviating effects of tradi-tional Chinese medicine on diabetes are related to

the inhibition of oxidative stress18. However, the role of oxidative stress-mediated mitochondrial damage in diabetic rats has not been reported.

ROS is an important product of oxidative stress19. MDA is one of the key products of mem-brane lipid peroxidation, and its production can exacerbate membrane damage. Therefore, the content of MDA is a common indicator in phy-siological and physiological research of plant senescence, and can indirectly assess the degree of membrane damage20. In the present work, we found ROS and MDA elevation in the kidneys of diabetic rats, confirming the existence of hype-roxidative stress in the model. NAC, a precursor of GSH, is a thiol-containing antioxidant that exerts tissue protection in vivo through converting to GSH21. In early researches, NAC inhibited the expression of adhesion molecules on neutrophils and vascular endothelial cells, thus alleviating lung injury22. It also entered leukocytes and con-verted to physiological antioxidants to increase intracellular levels of reduction, leading to inacti-vation of ROS in cells and media. In our work, ROS and MDA levels were significantly inhibited by the oxidative stress inhibitor NAC, indicating that NAC markedly suppressed oxidative stress.

It is well-known that the increase of ROS sour-ce and the decrease of antioxidant capacity are the important factors leading to the enhancement of oxi-

Figure 4. ColIV and fibronectin protein expressions in renal tissue.

Figure 5. ROS and MDA levels in renal tissue. *p<0.05, compared with NC; #p<0.05, compared with DM.

Oxidative stress in diabetic nephropathy

5253

dative stress. It has been reported that an increasing production of ROS in DN patients is accompanied by diminished antioxidant capacity. The main sources of ROS production are oxidative respiratory chain and NADPH oxidase pathway, accompanied by the activation of multiple oxidative signaling pathways and looping between multiple signaling pathways. The oxidative signaling cascade results in oxida-tion damage continue to expand. Increased urinary protein level and decreased serum creatinine clea-rance are important features of diabetic renal injury. Diabetic nephrotic syndrome without treatment can gradually developed to chronic renal failure within four years23. In the present investigation, the admini-stration of NAC inhibited ROS level, urinary protein level, and serum creatinine level, and improved glo-merular morphology and function upon HE staining and periodate-Schiff staining.

Cyto C is an important protein of mitochondrial damage induced by oxidative stress. Elevated cyto C levels in cytoplasm are crucial markers of mitochon-drial damage. Maintaining mitochondrial integrity is necessary to maintain normal renal cell function. It was showed24,25 that mitochondria play a key role in acute kidney injury caused by surgery, chemothe-rapy, and shock. In this report, we used NAC to inhi-bit ROS levels, which significantly decreased the le-vel of cyto C and improved mitochondrial function.

The kidneys are one of the organs that is sensiti-ve to oxidative stress. In this study, the use of ROS inhibitor NAC confirmed that inhibition of oxidative stress in diabetes can reduce renal mitochondrial da-mage and thereby improve renal function. The inhi-bition of ROS production is an important target for the treatment of diabetic nephropathy.

Conclusions

Oxidative stress-mediated mitochondrial da-mage is a key part of the pathogenesis of DN. The suppression of oxidative stress alleviated mito-

chondrial damage and improved renal injury. The inhibition of the oxidative stress processes or sup-pression of the ROS production can be a potential target for the treatment of DN in clinical. There-fore, we recommend targeting oxidative stress pa-thway as an optimal anti-diabetic therapy.

Conflict of InterestThe Authors declare that they have no conflict of interest.

References

1) Yuan D, LiuXM, Fang Z, Du LL, Chang J, Lin Sh. Protective effect of reseratrol on kidney in rats with diabetic nephropathy and its effect on endo-plasmic reticulum stress. Eur Rev Med Pharmacol Sci 2018; 22: 1485-1493.

2) herMan Wh, PeterSen M. Response to comment on American Diabetes Association. Standards of Medical Care in Diabetes-2017. Diabetes Care 2017; 40: S1-S135.

3) Xiao L, Wang M, Yang S, Liu F, Sun L. A glimpse of the pathogenetic mechanisms of Wnt/beta-ca-tenin signaling in diabetic nephropathy. Biomed Res Int 2013; 2013: 987064.

4) KanWar YS, Sun L, Xie P, Liu FY, Chen S. A glimpse of various pathogenetic mechanisms of diabetic nephropathy. Ann Rev Pathol 2011; 6: 395-423.

5) Zhang Jh, Liu Xr, Sheng CX, Liu Ye. Analysis on difference in gastrointestinal hormoe levels of pa-tients with the history of diabetes and concurrent nephropathy and study on the role of liraglutide. Eur Rev Med Pharmacol Sci 2017; 21: 3523-3529.

6) KiM KM, KiM eY, Lee SY, Ko JJ, Lee Ka. Oocyte cyto-plasmic Gas6 and heparan sulfate (HS) are requi-red to establish the open chromatin state in nuclei during remodeling and reprogramming. Cell Phy-siol Biochem 2017; 45: 37-53.

7) Montero rM, herath a, QureShi a, eSFanDiari e, Pu-SeY CD, FranKeL ah, taM FWK. Defining phenotypes in diabetic nephropathy: a novel approach using a cross-sectional analysis of a single centre cohort. Sci Rep 2018; 8: 53.

8) Sagoo MK, gnuDi L. Diabetic nephropathy: is there a role for oxidative stress? Free Radic Biol Med 2018; 116: 50-63.

