Lysyl Oxidase (LOX) Family Members: Rationale and Their Potential as Therapeutic Targets for Liver Fibrosis

Abstract

The cross‐linking of structural extracellular matrix (ECM) components, especially fibrillar collagens and elastin, is strongly implicated in fibrosis progression and resistance to fibrosis reversal. Lysyl oxidase family members (LOX and LOXL1 [lysyl oxidase‐like 1], LOXL2 [lysyl oxidase‐like 2], LOXL3 [lysyl oxidase‐like 3], and LOXL4 [lysyl oxidase like 4]) are extracellular copper‐dependent enzymes that play a key role in ECM cross‐linking, but have also other intracellular functions relevant to fibrosis and carcinogenesis. Although the expression of most LOX family members is elevated in experimental liver fibrosis of diverse etiologies, their individual contribution to fibrosis is incompletely understood. Inhibition of the LOX family as a whole and of LOX, LOXL1, and LOXL2 specifically has been shown to suppress fibrosis progression and accelerate its reversal in rodent models of cardiac, renal, pulmonary, and liver fibrosis. Recent disappointing clinical trials with a monoclonal antibody against LOXL2 (simtuzumab) in patients with pulmonary and liver fibrosis dampened enthusiasm for LOX family member inhibition. However, this unexpected negative outcome may be related to the inefficient antibody, rather than to LOXL2, not qualifying as a relevant antifibrotic target. Moreover, LOX family members other than LOXL2 may prove to be attractive therapeutic targets. In this review, we summarize the structural hallmarks, expression patterns, covalent cross‐linking activities, and modes of regulation of LOX family members and discuss the clinical potential of their inhibition to treat fibrosis in general and liver fibrosis in particular.

Potential conflict of interest: Dr. Schuppan consults for, advises for, received grants from, and holds intellectual property rights with NorthSea. He consults for, advises for, and received grants from Boehringer Ingelheim. He consults for and advises for Pliant, UCB, Inversago, and Prometik.

Liver fibrosis (LF) is the result of chronic inflammation and a continuous wound‐healing process. It is characterized by excessive accumulation and reduced degradation of extracellular matrix (ECM) components and can progress to cirrhosis, which is characterized by severe liver architectural distortion, and the development of portal hypertension, liver functional impairment, and hepatocellular carcinoma (HCC).⁽¹⁾ Only recently, experimental and clinical studies indicate that even advanced LF and early‐stage cirrhosis are reversible once the prime injurious trigger, such as viral hepatitis B and C, is suppressed or eliminated.^(2‐4) Spontaneous reversal usually begins with resolution of profibrogenic inflammation and a halt of the ongoing fibrotic repair, with subsequent deactivation and/or apoptosis/necroptosis of myofibroblasts and, ultimately, ECM degradation, especially with still ill‐defined “pro‐resolution macrophages.”^(5,6) This corresponds to the regression criteria that were previously established prospectively by the Beijing classification based on liver biopsies before and after effective antiviral therapy, including the emergence of thin and delicate fibrotic septa, with little or no inflammation.⁽⁷⁾ The speed and extent of reversal depend on the etiology and degree of fibrosis; in general, advanced fibrosis reverses more slowly than mild fibrosis.⁽⁸⁾ However, despite successful elimination or suppression of causative factors, a substantial proportion, up to 25%, of patients with advanced cirrhosis do not regress.⁽²⁾ Therefore, apart from the removal of causative factors, pharmacological approaches that specifically target fibrosis to effectively slow down fibrosis progression or even induce regression of advanced fibrosis or cirrhosis are urgently required.^(9‐11)

