Patent Application
20240117436
The application contains a Sequence Listing which has been submitted electronically in .XML format and is hereby incorporated by reference in its entirety. Said .XML copy, created on Nov. 17, 2023, is named “MAGI CIP.xml” and is 80,206 bytes in size. The sequence listing contained in this .XML file is part of the specification and is hereby incorporated by reference herein in its entirety.
This invention relates to methods to prevent and treat human lipedema and to methods for diagnosis of lipedema or the determination of a predisposition for lipedema.
Lipedema is a chronic and progressive pathologic condition mainly characterized by an abnormal body fat distribution. It affects extremities with abnormal fat deposition in thighs and legs and in some cases also the arms, while the trunk, hands and feet remain unaffected [Kruppa et al., 2020]. Lipedema is an autosomal dominant genetic disease that mainly affects women. It is characterized by excess deposition of subcutaneous adipose tissue, pain, and anxiety [Paolacci et al., 2019; Precone et al., 2019; Michelini et al., 2022]. The genetic and environmental etiology of lipedema is still largely unknown, and while rare, it is suggested to be misdiagnosed as obesity or lymphedema [Warren et al., 2007; Forner-Codero, 2012; Fife et al., 2010]. Lipedema patients can be distinguished from these two conditions by a series of features such as body disproportion, bilateral symmetry, hematoma tendency and scarce influence of diet, exercise and bariatric surgery.
Contrary to obese patients, the increase of fat in lipedema often causes symptoms such as pain and increased vascular fragility and is not responsive to diet or exercise. Contrary to lymphedema, the tissue in lipedema is soft to the touch. Involvement of steroid hormones was postulated. Indeed, its manifestations commonly arise in females in phases of hormonal changes, and steroid hormones are known to be involved in adipogenesis, anxiety, and pain, three common features of lipedema.
It has been estimated that about 10% of the woman are affected by lipedema worldwide [Buck D W and Herbst K L, 2016]. Male cases have been described in very few reports. For this reason, the involvement of sexual hormones in the etiology of the disease has been postulated several times [Torre et al., 2018; Bauer et al., 2019]. In line with this hypothesis, manifestations commonly arise in phases of hormonal changes (puberty, pregnancy or menopause) in females [Torre et al., 2018]. There is very strong evidence of a genetic base for the condition, since an autosomal dominant hereditary pattern was found in many families [Buso et al., 2019].
While genetic factors apparently regulate subcutaneous adipose tissue distribution, so far, no monogenic cause of non-syndromic primary lipedema have been discovered until the inventors' work.
Generally, patients with lipedema undergo differential diagnosis from other disorders, and other genes are screened to exclude the patient as having a known diagnosis of another disorder of subcutaneous adipose tissue (ADRA2A, AKT2, ALDH18A1, CIDEC, LIPE, LMNA, MFN2, NSD1, PALB2, PLIN1, POU1F1, PPARG, TBL1XR1) and of localized lipodystrophies (AGPAT2, AKT2, BSCL2, CAV1, CAVIN1 (PTRF), CIDEC, LIPE, LMNA, PLIN1, PPARG, ZMPSTE24).
Therapies are performed to help relieve symptoms and prevent frustration. When possible, a conservative management is suggested and this includes manual lymph drainage, appropriate compression therapy with custom-made, flat-knitted compressive clothing, psychosocial therapy, patient education on self-management, physiotherapy and exercise therapy (such as low impact, cycling, walking or other exercise or movements), dietary counseling and weight management.
Lipedema fat is resistant to diet therapy. Current dietary approaches are aimed at lowering body weight through a hypocaloric diet, inhibiting systemic inflammation with antioxidant and anti-inflammatory components and reducing water retention [Di Renzo et al., 2021].
In some cases, if symptoms impair quality of life, the potential indication for surgery should be evaluated. Liposuction therapeutic benefit has not yet been evaluated in any randomized, controlled trials. Liposuction can reduce leg circumference, pain, feeling of tightness, tendency to form hematomas, improving quality of life. In highly advanced stages of the disease (i.e. in presence of lymphedema and fibrosis) dermato-fibro-lipectomy may be indicated.
