Review Articles Open Access CC BY 4.0

Flavonoids in Sepsis: Mechanistic Modulation of Inflammatory Pathways and Therapeutic Potential-A Systematic Review of Preclinical Studies

Yunita Dewani , Gusbakti Rusip , Boyke Marthin Simbolon
First published: 31 May 2026 |https://doi.org/10.71197/jsocmed.v5i5.255
PDF HTML XML

Abstract

Introduction: Sepsis is a life-threatening syndrome driven by dysregulated host immunity, excessive inflammation, oxidative stress, endothelial injury, and immunosuppression. Flavonoids are bioactive polyphenols with anti-inflammatory and antioxidant effects; however, their therapeutic relevance in sepsis remains primarily preclinical.

Methods: A systematic review was conducted using PubMed/MEDLINE, Scopus, Web of Science, and Embase to identify controlled in vivo studies that evaluated the effects of flavonoids in experimental sepsis or endotoxemia. Eligible studies compared flavonoid-treated animals with septic controls and reported survival, organ injury, inflammatory, oxidative, and mechanistic outcomes. Evidence was qualitatively synthesized, with survival findings contextualized using relevant preclinical meta-analytic data.

Results: Eighty eligible studies were synthesized, predominantly rodent models using lipopolysaccharide-induced endotoxemia or cecal ligation and puncture. More than 30 flavonoids have been reported, including quercetin, kaempferol, luteolin, apigenin, fisetin, and orientin. Flavonoids reduce TNF-α, IL-6, IL-1β, oxidative stress, and organ injury in the pulmonary, renal, hepatic, and cardiovascular systems. Aggregated evidence suggests approximately 50% higher survival in flavonoid-treated animals. The mechanisms included NF-κB and MAPK inhibition, Nrf2/HO-1 activation, endothelial protection, and macrophage polarization. The limitations of this study include the prophylactic designs, heterogeneity, and limited clinical evidence.

Conclusion: Flavonoids exhibit consistent multi-target immunomodulatory and organ-protective effects in experimental sepsis. Translation requires standardized post-insult studies, improved bioavailability, pharmacokinetic evaluation, and early phase clinical trials.

Keywords: Sepsis, Flavonoids, inflammation, Oxidative Stress, NF-kB, Nrf2, MAPK, Macrophage Polarisation, Organ Injury

Downloads

Download data is not yet available.

INTRODUCTION

Sepsis is defined as life-threatening organ dysfunction caused by dysregulated host response to infection. It remains one of the leading causes of mortality in intensive care units worldwide, accounting for an estimated 11 million deaths annually, with hospital mortality exceeding 40% in septic shock [1-3]. Contemporary understanding recognises sepsis as a complex, time-dependent syndrome characterised not only by excessive inflammation but also by concurrent immunosuppression, endothelial dysfunction, mitochondrial injury, and maladaptive immunometabolic reprogramming [4,5]. This multidimensional pathobiology contributes to marked clinical heterogeneity and inconsistent therapeutic responses.

Despite advances in early recognition, antimicrobial therapy, source control, hemodynamic optimization, and organ support, no sepsis-specific pharmacological intervention has consistently demonstrated survival benefits in large phase III trials. Targeted strategies directed at single mediators, including anti-TNF-α, anti-IL-6, high-dose corticosteroids, and activated protein C, have produced disappointing or context-dependent results [6]. These limitations highlight the need for multi-target therapeutic approaches capable of modulating the interconnected inflammatory, oxidative, endothelial, and metabolic pathways. Flavonoids are structurally diverse plant-derived polyphenolic compounds characterized by a benzo-γ-pyrone core and classified into major subclasses such as flavonols, flavones, flavanones, flavanonols, isoflavones, and anthocyanins. Representative compounds include quercetin, kaempferol, fisetin, luteolin, apigenin, naringenin, pinocembrin, genistein, and orientin [7,8]. Extensive experimental evidence has demonstrated their anti-inflammatory, antioxidant, vasoprotective, endothelial-stabilizing, and immunomodulatory properties across inflammatory and infectious disease models. Mechanistically, flavonoids interact with key signalling pathways central to sepsis pathogenesis, including inhibition of nuclear factor kappa B (NF-κB) and mitogen-activated protein kinase (MAPK) signalling, activation of the nuclear factor erythroid 2-related factor 2/heme oxygenase-1 (Nrf2/HO-1) antioxidant axis, attenuation of inflammasome activation, and modulation of macrophage polarisation [9-12]. Several systematic reviews and meta-analyses have evaluated individual flavonoids or selected flavonoid subclasses in experimental sepsis, and pooled preclinical data suggest a favourable survival signal in treated animals [13-19]. However, existing syntheses remain fragmented, often focusing on single compounds, single mechanisms, or limited outcome domains. A comprehensive class-wide review integrating recent preclinical evidence, organ-specific protective effects, mechanistic convergence, and translational readiness is lacking.