Lysyl oxidase family members (LOX and LOXL1 [lysyl oxidase‐like 1], LOXL2 [lysyl oxidase‐like 2], LOXL3 [lysyl oxidase‐like 3], and LOXL4 [lysyl oxidase like 4]) are extracellular copper‐dependent enzymes that catalyze the cross‐linking of structural ECM components in fibrotic organs, including the liver.^(12‐17) The resultant ECM stabilization by cross‐linking contributes to fibrosis progression and retards fibrolysis, for example, through matrix metalloproteinases‐induced proteolytic degradation.⁽⁸⁾ Experimental evidence suggests that therapeutic targeting of LOX, LOXL1, or LOXL2 may represent an attractive strategy to target LF.^(12‐14) However, in recent clinical trials, a selective LOXL2‐blocking monoclonal antibody (mAb), simtuzumab, failed to reduce myelofibrosis, pulmonary fibrosis, and especially LF in patients with human immunodeficiency virus (HIV) and hepatitis C virus (HCV) coinfection, nonalcoholic steatohepatitis (NASH), and primary sclerosing cholangitis (PSC).^(18‐22) However, lack of potency of this allosteric LOXL2 antibody and insufficient target engagement within the fibrotic scar are likely causes, and there remains great potential for LOX and LOXL antagonistic therapies in fibrosis. We therefore systematically summarize the structures, expression patterns, modes of regulation, and the often multiple functions, apart from covalent cross‐linking activities and specificities, of the LOX family members, with a special focus on the treatment of LF.

Structural Basis of LOX Family Members for Catalyzing ECM Cross‐Linking

LOX enzymes are secreted, copper‐dependent amino oxidases. The five members, LOX and LOXL1‐4, share a highly conserved carboxyl terminus and a nonconserved amino terminus, with a copper‐binding motif, a lysyl‐tyrosyl‐quinone cofactor (LTQ) domain, and in the case of LOXL2‐4, several scavenger receptor cysteine‐rich (SRCR) domains⁽²³⁾ (Fig. 1). Details of the structural description of LOX family members are provided in Supporting Information S1.

Fig. 1: Domain structure of the LOX family members. LOX and LOXL1 differ from LOXL2 to LOXL4 mainly in their diverse N terminal regions. SRCR domains; LTQ (lysyl‐tyrosyl‐quinone cofactor); and the signal peptide, propeptide, and copper‐binding motif with the sequence, His‐X‐His‐X‐His (H), are color coded.

Expression Patterns of LOX Family Members in Healthy and Fibrotic Livers

Expression of LOX family members is strictly controlled during liver development, whereas they show aberrant expression and enhanced collagen and elastin cross‐linking activity in fibrotic livers.^(12‐14,16)

Restricted Expression in Healthy Livers

LOXL1 enzyme is abundantly expressed in multiple fetal organs, including liver of mice, but reduced and locally restricted in liver of normal adult mice,⁽²⁴⁾ indicating a spatiotemporal expression pattern. The increased expression of LOXL1 is paralleled by increased type I procollagen and tropoelastin expression during later stages of liver fibrogenesis.^(12,25) In contrast, the high abundance of LOXL1 expression and enzymes in the healthy fetal liver coincides with limited substrate availability (type I procollagen and tropoelastin) and may therefore not contribute significantly to difficult‐to‐reverse ECM deposition in the fetal liver. We analyzed the publicly available transcriptome data of the LOX family members in multiple adult normal tissues.^(26,27) In agreement with most earlier studies that were focused on certain organs,^(24,28,29) the five LOX family members are constitutively expressed at relative high levels in tissues such as bladder, lung, heart, kidney, stomach, gallbladder, placenta, and endometrium and at particularly low levels in the central nervous system, bone marrow, and normal liver (Fig. 2).

Fig. 2: Transcript levels of LOX family members in normal human and mouse tissues. Expression levels are shown as log₂‐transformed reads per kilobase per million mapped reads (RPKM). Data have been extracted from several studies deposited at the National Center for Biotechnology Information (NCBI; https://www.ncbi.nlm.nih.gov). The heatmap diagram was generated using MeV software (https://sourceforge.net/projects/mev‐tm4/). Blue, white, and red shades represent low, median, and high abundance, respectively. The deeper the blue, the lower the expression; the deeper the red, the higher the expression. In human tissues, RPKM ranges from 0.066 ± 0.013 to 21.207 ± 4.836 (mean ± SD, n = 2‐7 studies); in mouse tissues, RPKM ranges from 0.036 to 113.248 (n = 1 study).

LOX Family Members Are Highly Up‐regulated in Fibrotic Livers

Several LOX family members have been reported to be up‐regulated in actively fibrosing tissues, such as lung, bone marrow, heart, peritoneum, kidney, mucus membranes, and liver. Here, activated (myo‐)fibroblasts, endothelia, and epithelia are major cellular sources (Supporting Table S1).