This invention provides methods for diagnosing lipedema or identifying agents for treating a patient having lipedema or a predisposition for lipedema. The methods comprise one or more of the following steps:
Only the identification of AKR1C1 and AKR1C2 as the first lipedema-associated genes rendered the diagnostic and therapeutic approaches herein described possible. The identification of the gene and its linkage to lipedema opened the way to diagnose and treat the disease of lipedema. Since AKR1C1 and AKR1C2 are the first genes associated with the molecular diagnosis of non-syndromic lipedema, there are currently no molecular diagnostic alternatives.
Indeed, with their study, the inventors argue in favor of the involvement of AKR1C1 and AKR1C2 in lipedema [Michelini et al., 2020; reference herein]. AKR1C1 is a gene highly expressed in the subcutaneous tissue and it has been suggested that its activity in the regulation of steroid hormone levels plays an important role in the accumulation of subcutaneous fat depots. The enzyme expressed by this gene, the 20α-hydroxysteroid dehydrogenase (20α-HSD), metabolizes progesterone, a hormone that prompts lipogenesis [Blanchette et al., 2005]. Up to the inventors' previous and current findings, AKR1C1 or AKR1C2 have not been implicated in any genetic condition characterized by or including lipedema among its clinical manifestations.
The association may be due to rare genetic variants or common polymorphisms that alter enzymatic function and can also be caused by epigenetic alterations.
In fact, the inventors previously found that AKR1C1, an essential enzyme for steroid hormone regulation, is mutated in a family affected by lipedema [Michelini et al., 2020], suggesting that these hormones may play a role in the pathogenesis of the disease. It is known that sex hormones also determine the anatomical site of the accumulation of adipose tissue, and dysfunction of sex steroids result in abnormal fat distribution in predisposed subjects, especially in females at the time of puberty [Grigoriadis et al., 2021; Gavin et al., 2013]. The homeostasis of steroid hormones is finely regulated by enzymes such as aldo-keto reductases (ARK1C), hydroxysteroid dehydrogenase (HSD) and aromatases expressed in adipocytes, preadipocytes, and mature adipose tissue [Tchernof et al., 2015; Blouin et al., 2009].
The four human AKR1Cs are multifunctional enzymes with overlapping activities on a broad range of substrates. They possess approximately 320 amino acid residues and share at least 84% amino acid sequence identity. AKR1C1 and 1C2 in particular differ by only seven amino acids, with only one amino acid difference at the active site [see description below]. All four AKR1Cs can exert 3-, 17- and 20-ketoreductase activity, though each has its own distinct preferences for position, stereochemistry and substrate. AKR1C1 is the major 20α-reductase that inactivates progesterone, whereas AKR1C2 preferentially acts as a 3α-reductase, with particular importance in the deactivation of dihydrotestosterone (DHT) to 3α-Adiol [Penning et al., 2019]. Progesterone and DHT play opposite roles with regard to fat accumulation, with progesterone prompting lipogenesis and DHT inhibiting adipogenesis [Kiani et al., 2021].
To support the claim that AKR1C2 overexpression may be a feature of lipedema, the inventors described the AKR1C1 L54V variant found in one of the studied lipedema patients, which turns AKR1C1 into an enzyme possessing AKR1C2 activity [Zhang et al., 2014], as well as the AKR1C2 Ser320PheTer2 variant causing a deletion in the C-term tail of the protein. The inventors then decided to study the expression levels of AKR1C2 in a separate cohort of lipedema patients, and indeed, alteration in AKR1C2 mRNA expression was detected.
Finally, the inventors observed that variants gathered from the GWAS catalog database that were typical to obesity patients, lied in AKR1C2 promoter regions (known to affect fat deposition). Considering that lipedema is often misdiagnosed as obesity, these variants found in AKR1C2 promoters provided new insights into the correlation of AKR1C genes to lipedema.