Nevertheless, the translation of flavonoids from experimental sepsis models into clinical practice remains constrained by poor aqueous solubility, limited bioavailability, rapid metabolism, heterogeneous dosing regimens, and frequent reliance on prophylactic rather than therapeutic intervention designs. Furthermore, variability across sepsis models, animal species, timing of administration, formulation strategies, and outcome definitions complicates interpretation and limits the direct extrapolation to human sepsis. Therefore, this systematic review aimed to synthesize contemporary preclinical evidence on flavonoids in experimental sepsis, with an emphasis on in vivo models, survival outcomes, organ protection, and clinically relevant mechanistic pathways. This review also seeks to identify key translational barriers and propose strategic directions for future therapeutic development.

METHOD

This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines. Given the predominance of preclinical evidence, additional methodological considerations specific to animal research were incorporated, including the appropriateness of the sepsis model, intervention timing, route of administration, outcome assessment windows, and clinical relevance of reported endpoints. Eligibility was structured according to the PECO framework. The population comprised mammalian in vivo models of sepsis or endotoxemia. Exposure was defined as the administration of a chemically defined flavonoid compound or a well-characterized flavonoid-rich preparation. The comparator was a septic control group that received a vehicle, placebo, or no flavonoid treatment. The outcomes included survival, organ injury, inflammatory biomarkers, oxidative stress indices, endothelial or vascular dysfunction, and mechanistic endpoints. Studies were eligible if they were controlled in vivo experiments using established models of sepsis, including lipopolysaccharide (LPS)-induced endotoxemia, cecal ligation and puncture (CLP), cecal inoculum, or defined bacterial challenge. Studies were excluded if they were exclusively in vitro, lacked an appropriate septic control group, evaluated uncharacterized multi-component interventions without adequate controls, or were conference abstracts without sufficient extractable data.

A comprehensive literature search was performed in PubMed/MEDLINE, Scopus, Web of Science, and Embase from database inception to March 2026. The search strategy combined terms related to flavonoids, sepsis, inflammatory pathways, organ injury, and animal models using the Boolean operators. No language restrictions were applied. The core search string included combinations of flavonoid, flavonol, flavone, quercetin, kaempferol, luteolin, apigenin, fisetin, naringin, sepsis, septic shock, endotoxemia, cecal ligation and puncture, lipopolysaccharide (LPS), inflammation, NF-κB, cytokine, organ dysfunction, survival, animal, mice, “rat, and murine. Reference lists of relevant reviews and meta-analyses were manually screened. Additional targeted searches were performed to identify recent studies published between 2020 and 2026, formulation and bioavailability studies, and relevant primary studies not captured by the initial database search.

After the removal of duplicates, titles and abstracts were screened against predefined eligibility criteria. Full texts were retrieved for all potentially relevant records and assessed for their final inclusion. Studies identified from systematic reviews, meta-analyses, or reference lists were included if they fulfilled the eligibility criteria and provided sufficient extractable data. Any uncertainty during the screening or eligibility assessment was resolved through discussion and consensus. Data were extracted using a standardized framework. The extracted variables included the year of publication, country, study design, animal species, strain, sex, sample size, sepsis model, method of induction, model severity, and timing of outcome assessment. Intervention-related variables included flavonoid compounds, subclasses, sources or purities, doses, routes of administration, timings relative to sepsis induction, frequencies of administration, and formulation strategies. The outcome variables included survival or mortality, organ injury, inflammatory response, oxidative stress, endothelial or vascular function, and mechanistic readouts. Organ injury outcomes included histopathological findings and biochemical markers, such as serum creatinine, blood urea nitrogen, liver transaminases, and tissue injury scores. The inflammatory mediators included TNF-α, IL-6, IL-1β, HMGB1, and related cytokines. The oxidative stress indices included reactive oxygen species, malondialdehyde, superoxide dismutase, glutathione, and catalase. Safety and toxicity data were extracted when they were available. The risk of bias in animal studies was assessed using the SYRCLE Risk of Bias Tool. The evaluated domains included sequence generation, baseline comparability, allocation concealment, random housing, blinding of investigators and outcome assessors, incomplete outcome data, selective outcome reporting, and other potential sources of bias. Because methodological reporting in preclinical studies is often incomplete, the certainty of the evidence was interpreted conservatively. Greater emphasis was placed on the consistency of the direction of effect across independent studies, biological plausibility of mechanisms, reproducibility across different sepsis models, and concordance between functional outcomes and mechanistic findings.