In fibrotic livers of patients with Wilson’s disease or primary biliary cirrhosis (PBC), LOX and LOXL2 mRNA, but not LOXL1 and LOXL3 mRNA, are found around and within the nuclei of hepatocytes, and the LOXL2 enzyme is located not only in hepatocytes, but also in fibrotic liver tissue.⁽²⁹⁾ In LF attributed to hepatitis B and C, LOX enzyme transcripts are strongly expressed in myofibroblasts of the fibrotic lesions, but not hepatocytes.⁽²⁹⁾ In NASH mouse models, similar to the human biopsy studies in Wilson’s disease and PBC,⁽²⁹⁾ LOXL2 expression was not only increased in myofibroblasts, but also in the cytoplasm and nucleus of perivenular hepatocytes.⁽¹⁶⁾ Expression levels of LOX family members in liver biopsies of patients with NASH cirrhosis has not been reported, but serum levels of LOXL2 positively correlate with fibrosis and portal pressure.⁽¹⁹⁾ Serum LOXL2 levels were also reported to be positively correlated with liver stiffness measured by transient elastography in HCV patients with cirrhosis with a sustained virological response to antiviral therapy at both baseline and at the end of follow‐up,⁽³⁰⁾ decreasing after viral eradication.^(31,32) Great variability in serum LOXL2 was found among fibrotic patients, possibly attributable to the involvement of LOXL2 in multiple processes, apart from fibrogenesis, such as promoting tumor growth and metastatic niche formation in HCC.^(32,33) Moreover, in a small cohort of pretreatment HCV patients with cirrhosis, hepatic LOXL2 expression was correlated with LOXL2 serum levels, but not with fibrosis stage, posttreatment.^(31,32) We previously showed that in advanced CCl₄‐induced LF, LOX and LOXL1 transcripts were up‐regulated 3‐ and 34‐fold, respectively, whereas LOXL2 and LOXL3 expression remained unchanged and LOXL4 transcripts were even down‐regulated.⁽¹²⁾ These findings are in line with data by Perepelyuk et al.,⁽³⁴⁾ who only found LOX and LOXL1 up‐regulated in experimental CCl₄‐induced and biliary fibrosis, whereas LOXL2 and LOXL3 were down‐regulated and LOXL4 remained undetectable. To further clarify the expression patterns of the LOX family enzymes in fibrotic livers, we further analyzed the publicly available gene expression profiles associated with human and rodent LF in the Gene Expression Omnibus database (https://www.ncbi.nlm.nih.gov/geo/). LOX and LOXL1 mRNA expression is markedly up‐regulated in fibrotic liver tissues across humans and mice, regardless of the underlying etiology; however, the extent of mRNA expression changes of LOXL2, LOXL3, and LOXL4 vary among different studies, which requires further confirmation (Table 1). In summary, the expression of LOXL1>LOX>LOXL2>LOXL3/4 is increased with advancing fibrosis, whereas changes in mild fibrosis are modest.⁽¹²⁾ However, more protein and functional data are needed, to establish a hierarchy of LOX targets for therapeutic intervention.

Table 1 - Expression of LOX Family Members in Microarray Data Sets From Fibrotic Livers

Data Set ID	Sample Size	Etiology	mRNA Expression of LOX Family Members
			LOX	LOXL1	LOXL2	LOXL3	LOXL4
GSE6764 (H)	N = 10, D = 13	HCV	Stable	Up	Stable	Stable	Up
GSE14323 (H)	N = 19, D = 41	HCV	Up	Up	Up	ND	ND
GSE33814 (H)	N = 13, D = 19	NAFLD	Up	Up	Stable	Up	Up
GSE49541 (H)	N = 40, D = 32	NAFLD	Up	Up	Up	Stable	Up
GSE84044 (H)	N = 43, D = 81	HBV	Up	Up	Up	Up	Up
GSE55747 (M)	N = 4, D = 6	CCl4	Up	Up	Up	Stable	Stable
GSE27640 (M)	N = 2, D = 2	CCl4	Up	Up	Stable	Stable	Stable
GSE80601 (M)	N = 5, D = 5	CCl4	ND	Up	Up	Stable	Stable

NAFLD indicates nonalcoholic fatty liver disease, a term comprising NASH, but also mild disease. The fibrotic or cirrhotic groups (D) are compared to normal groups (N). Gene expression profiles are from human livers (H) and mouse (M) liver. “Up” represents significant up‐regulation (adjusted P value, <0.05); “stable” indicates no change.