The role of leucine 54 in substrate selectivity has been already elucidated [Penning et al., 2019; Couture et al., 2003]. Its bulky side chain significantly confines the spatial movement of the steroid in its cavity, restricting the flexibility provided by valine 54 in AKR1C2 [Zhang et al., 2014; Couture et al., 2003]. Of the seven amino acids that are different between AKR1C1 and AKR1C2, one is located in the active site at position 54. The replacement of leucine 54 in AKR1C1 with valine 54 found in AKR1C2, is the only mutation that generated an enzyme with identical properties to AKR1C2, further indicating the importance of residue 54 in determining the substrate specificity of the two enzymes [Matsuura et al., 1997] while the reverse mutation V54L in AKR1C2 converts the enzyme into AKR1C1 (
The AKR1C enzymes exert their HSD activity mainly in subcutaneous adipose tissue as the reduction and inactivation of steroid hormones [Penning, Trevor M., 1997; Blouin et al., 2006]. AKR1C1, which is highly expressed in adipocytes and subcutaneous fat, is responsible for inactivating progesterone to 20α-hydroxyprogesterone, a main metabolite in isolated mature adipocytes. In this way, AKR1C1 decreases the levels of progesterone in peripheral adipose tissue [Brozic et al., 2009; Michelini et al., 2020]. AKR1C1 also reduces DHT to its 3β metabolite, 5α-androstane-3β,17β-diol (3β-Adiol), which is a potent agonist of the estrogen receptor beta (ERβ). ERβ may enhance the lipid burning process in adipose tissue [Katzer et al., 2021], suggesting that loss of this function of AKR1C1 may further contribute to fat deposition. In rats, progesterone reverses the weight-reducing actions of estradiol [Gray et al., 1981; Wade, G N, 1975]. This suggests that a diminished activity of AKR1C1 could lead to high levels of progesterone in adipocytes and a consequent increase of lipogenesis mediated by this hormone [Michelini et al., 2020]. Meanwhile, AKR1C2 is important for DHT inactivation in preadipocytes, and overexpression of it might lead to the elimination of the androgen inhibitory effects on adipogenesis [Penning et al., 2019]. Therefore, this single mutation is involved in lipedema by affecting the steroid homeostasis two-fold: diminishing the activity of AKR1C1 on the reduction of progesterone which mediates lipogenesis, while enhancing the activity of AKR1C2 on the reduction and inactivation of DHT which inhibits adipogenesis.
As previously described, in one method for the diagnosis of lipedema and/or for the individuation of treatments thereof, the variants of step (i) are detected from gDNA, in particular by single nucleotide polymorphism (SNP) analysis for detecting differences between alleles of AKR1C1 genes, that reside within a region of human chromosome 10, or detected through NGS (Next Generation Sequencing) or Sanger technologies.
In an advantageous embodiment of the method for the diagnosis of lipedema and/or for the individuation of treatments thereof, the variants of step (i) are selected from known loss-of-function (LoF) SNPs indicated in table 1 or from a list of selected SNPs as indicated in table 2 or 4. The SNPs can, for example, be selected on the basis of the following criteria: only missense variants; absent in homozygous state; frequency below 0.1%. The selected variants are subsequently preferably studied by functional modelling to verify their impact, for example in terms of binding affinity to certain compounds. This permits to study for one or more particular variants found in a patient the binding affinity to pharmaceutically active compounds, to find the compound that best fits for the particular variant and thus for the patient being affected by this variant.
Preferably, the AKR1C1 variants are selected from the group consisting of: c.840C>A (p.Asn280Lys), or are selected from c.160T>G (p.Leu54Val), c.162A>T (p.Leu54Phe), c.638T>A (p.Leu213Gln), the p.Leu54 and p. Leu213 variants being particularly preferred. The four variants above are particularly interesting as they have been found in lipedema patients.
In particular, the missense variant p.(Leu213Gln) in AKR1C1, the gene encoding for an aldo-keto reductase catalyzing the reduction of progesterone to its inactive form, 20-α-hydroxyprogesterone, suggests a partial loss-of-function resulting in a slower and less efficient reduction of progesterone to hydroxyprogesterone and an increased subcutaneous fat deposition in variant carriers. The p.(Leu213Gln) variant, to the knowledge of the inventors, is the first one ever identified in a lipedemia family.
Being an inducible gene [Pallai et al., 2010], AKR1C1 expression in the blood can be a marker of the disease. Similarly, urinary and blood plasma or serum metabolites can be used as disease markers and have diagnostic value.
In another embodiment of the method for the diagnosis of lipedema and/or for the individuation of treatments thereof, the mRNA of step (ii) or the enzymatic substrate or product or metabolite of step (iii) is detected in a biological sample, in particular in blood, urine and/or adipose tissue specimens.
In a preferred embodiment, the enzymatic substrate or product or metabolite of step (iii) is a steroid derivative or a prostaglandin.
Preferably, the biological sample of step (iii) is screened with an antibody that specifically binds to the AKR1C1 enzymatic substrate or product or metabolite or the biological sample is treated or converted by AKR1C1 enzyme.