Given the substantial heterogeneity in flavonoid compounds, doses, routes of administration, treatment timing, sepsis models, animal species, and outcome definitions, a formal de novo meta-analysis was not performed. The findings were synthesized narratively according to major outcome domains, including survival, organ protection, inflammatory modulation, oxidative stress attenuation, endothelial protection, and mechanistic pathways. When multiple studies evaluated the same flavonoid, the results were integrated to identify compound-specific patterns. Quantitative survival effects were contextualized using available preclinical meta-analytic evidence rather than being recalculated. The final synthesis prioritized consistency, biological coherence, translational relevance, and limitations affecting applicability to human sepsis.

RESULTS

A total of 80 studies were included, comprising predominantly preclinical animal experiments, supplemented by a limited number of systematic reviews and meta-analyses. The evidence base was overwhelmingly derived from rodent models, with only a single canine study identified and no large-scale human clinical trials. Across the included studies, more than 30 distinct flavonoids or flavonoid-rich preparations were evaluated. Quercetin emerged as the most extensively investigated compound, followed by kaempferol, fisetin, luteolin, and apigenin, reflecting a concentration of evidence around flavonol and flavone sub-classes. The predominant experimental models were lipopolysaccharide (LPS)-induced endotoxemia and cecal ligation and puncture (CLP), which together accounted for the majority of the study designs. A smaller subset employed live bacterial infection models, including Escherichia coli, CRAB, and MRSA, thereby enhancing translational relevance through the incorporation of pathogen-driven immune responses.

Table 1. summarizes the characteristics of all the included studies.

Study

Full text?

Study type

Flavonoid(s) tested

Animal model / cell line

Sepsis induction

Primary organ(s) studied

Qian Ren et al., 2019

No

In vivo

Fisetin [20]

Male C57BL/6J mice [20]

LPS i.p. (10 mg/kg) [20]

Kidney [20]

H. J. Park et al., 2018 F. Koç et al., 2020

Yes

No

In vitro + in vivo In vivo

Tamarixetin [21] Chrysin [22]

C57BL/6 and BALB/c mice; BMDCs [21] Rats [22]

LPS and E. coli K1 infection [21] LPS i.p. [22]

Lung, liver, kidney [21] Liver, lung, kidney [22]

A. Chauhan et al., 2019 Yi-Ru Liao & Jin-Yuarn

Yes

No

In vivo + in vitro In vivo

Isorhamnetin [15] Quercetin, Quercetin-3-glucuronide

Female BALB/c mice; HEK-Blue hTLR4 cells [15] Mice [23]

E. coli K1 infection [15] LPS i.p. [23]

Lung, liver, kidney [15] Peritoneal cavity [23]

Lin, 2015 W. Cui et al., 2019

Yes

In vivo

[23] Quercetin [24]

Wistar albino rats [24]

CLP [24]

Lung [24]

Haifeng Zhang et al.,

No

In vivo + in vitro

Fisetin [7]

Mice; BMDMs [7]

CLP [7]

Lung, liver, kidney [7]

2020 M. Karamese et al., 2016

No

In vivo

Apigenin [5]

Female Wistar albino rats

CLP [5]

Spleen [5]

Lichao Sun et al., 2017

No

In vivo

Acacetin [25]

(n=64) [5] Mice [25]

Sepsis-induced ALI [25]

Lung [25]

A. Shehata et al., 2024

No

In vivo

Morin [26]

Male mice (n=80) [26]

Not specified [26]