Abbreviation: ND, not determined.

Interplay Among LOX Family Members

Synergistic or compensatory effects must be considered when one or multiple LOX family members are selected as therapeutic target(s). Thus, small interfering RNA (siRNA)‐mediated silencing of LOX in primary mouse lung fibroblast significantly reduced the expression of LOXL2, but not LOXL1; LOXL2 knockdown promoted the expression of LOX, but had no influence on the expression of LOXL1; LOXL1 knockdown increased the expression of LOX, but decreased that of LOXL2.⁽³⁵⁾ In embryonic lungs of LOXL3^–/– mice, expression of pulmonary LOX and LOXL1 remained unchanged, that of LOXL2 was down‐regulated, and that of LOXL4 was up‐regulated.⁽³⁶⁾ However, in an in vivo study of bleomycin (BLM)‐induced pulmonary fibrosis, no interdependence of LOX family enzymes could be shown, given that knockdown of individual members by specific siRNAs did not significantly affect the expression of other members.⁽¹⁷⁾ The mutual regulation among the LOX family members in the liver is even less clear, which may limit the utility of targeting a single LOX member to treat LF. The potential interplay among LOX members can be visualized using the Search Tool for the Retrieval of Interacting Genes (STRING) database (version 10.5; https://string‐db.org/). All the LOX family members are predicted to interact with each other, implying that targeting one member may perturb the expression or activity of others, either in a single cell or between different cell types (Fig. 3). From this analysis, it can also be seen that the first and second neighbors adjacent to LOX family members constitute two closely connected clusters covering key elements in elastic and collagen fiber synthesis, stressing the role of LOX family members in the formation and maintenance of elastic and collagenous fibers.

Fig. 3: Potential interactions among LOX family members. The interactive network is derived from the STRING database and has been visualized with the Cytoscape software (https://cytoscape.org/). The nodes represent LOX family members and their neighbors; and the lines between any two nodes represent potential interactions between two members (score, >0.9). Abbreviations: BMP1, bone morphogenetic protein 1; COL1A1, collagen type I alpha 1 chain; COL1A2, collagen type I alpha 2 chain; COL2A1, collagen type II alpha 1 chain; COL3A1, collagen type III alpha 1 chain; COL5A1, collagen type V alpha 1 chain; COL5A2, collagen type V alpha 2 chain; COL7A1, collagen type VII alpha 1 chain; EFEMP2, EGF containing fibulin extracellular matrix protein 2; ELN, elastin; FBLN1, fibulin 1; FBLN2, fibulin 2; FBN3, fibrillin 3; FBN1, fibrillin 1; FBN2, fibrillin 2; HSPG2, heparan sulfate proteoglycan 2; JAG1, jagged 1; LAMA3, laminin subunit alpha 3; LAMB3, laminin subunit beta 3.; LAMC2, laminin subunit gamma 2; MFAP1, microfibril‐associated protein 1; MFAP2, microfibril‐associated protein 2; MFAP3, microfibril‐associated protein 3; MFAP4, microfibril‐associated protein 4; MFAP5, microfibril‐associated protein 5; PCOLCE, procollagen C‐endopeptidase enhancer; TLL1, tolloid‐like 1; TLL2, tolloid‐like 2.

Covalent Cross‐Linking of Structural ECM Proteins

LF, as fibrosis of other organs, is characterized by ECM remodeling and stiffening. Cirrhotic livers accumulate up to 10‐fold increased collagenous and noncollagenous ECM components compared to normal livers.^(37,38) Moreover, covalent cross‐linking occurring among intra‐ and intermolecular protein chains of ECM proteins results in the stabilization and attenuation of ECM turnover, a process that prominently requires LOX family members. A schematic diagram of the roles of LOX family members, especially LOX, LOXL1, and LOXL2, in liver collagen and elastin cross‐linking is shown in Fig. 4. Detailed description is provided in Supporting Information S2.

Fig. 4: Schematic diagram of the roles of LOX family members in collagen and elastin cross‐linking. Abbreviations: N, N‐terminal; C, C‐terminal. +, activation; ?, no data.