Preferably, the enzymatic substrate or product in step (iii) is selected among 20α-hydroxysteroid dehydrogenase (20α-HSD); PGF2α and its derivatives, in particular by measurement of 15-keto-13,14-dihydro-PGF2α, the major metabolite of PGF2α in plasma; or isoprostane 8-iso-Prostaglandin F2α (8-iso-PGF2α).
In a preferred embodiment of the method for the diagnosis of lipedema, in step (iii) the levels of at least one of the following metabolites 3α-Hydroxy-5α-pregnan-20-one, 3α-Hydroxy-5β-pregnan-20-one, 3β-Hydroxy-5α-pregnan-20-one, 3β-Hydroxy-5β-pregnan-20-one, 5α-Pregnane-3,20-dione, 5β-Pregnane-3,20-dione, Pregn-4-ene-3,20-dione, 20α-Hydroxy-pregn-4-ene-3-one, 5α-Pregnane-3α,20α-diol, 5β-Pregnane-3α,20α-diol, 5α-Androstan-17β-ol-3-one, 5α-androstane-3α,17β-diol, 21-hydroxy-5α-pregnan-20-one, 3α,21-dihydroxy-5α-pregnan-20-one, Pregnanetriol/17-hydroxypregnanol one, 15-keto-13,14-dihydro-PGF2α, in particular 8-iso-Prostaglandin F2α progesterone and/or 5alpha-dihydrotestosterone is determined in a body fluid.
In another embodiment of the method for the diagnosis of lipedema, in step (iii) the ratio (androstanediol1.5×20β-DH-cortisone)/(20β-DH-cortisone+[cortisol×log(estriol)] in a body fluid is determined.
AKR1C1 is a target of natural and synthetic molecules capable of modulating its activity. Benzodiazepines such as medazepam represent a class of non-competitive inhibitors of AKR1C1. Synthetic derivatives of pyrimidine, phthalimide and anthranilic acid potently inhibited AKR1C1 (Brozic et al., 2009). Compounds provided with a core structure of steroid carboxylate and flavones are instead AKR1C1 competitive inhibitors. Among natural compounds, liquiritin has been discovered as a selective and potent AKR1C1 inhibitor capable of reducing the progesterone metabolism in cells [Zeng et al., 2019].
Prostanoids, acting via peroxisome proliferator-activated receptor gamma (PPARγ), a fundamental receptor in fatty acid storage and glucose homeostasis, have been proposed as potent regulators of fat cell differentiation. Indeed, in vitro studies showed that prostaglandin J2 (PGJ2) binds and activates PPARγ acting as a potent adipogenic hormone; inversely, prostaglandin F(2α) (PGF2α), which has PPARγ antagonist properties, is a potent antiadipogenic factor [Quinkler et al., 2006; Volat et al., 2012]. Another proof of the involvement of prostaglandins (PG) in the regulation of adipocyte differentiation came from the use of PG analogues as hypotensive agents in the treatment of glaucoma, extensively described in literature reports. Indeed, patients treated with topical therapies based on PG analogues showed periorbital fat changes as an adverse effect. These molecules can directly lead to reduced orbital fat by inhibiting adipogenesis [Taketani et al., 2014]. Aldo-keto reductases have been reported as major regulators of white adipose tissue development with antiadipogenic properties supported by PGF2α synthase activity [Quinkler et al., 2006; Volat et al., 2012]. Indeed, PGF2α can be synthesized from PGD2 and PGE2 by the enzymes AKR1C (1, 2 and 3) [Quinkler et al., 2006; Dozier et al., 2008] and Akr1b7 [Volat et al., 2012]. In vitro studies demonstrated that PGD2 enhances adipocyte differentiation while PGE2 and PGF2α suppress adipogenesis [Miller et al., 1996].
A further aspect refers to a method of treating and/or preventing of human lipedema in a subject, the method comprising administering or applying to a subject in need thereof a therapeutically effective amount of a compound of natural or synthetic origin, preferably contained in a food supplement, cream or ointment, suitable for modulating the activity of AKR1C1 or of prostaglandins.