Kidney [26]

Yuanshuo Ouyang et al.,

2021

No

In vivo + in vitro

Acacetin [27]

Mice [27]

LPS injection [27]

Liver, lung [27]
Shanting Liao et al., No

In vivo

Baicalin [28]

Mice [28]

LPS injection [28]

Liver, kidney [28]

2016 P. Bayram et al., 2023

No

In vivo

Baicalein, Naringin [29]

Wistar albino rats (n=66)

CLP [29]

Not specified [29]

H. Lee et al., 2022

Yes

In vitro + in vivo

Rhamnetin [8]

[29] Female ICR mice; RAW

CRAB and E. coli infection

Lung, liver, kidney [8]

Y. D. Rattmann et al.,

No

In vivo

Myricetin and quercetin rhamnosides

264.7, HEK cells [8] Mice [2]

[8] CLP [2]

Lung, ileum [2]
2012

[2]

Shan Lu et al., 2021

No

In vivo + in vitro

Quercetin (nanoparticle) [18]

Mice; HK-2 cells [18]

LPS [18]

Kidney [18]

Zuqing Xu et al., 2023

No

In vivo

Kaempferol [30]

Mice [30]

CLP [30]

Kidney [30]

Murat Bıçakçıoğlu et al., 2023 M. Doğukan et al., 2021

Yes

No

In vivo In vivo

Quercetin (20 mg/kg) [31] Quercetin (20 mg/kg) [32]

Male Sprague Dawley rats (n=32) [31] Male rats (n=32) [32]

CLP [31] Cecal ligation [32]

Lung [31] Liver [32]

Yukun Liu et al., 2021

No

In vivo

Alpinetin (50 mg/kg IV) [33]

Mice [33]

CLP [33]

Multiple organs [33]

Lili Feng et al., 2014

No

In vivo + in vitro

Pentamethoxyflavanone (PMFA) [12]

Mice; M1 macrophages [12]

LPS and CLP [12]

Lung [12]

Xuan Zhu et al., 2022 Gaoxiang Li et al., 2024

No

No

In vitro + in vivo In vivo + in vitro

Kaempferol [16] Pinocembrin [34]

Mice; RAW264.7, HUVECs [16] Mice [34]

LPS [16] CLP and LPS [34]

Pulmonary v asculature [16] Vascular (thrombosis)

Yilin Wang et al., 2018

No

In vivo

Mangiferin [35]

Mice (n=24) [35]

CLP [35]

[34]

Lung [35]

D. Rabha et al., 2018

No

In vivo

Kaempferol (100 mg/kg oral) [36]

Mice [36]

CLP [36]

Lung [36]

Y. Zong & Huali Zhang, 2017

No

In vivo

Amentoflavone [9]

Rats [9]

CLP [9]

Lung [9]

Table 1. (continued 2)

Study

Full text?

Study type

Flavonoid(s) tested

Animal model / cell line

Sepsis induction

Primary organ(s) studied

Lili Feng et al., 2019

No

In vitro + in vivo

5,7,2',4',5'-Pentamethoxyflavanone

Mice [37]

LPS [37]

Lung [37]

Hong-bo Zhang et al., No

In vivo

[37] Astilbin [38]

Rats [38]

CLP [38]

Lung [38]

2017

Yuanfeng Zhu et al.,

No

In vitro + in vivo

Quercetin [11]

Mice; peritoneal

LPS [11]

Lung [11]

2019 S. Rungsung et al., 2022

No

In vivo

Luteolin (0.2 mg/kg IP) [39]

macrophages [11] Mice [39]

CLP [39]

Vascular (aorta) [39]

G. Kim et al., 2023

Yes

In vivo

Procyanidin B2 (0.5–2 mg/kg IV) [40]

Male C57BL/6 mice [40]

LPS i.p. [40]

Lung [40]

Mevlüt Doğukan et al., 2021 Yu-Ge Zhou et al., 2025

Yes

No

In vivo In vitro + in vivo

Quercetin (20 mg/kg oral) [41] Quercetin-3-β-aminobutyrate (HPS-β)

Male Sprague Dawley rats (n=32) [41] Mice; RAW264.7 [19]

Cecal ligation [41] LPS [19]

Liver [41] Lung, intestine [19]

Liangyong Deng et al.,

No

In vivo + in vitro

[19] Luteolin [6]