Crosstalk of LOX Family Enzymes With Regulators of LF

The SRCR domains of LOX family members provide the structural basis for a crosstalk with other regulators of fibrosis. Through these interactions, and in addition to their function in collagen and elastin crosslinking, LOX family members are involved in cellular proliferation and differentiation, in angiogenesis and in cancer invasion/metastasis.^(39,40) These functions may even be more relevant for fibrogenesis than their cross‐linking activities. In the following, we summarize these other functions and the upstream and downstream modulators of LOX family member expression (Supporting Information S3 with Tables S2‐S4 and Figs. S1‐S2).

Effects of LOX Enzymes on Fibrogenic Activation of Myofibroblasts

Activated hepatic stellate cells (HSCs) and myofibroblasts, the central ECM‐producing cells in LF, are themselves a major cellular source of most LOX family members in liver fibrogenesis.^(34,41) Most LOX enzymes, primarily LOX, LOXL1, and LOXL2, are involved in fibroblast‐to‐myofibroblast transition in organs like the lung,⁽¹⁷⁾ heart,⁽⁴²⁾ kidney,⁽⁴³⁾ intestine,⁽⁴⁴⁾ and liver.⁽¹⁴⁾ In rodent models of parenchymal LF, LOX and LOXL inhibition by β‐aminopropionitrile (BAPN) decreased liver stiffness⁽⁴¹⁾ and liver HSC/myofibroblast activation, paralleled by a reduction of alpha‐smooth muscle actin (α‐SMA)‐positive cells, by 63%,⁽¹³⁾ and silencing of LOXL1 by an adeno‐associated viral vector encoding LOXL1 short hairpin RNA (shRNA) in CCl₄‐induced LF reduced both α‐SMA mRNA and protein by 79% and 61%, respectively.⁽¹²⁾ In Mdr2 knockout mice (Mdr2^−/−) with severe biliary fibrosis, LOXL2 neutralization by the AB2003 mAb reduced α‐SMA protein levels by 57.4%.⁽¹⁴⁾ To date, it appears that LOXL3 and LOXL4 do not play a role in HSC/myofibroblast activation and in fibrosis of the liver and other organs. The potential mechanisms by how extracellular (by mechanotransduction of stiff ECM) and intracellular LOX regulate HSC/myofibroblast activation are illustrated in Fig. 5 and Supporting Information S4.

Fig. 5: Potential mechanisms underlying LOX family members on the activation of HSC/myofibroblasts. Arrows represent direct activation; T‐type arrow indicates inhibition. +Pi indicates phosphorylation. Abbreviations: SMAD2/3, SMAD family member 2/3; PI3Kα, phosphoinositide‐3‐kinase α; TGFβ2, transforming growth factor beta‐2; AKT, serine/threonine kinase; mTORC1, mammalian target of rapamycin complex 1; 4EBP1, eukaryotic translation initiation factor 4E binding protein 1; S6K1, ribosomal protein S6 kinase B1.

Lessons Learned From Simtuzumab Trials and Driving Forces of the Continuous Search for Effective LOX Therapeutics

LOXL2 became the first LOX family member to be specifically targeted, with development of the humanized mAb, simtuzumab (Gilead), and its broad evaluation in phase 2 clinical studies of patients with myelofibrosis,⁽²²⁾ idiopathic pulmonary fibrosis,⁽²¹⁾ HCV/HIV‐induced advanced LF,⁽¹⁸⁾ and NASH‐associated bridging fibrosis/compensated cirrhosis,⁽¹⁹⁾ and PSC⁽²⁰⁾ (Table 2).

Table 2 - Clinical Trials of LOXL2 mAb Simtuzumab in Fibrotic Diseases

Disease	Clinicaltrials.gov Number	Etiology	Sample Size	Populations	Administration Method	Dose and Group	Duration of Treatment
MF(22)	NCT01769196	Primary/PV/ET	N = 54	7 sites in the United States	Intravenous infusion	200 mg/2 weeks	24 weeks
						700 mg/2 weeks
IPF(21)	NCT01769196	Unknown	N = 544	183 sites in 14 countries	Intravenous infusion	125 mg/L week	254 weeks
ALF(18)	NCT01707472	HIV, HCV or HIV/HCV	N = 18	Single center	Intravenous infusion	700 mg/2 weeks	22 weeks
LF(19)	NCT01672866	NASH	N = 219	80 study sites in North America and Europe	Subcutaneous injection	75 mg/L week	96 weeks
						125 mg/L week
LC(19)	NCT01672879	NASH	N = 258	80 study sites in North America and Europe	Intravenous infusion	200 mg/2 weeks	96 weeks
						700 mg/2 weeks
						placebo
CLD(20)	NCT01672853	PSC	N = 234	61 sites in North America and Europe
	Subcutaneous injection	75 mg/L week	96 weeks
						125 mg/L week
						placebo