9) reXroth S, PoetSCh a, rogner M, haMann a, Wer-ner a, oSieWaCZ hD, SChaFer er, SeeLert h, DenCher na. Reactive oxygen species target specific tryp-tophan site in the mitochondrial ATP synthase. Biochim Biphys Acta 2012; 1817: 381-387.

10) Yang Z, Lv J, Lu X, Li X, an X, Wang J, Weng h, Li Y. Emulsified isoflurane induces release of cyto-chrome C in human neuroblastoma SHSY-5Y cells via JNK (c-Jun N-terminal kinases) signaling pathway. Neurotoxicol Teratol 2017; 65: 19-25.

Figure 6. Cytochrome C expression in renal tissue.

X. Chen, M. Fang

5254

11) van Der WeL MJ, DuitS LC, PouW re, SeLDenriJK Ca, oFFerhauS gJa, viSSer M, ten Kate FJ, BierMann K, BroSenS Laa, DouKaS M, huYSentruYt C, KarrenBeLD a, KatS-ugurLu g, van Der Laan JS, van LiJnSChoten g, MoLL FCP, ooMS a, van Der vaLK h, tiJSSen Jg, BergMan JJ, MeiJer SL. Improved diagnostic stratifi-cation of digitised Barrett’s oesophagus biopsies by TP53 immunohistochemical staining. Histopa-thology 2018; 72: 1015-1023.

12) CaSo F, DeL Puente a, oLiviero F, PeLuSo r, giroLiMetto F, BottigLieri P, FogLia F, Benigno C, taSSo M, PunZi L, SCarPa r, CoSta L. Metabolic syndrome in psoriatic arthritis: the interplay with cutaneous involvement. Evidences from literature and a recent cross-sectio-nal study. Clin Rheumatol 2018; 37: 579-586.

13) Bruno g, MerLetti F, Bargero g, noveLLi g, MeLiS D, SoDDu a, Perotto D, Pagano g, CavaLLo-Perin P. Estimated glomerular filtration rate, albuminuria and mortality in type 2 diabetes: the Casale Mon-ferrato study. Diabetologia 2007; 50: 941-948.

14) ninoMiYa t, PerKoviC v, De gaLan Be, ZoungaS S, PiLLai a, JarDine M, PateL a, CaSS a, neaL B, PouL-ter n, MogenSen Ce, CooPer M, Marre M, WiLLiaMS B, haMet P, ManCia g, WooDWarD M, MaCMahon S, ChaLMerS J; ADVANCE Collaborative Group. Albu-minuria and kidney function independently predict cardiovascular and renal outcomes in diabetes. J Am Soc Nephrol 2009; 20: 1813-1821.

15) YooWannaKuL S, harriS LS, DavenPort a. Peritoneal protein losses depend on more than just perito-neal dialysis modality and peritoneal membrane transporter status. Ther Apher Dial 2018; 22: 171-177.

16) tZioMaLoS K, athYroS vg. Diabetic nephropathy: new risk factors and improvements in diagnosis. Rev Diabet Stud 2015; 12: 110-118.

17) KiL YS, So YK, Choi MJ, han ar, Jin Ch, Seo eK. Cytoprotective dihydronaphthalenones from the wood of Catalpa ovata. Phytochemistry 2017; 147: 14-20.

18) nie t, Zhao S, Mao L, Yang Y, Sun W, Lin X, Liu S, Li K, Sun Y, Li P, Zhou Z, Lin S, hui X, Xu a, Ma CW, Xu Y, Wang C, DunBar Pr, Wu D. A natural compound formononetin derived from astragalus membranaceus increase adipocyte thermogene-sis by modulating PPARg activity. Br J Pharmacol 2018; 175: 1439-1450.

19) Xie h, Chen P, huang FW, Liu LP, Zhao F. Reactive oxygen specises: downregulate ARIDIA expres-sion via its promoter methylation during the pa-thogenesis of endometriosis. Eur Rev Med Phar-macol Sci 2017; 21: 4509-4515.

20) CarvaLho SB, Moreira aS, goMeS J, CarronDo MJt, thornton DJ, aLveS PM, CoSta J, PeiXoto C. A de-tection and quantification label-free tool to spe-ed up downstream processing of model mucins. PLoS One 2018; 13: e0190974.

21) LoWeth Ja, Singer BF, BaKer LK, WiLKe g, inaMine h, BuBuLa n, aLeXanDer JK, CarLeZon Wa, neve rL, veZina P. Transient overexpression of alpha-Ca2+/calmodulin-dependent protein kinase II in the nu-cleus accumbens shell enhances behavioral re-sponding to amphetamine. J Neurosci 2010; 30: 939-949.

22) rogerS LK, CiSMoWSKi MJ. Oxidative stress in the lung - The essential paradox. Curr Opin Toxicol 2018; 7: 37-43.

23) CZaJKa a, aJaZ S, gnuDi L, ParSaDe CK, JoneS P, reiD F, MaLiK an. Altered mitochondrial function, mito-chondrial DNA and reduced metabolic flexibility in patients with diabetic nephropathy. EBioMedicine 2015; 2: 499-512.

24) noBrega-Pereira S, FernanDeZ-MarCoS PJ. G6PD pro-tects from oxidative damage and improves health-span in mice. Nat Commun 2016; 7: 10894.

25) tran Mt, ZSengeLLer ZK, Berg ah, KhanKin ev, Bha-Sin MK, KiM W, CLiSh CB, StiLLMan ie, KaruManChi Sa, rhee eP, PariKh SM. PGC1alpha drives NAD biosynthesis linking oxidative metabolism to renal protection. Nature 2016; 531: 528-532.