WT and TLR4-deficient

Not specified [6]

Liver [6]

2025

Lichao Sun et al., 2019

No

In vivo

Luteolin (20–80 mg/kg oral) [42]

mice [6] Mice (n=50) [42]

Sepsis-induced ALI [42]

Lung [42]

M. Karamese, 2023

No

In vivo

Naringin [43]

Wistar albino rats (n=30)

CLP [43]

Kidney [43]

Yu-fei Li et al., 2025

No

In vitro + in vivo

Protocatechuic aldehyde [44]

[43] Mice; macrophages [44]

LPS [44]

Not specified [44]

Y. Jafari-khataylou et al.,

No

In vivo

Troxerutin [45]

Mice [45]

LPS [45]

Liver [45]

2020 A. Soltanian et al., 2019

No

In vivo

Quercetin (2 mg/kg IV) [46]

Mixed-breed dogs (n=15)

LPS (0.1 µg/kg IV) [46]

Heart, liver [46]

Zheng Lijun et al., 2025

Yes

In vivo + in vitro

Apigenin (50 mg/kg) [47]

[46] C57BL/6 male mice;

LPS (5 mg/kg) [47]

Intestine [47]

Yanjun Zheng et al.,

No

In vivo + in vitro

Orientin [48]

Caco-2 cells [47] Mice; BMDMs,

LPS [48]

Lung [48]

2026

Haifeng Zhang et al.,

No

In vivo + in vitro

Fisetin (10 mg/kg IP) [49]

RAW264.7 [48] Mice; BMDMs [49]

CLP [49]

Lung, liver, kidney [49]

2020a

Jianying Wang et al.,

No

In vivo

Afzelin [50]

Mice [50]

CLP [50]

Kidney [50]

2021

Protective effects of

Yes

In vivo

Quercetin (20 mg/kg oral) [51]

Sprague Dawley rats

Intestinal ligation and

Kidney [51]

quercetin, 2022 Aya Mohamed et al.,

No

In vivo

Morin (50 mg/kg) [3]

(n=31) [51] Mice [3]

puncture [51]

LPS (5 mg/kg) [3]

Kidney [3]

2023

Jiafu Li et al., 2025

Yes

In vitro + in vivo

Kakkalide (20 mg/kg IP) [52]

Male C57BL/6J mice

CLP [52]

Lung, kidney [52]

Amira Rifdatari, 2017

No

In vivo

Flavonoid-containing A. paniculata

(n=40); HUVECs [52] Male Wistar rats (n=20)

LPS [53]

Duodenum [53]

Almaz Zaki et al., 2024

No

In vivo + in vitro

extract [53] Vitexin [54]

[53] C57BL/6 mice; MLE-12,

LPS [54]

Lung [54]

Pradipta Reza Syahruna et al., 2020

No

In vivo

Mangosteen peel extract [55]

RAW264.7 [54] Mice (n=30) [55]

Shigella dysenteriae i.p. [55]

Not specified [55]

Table 1. (continued 4)