Abbreviations: ALF, advanced liver fibrosis; CLD, compensated liver disease; ET, essential thrombocythemia; IPF, idiopathic pulmonary fibrosis; LC, liver cirrhosis; MF, myelofibrosis; PV, polycythemia vera.

The initial excitement for this approach, based on preclinical evidence implicating LOXL2 in the pathogenesis of LF^(14,45,46) and on this novel concept of directly targeting the ECM, has been dampened by disappointing clinical trials with humanized anti‐LOXL2 antibody simtuzumab that did not demonstrate a clinical benefit.^(47,48) The following reasons may explain the clinical failure and help in the planning of LOX targeting in the postsimtuzumab era: (1) the extremely selective nature of targeting with specific antibody that lacks cross‐inhibition of other LOX isoforms except LOXL2 itself; (2) the pure extracellular nature of antibody action; and (3) the lack of evidence that the therapeutic antibody actually reaches its target in (human) liver scar tissue. The latter is a misconception in human trials, which interpreted serum LOXL2 levels as surrogate pharmacodynamic efficacy markers instead of assessing the actual target engagement within the liver ECM. In a small study of 8 HCV‐infected patients, serum LOXL2 levels showed a good correlation with LOXL2 expression in the liver, but a high variability between patients with the same fibrosis stage^(31,32); this may suggest that LOXL2 inhibition may achieve best clinical results in patients with higher serum LOXL2 levels. The unfortunate outcome has been a (temporary) loss of enthusiasm among clinicians and drug developers for targeting LOXL2 and the entire LOX family. This emphasizes the need to clearly evaluate target engagement preclinically and in future clinical trials, in the public health interest rather than private drug development.

There are several other potential reasons for the lack of efficacy observed in the simtuzumab trial, in stark contrast to strong data supporting an important role of LOXL2 in fibrosis,^(29,30,32) including a reasonable efficacy profile of anti‐LOXL2 mAb in preclinical models.^{(14,42,45,46)} Thus, there appears to be a problem with the drug in relation to its target, given that the simtuzumab antibody itself is only a weak LOXL2 antagonist because of its indirect, allosteric mechanism of action, achieving only 50% inhibition of enzymatic activity in vitro.⁽⁴⁵⁾ Combined with the lack of evidence that it penetrates the liver scar, this may be insufficient to ensure sufficient blocking of local LOXL2 activity, especially in the 5‐fold more dense ECM of human versus rodent LF.^(38,49) Another important consideration is that important synergistic pathogenic activities of intracellular LOXL2 have been described, and that intracellular drug activity may be required for efficacy.^(14,50) Here, cell‐permeable small molecules targeting LOXL2 may achieve much greater potency than an antibody that is limited to the extracellular space and that may not penetrate well into scar tissue. Accordingly, our preliminary data suggest that the small‐molecule LOXL2 inhibitor, PAT‐1251, is significantly more effective than simtuzumab (mAb AB0023) in the BALBc Mdr2^–/– model of severe biliary LF.⁽⁴⁶⁾ In addition, there is the hypothetical possibility that LOXL2 may be a wrong target in humans, given that fundamental species differences may exist regarding the pathological role of LOXL2 in human versus rodent LF. In summary, all these points need to be addressed for novel LOX enzyme inhibitors in an experimental approach before considering future clinical studies.

Despite the disappointing results of clinical studies with the LOXL2‐targeted mAb, simtuzumab, other LOX family members, which demonstrate a different tissue distribution and relative contribution to collagen cross‐linking, particularly LOX and LOXL1, remain promising antifibrotic drug targets. As mentioned above, the much greater up‐regulation of LOX and LOXL1, rather than LOXL2‐4, expression in advanced LF suggests that they may play an important role in fibrogenesis,⁽¹²⁾ Moreover, LOX and LOXL1, not other lysyl oxidases, bind directly to tropoelastin with the help of fibulin 4 and fibulin 5 (FLBN5), accelerating elastogenesis that is implicated as a key factor in retarding fibrosis reversal.⁽²⁵⁾

Beyond Simtuzumab: Novel Therapeutic Agents Specifically Addressing LOX Enzymes

This section will systematically summarize the potential agents targeting LOX enzymes, including small‐molecule inhibitors, oligonucleotides, and blocking antibodies (Table 3).