Study Full text? Study type Flavonoid(s) tested Animal model / cell line Sepsis induction Primary organ(s) studied
N. Aisyah, 2017 No In vivo Flavonoid-containing A. paniculata extract [56] Rats [56] LPS [56] Ileum [56]
Edinildo de Oliveira Rodrigues Junior et al., 2023 No Systematic review + meta-analysis 30 different flavonoids [1] Various [1] Various [1] Multiple [1]
Jiawei Zhou et al., 2018 Yes Systematic review + meta-analysis Resveratrol [13] Various rodents [13] LPS and CLP [13] Multiple organs [13]
Yu-Cheng Chang et al., 2013 Yes In vitro + in vivo Quercetin [4] Male C57BL/6J mice; RAW264.7 [4] LPS i.p. (10 mg/kg) [4] Systemic [4]
M. Berköz et al., 2021 No In vivo Myricetin, Apigenin (100–200 mg/kg oral) [57] Mice (n=36) [57] LPS [57] Liver [57]
Weichao Ding et al., 2024 No In vitro + network pharmacology Kaempferol [58] MH-S cells [58] LPS [58] Lung (ARDS) [58]
Xia Cao et al., 2024 No In vivo + metabolomics Quercetin, Acacetin, Diosmetin (in YZC extract) [59] Mice [59] Not specified [59] Lung, intestine [59]
Bo-tao Chang et al., 2023 No In vitro + in vivo Mangiferin (20 mg/kg) [60] Mice; RAW264.7 [60] LPS [60] Liver, intestine [60]
Liuye Yang et al., 2024 No In vitro + in vivo Pimpinellin [61] C57 and PARP1 knockout mice [61] LPS [61] Not specified [61]
Lisa Savitri & Maria Do Carmo Da Costa Freitas, 2024 No In vivo Flavonoid-containing P. foetida extract [62] White male mice (n=24) [62] E. coli injection [62] Liver [62]
Rezya Salsabela et al., 2023 Yes In vivo Flavonoid-containing A. paniculata extract [63] Male Wistar rats (n=25) [63] LPS (5 mg/kgBW) [63] Systemic (CRP, ferritin) [63]
Naelaturroja Naelaturroja et al., 2020 No In vivo Flavonoid-containing I. cylindrica extract [64] Male DDY mice [64] LPS [64] Liver [64]
Hilal Üstündağ et al., 2025 No In vivo Propolis-based nanocomposites [65] Sprague-Dawley rats (n=42) [65] LPS (5 mg/kg) [65] Lung [65]
Mutiara Indah Sari et al., 2023 No In vivo Flavonoid-containing C. amboinicus extract [66] Male R. norvegicus (n=28) [66] Not specified [66] Liver [66]
Yang (楊斯皓), 2013 No In vitro + in vivo Tetramethoxyflavone (TMF) [67] C57BL/6 mice; RAW264.7 [67] LPS [67] Systemic [67]
Diding Heri Prasetyo & E. L. Suparyanti, 2013 Yes In vivo Propolis ethanol extract [14] Male R. norvegicus (n=40) [14] Cecal inoculum [14] Intestine [14]
Xiaoxue Bai et al., 2022 No In vitro + in vivo Maackiain [10] Mice; RAW264.7 [10] CLP and LPS [10] Multiple organs [10]
Wafiq Azizah et al., 2023 Yes In vivo Flavonoid-containing S. album extract [68] Male mice (n=27) [68] MRSA injection [68] Immune cells [68]
A. Esmat et al., 2019 Yes In vivo Propolis extract (250 mg/kg oral) [69] Male albino rats (n=40) [69] Cecal slurry [69] Liver, brain [69]
Riswanto Riswanto et al., 2020 No In vivo Flavonoid-containing M. oleifera extract [70] Male Wistar rats (n=30) [70] LPS [70] Liver [70]
Jingqian Su et al., 2024 No In vivo Turmeric kombucha (flavonoid-containing) [71] Mice [71] LPS [71] Lung [71]

Table 1. (continued 4)

Similar Articles

1 2 3 4 5 6 7 > >> 

You may also start an advanced similarity search for this article.

Study Full text? Study type Flavonoid(s) tested Animal model / cell line Sepsis induction Primary organ(s) studied
Kaiyuan Liu et al., 2024 No In vivo + in vitro Myricanol [72] Mice (incl. SIRT1-knockout) [72] LPS [72] Lung [72]
Shod Abdurrachman Dzulkarnain et al., 2024 Yes In vivo Flavonoid-containing P. betle extract [73] BALB/c mice [73] ESBL-producing E. coli [73] Lung, kidney, liver [73]
Devika Yuldharia, 2012 No In vivo Propolis ethanol extract [74] White male rats (n=40) [74] Cecal inoculum i.p. [74] Intestine [74]
Berty, 2010 No In vivo Angkak (flavonoid-containing) [75] Male BALB/c mice (n=24) [75] Cecal inoculum [75] Systemic (neutrophils) [75]
Danar Dwi Anandika, 2009 No In vivo Garlic extract (flavonoid-containing) [76] Male BALB/c mice (n=27) [76] S. aureus infection [76] Systemic (leukocytes) [76]
Kusni Kurnia Putri, 2012 No In vivo Propolis ethanol extract [77] Male rats (n=40) [77] Cecal inoculum (40 mg i.p.) [77] Systemic (lymphocytes) [77]
S. Abdul-Rahman et al., 2026 No Systematic review Genistein [17] Various (10 in vitro, 19 animals, 1 human) [17] Various [17]