Table 3 - Inhibitors for LOX Family Members and Their Antifibrotic Effect in Preclinical Models

Agents	Characterization	Targets	Cell/Animal Models	Diseases	Antifibrotic Efficacy	References
d‐Penicillamine	Copper chelator	LOXL2	A549 lung cells	Tumor fibrosis	—	(51)
		LOXL2	HepG2 cells	LF	—	(29)
Tetrathiomolybdate	Copper chelator	LOX/LOXL1‐2	BLM treatment	Pulmonary fibrosis	++	(52)
BAPN	Small‐molecule inhibitor	LOX/LOXL1‐4	hSEs	Cutaneous fibrosis	++	(53)
			NT‐treated mice	Peritoneal fibrosis	++++	(15)
			BLM‐treated mice	Pulmonary fibrosis	++	(54)
			CCl4‐treated mice	LF	++	(13)
			Rats with BCNI	Cavernosal fibrosis	++	(55)
			Rats with IP	Cavernosal fibrosis	—	(56)
PXS‐5153A	Small‐molecule inhibitor	LOXL2/3	CCl4‐treated mice	LF	+++	(57)
			Streptozotocin/high‐fat diet–fed mice	LF	++	(57)
PXS‐S2B	Small‐molecule inhibitor	LOXL2	Diabetic mice	Glomerulosclerosis	++	(58)
PAT‐1251	Small‐molecule inhibitor	LOXL2	COL4A3–/– Alport mice	Renal fibrosis	+++	(43)
N‐acetylcysteine	Small‐molecule inhibitor	LOX	BLM‐treated mice	Pulmonary fibrosis	—	(59)
siRNA	Small‐interfering RNA	LOXL2	Mouse lung fibroblasts	Pulmonary fibrosis	+++	(60)
		LOXL1	LX‐2 cells	LF	+++	(12,61)
		LOX2/3	Human lung fibroblasts	Pulmonary fibrosis	++++	(17)
shRNA	Short hairpin RNA	LOXL1	CCl4‐treated mice	LF	+++	(12)
		LOX/LOXL4	Foreskin fibroblasts	Cutaneous fibrosis	+/++	(53)
		LOX	BLM treatment	Pulmonary fibrosis	+++	(54)
miRNA constructs	Small noncoding RNAs	LOX	NT‐induced mice	Peritoneal fibrosis	+++	(15)
AB0023
	mAb	LOXL2	CCl4‐treated mice	LF	++	(45)
			TAA‐treated mice	LF	+	(14)
			MDR2 (ABCB4)−/− mice	LF	++	(14)
			DDC‐fed mice	LF	++	(14)
			Mice with TAC	Myocardial fibrosis	+++	(42)
			BLM‐treated mice	Pulmonary fibrosis	++	(45)

+, antifibrotic efficacy ≤25%; ++, antifibrotic efficacy ≤50% but >25%; +++, antifibrotic efficacy ≤75% but >50%; ++++, antifibrotic efficacy ≤100% but >75%; —, unknown. Antifibrotic efficacy is defined as the reduction percentage of fibrosis area in vivo study or the decrease percentage of LOXs expression in vitro study after relative inhibitors treatment.

Abbreviations: ABCB4, ATP‐binding cassette subfamily B member; BCNI, bilateral cavernous nerve injury; COL4A3, collagen type IV alpha 3 chain; DDC, 3,5‐diethoxycarbonyl‐1,4‐dihydrocollidine; hSEs, human skin equivalents; IP, ischemic priapism; MDR2, multidrug‐resistance transporter; NT, carbon nanotubes; TAA, thioacetamide; TAC, transverse aortic constriction.

Small‐Molecule Inhibitors

The copper‐binding motif and LTQ domain of LOX enzymes are key structural targets for these agents. Conventional copper chelators include, for example, d‐penicillamine, tetrathiomolybdate, disulfiram, dithiocarbamates, and 2‐mercaptopyridine‐N‐oxide, largely being nonspecific and broad‐spectrum LOX inhibitors, with unwanted off‐target effects.⁽⁶²⁾d‐penicillamine is widely used in patients with Wilson’s disease,^(63,64) pulmonary fibrosis,⁽⁶⁵⁾ and Indian childhood cirrhosis,^(66,67) with clear clinical benefits in preventing organ toxicity and fibrosis progression. Inhibition of LOX activity and collagen accumulation was implicated as a major underlying mechanism.^(29,68) Tetrathiomolybdate has been confirmed preclinically to protect against liver and pulmonary fibrosis,^(52,69) but also in experimental multiple sclerosis.⁽⁷⁰⁾ Except for d‐penicillamine and tetrathiomolybdate, the antifibrotic potential of the other copper chelators is unknown, but would deserve further investigation. BAPN, PXS‐5153A, PXS‐S2B, and PAT‐1251 are inhibitors of the LTQ with which they form an irreversible covalent complex. BAPN has been extensively used as a preclinical experimental tool in multiple fibrosis models serving as a pan‐inhibitor blocking all five LOX enzymes. However, a recent report suggests that BAPN may be ineffective in inhibiting LOX and LOXL4 in vitro.⁽⁶²⁾ Unlike the pan‐LOX inhibitors like BAPN, several rationally designed and selective small‐molecule inhibitors demonstrated promising activity in multiple organ fibrosis models. These include Pharmakea’s PAT‐1251 (anti‐LOXL2) and several inhibitors developed by Pharmaxis with pan‐LOX, LOXL2/3, or LOX/LOXL2 specificities.^{(43,46,57,58)} These new molecules are promising antifibrotic agents. Thus, in contrast to antibody, they are orally available and can more easily reach intracellular and dense extracellular compartments, as found in advanced fibrosis.^{(43,46,57,58)}

Oligonucleotide‐Based Therapies

This is an active area of drug development that is particularly suited to treat liver diseases. When conjugated to certain trimeric branched N‐acetyl‐galactosamine moieties, oligonucleotides are almost exclusively taken up by the hepatocyte‐specific asialoglycoprotein receptor; also, underivatized lipoplexes and nanoparticles are preferentially engulfed by nonparenchymal liver cells, including HSCs and myofibroblasts, which makes them attractive carriers for oligonucleotides targeted at genes that are prominently expressed in these cells, such as LOX family members, with knockdown efficiency up to 95% in vivo.⁽¹¹⁾ Antisense oligonucleotides (ASOs), siRNAs, and microRNA (miR) inhibitors (anti‐miRs) or mimics (pro‐miRs) serve as therapeutic platforms.⁽⁷¹⁾ To date, oligonucleotide therapies aimed at LOX enzyme expression in fibrosis are in their infancy.

Summary and Perspective

It is well documented that certain LOX family members, especially LOX, LOXL1, and LOXL2, are up‐regulated in fibrosis in general and in fibrotic livers in particular. Their increased activity stabilizes collagen and elastin by covalent cross‐linking, rendering the fibrotic scar less reversible and increasing ECM stiffness, which, in itself, promotes fibrogenic activation of the central fibrogenic effector cells, HSCs, and myofibroblasts. Moreover, the functions of LOX family members go far beyond mere cross‐linking and creation of a stiffer matrix, including the modulation of various cytokines, and yet little explored intracellular activities that affect cellular differentiation and proliferation. To date, small‐molecule inhibitors, siRNAs, and ASOs that target specific LOX family members provide a growing armamentarium to explore how far their inhibition may be useful to treat fibrosis of multiple organs and, prominently, the liver. Importantly, they come close to the class of direct antifibrotics, that is, agents that specifically address the ECM or fibrogenic cells.⁽¹⁰⁾ In summary, the LOX family members will continue to represent attractive antifibrotic drug targets, especially in combination with other antifibrotic agents.

Author Contributions

H.Y., D.S., Y.P., J.D.J. and W.C. substantially contributed to the conception and design. W.C. performed the acquisition of data. W.C., H.Y., D.S. and A.T.Y. analyzed and interpreted the data. W.C. drafted the article. D.S., H.Y., Y.P. and J.D.J. revised it critically for important intellectual contents. H.Y. and D.S. finally approved the version to be published.