In one embodiment of the method of treating and/or preventing of human lipedema in a subject, the compound modulates the catalytic activity of the AKR1C1 enzyme, and comprises at least one of the compounds indicated in table 6, in particular benzodiazepines, such as medazepam, derivatives of pyrimidine, phthalimide and anthranilic acid, competitive inhibitors with a core structure of steroid carboxylate and flavones, and liquiritin. Advantageously, the compound is selected from the group consisting of flavanone, flavone, 3-hydroxyflavone, 5-hydroxyflavone, equilin, diazepam, 20α-hydroxydydrogesterone, coumarin, glycyrrhetinic acid, 7-hydroxyflavone and 3,7-dihydroxyflavone.
In another embodiment of the method of treating and/or preventing of human lipedema in a subject, the compound is suitable for modulating prostaglandins and comprises at least one of the compounds indicated in table 7.
Variants of the method of treating and/or preventing of human lipedema in a subject foresee, that the step of administering or applying to a subject in need thereof a therapeutically effective amount of a compound of natural or synthetic origin is preceded by a step for the diagnosis of lipedema according to the invention that confirmed the tested person is affected by lipedema. Advantageously, the confirmation of the fact that the tested person is affected by lipedema is obtained by the detection of a biomarker in a body fluid in a concentration exceeding a determined limit value.
A further aspect relates to a composition for the treatment of human lipedema, in particular in the form of a food supplement, cream or ointment, comprising a compound that modulates the catalytic activity of the AKR1C1 enzyme or of prostaglandins, in particular at least one of the components indicated in tables 6-10.
An additional aspect of the invention relates to a food supplement comprising the composition according to the invention. Another aspect of the invention relates to a cream comprising the composition according to the invention. A final aspect of the invention refers to an ointment comprising the composition according to the invention.
the distance between C3 of 5α-DHT and C4N of NADPH.
Diagnostic methods can comprise the sequencing (through next generation sequencing [NGS] or Sanger technologies) of the AKR1C1 gene, or portions of it, or through whole genome and whole exome approaches for the diagnosis of lipedema. Single nucleotide polymorphism (SNP) analysis is also useful for detecting differences between alleles of AKR1C1 genes that reside within a region of human chromosome 10. Within this region, about 700 known SNPs have been reported to date. A list of known loss-of-function (LoF) SNPs is shown in table 1. In addition, a series of SNPs to have effect on protein function and an association with lipedema selected on the basis of the following criteria are listed in table 2: only missense variants; absent in homozygous state; frequency below 0.1%.
1identified in lipedema patients.
2These variants have been created in a mutagenesis experiment described by Couture et al.
3The enzyme activity parameters described by Couture et al. were used to calculate those of the first variant identified by Michelini et al. in a family with lipedema, p.(Leu213Gln).
The complete sequence of AKR1C1 and AKR1C2 genes are well known and documented in literature. The following links take to a database (https://www.ensembl.org/) that discloses details about both genes and whole sequences:
AKR1C1 gene summary:
https://www.ensembl.org/Homo_sapiens/Gene/Summary?db=core;g=ENSG00000187134;r=10:4963253-4983283
AKR1C1 transcript sequence (MANE select):
https://www.ensembl.org/Homo_sapiens/Transcript/Exons?db=core;g=ENSG00000187134;r=10:4963253-4983283;t=ENST00000380872
The sequence listing reports the complete sequence of the AKR1C1 gene (Homo sapiens) as SEQ ID NO 1, the corresponding coding sequence (cDNA) as SEQ ID NO 2 and two isoform corresponding proteins as SEQ ID NO 3 and SEQ ID NO 4. The DNA and corresponding protein sequence of the variant c.928A>C (p.(Ile310Leu)) are depicted as SEQ ID NO 5 and SEQ ID NO 6, respectively.
AKR1C2 gene summary:
https://www.ensembl.org/Homo_sapiens/Gene/Summary?db=core;g=ENSG00000151632;r=10:498777 5-5018031
AKR1C2 transcript sequence (MANE select):
https://www.ensembl.org/Homo_sapiens/Transcript/Exons?db=core;g=ENSG00000151632;r=10:4987 775-5018031;t=ENST00000380753
The sequence listing reports the sequence of the AKR1C2 gene (Homo sapiens) as SEQ ID NO 7, the corresponding exons as SEQ ID NO 8 and the protein corresponding to the exons as SEQ ID NO 9.
Further details regarding the identification of missense AKR1C1 variants in lipedema patients, sequencing, molecular modelling etc. are described in Michelini S, Chiurazzi P, Marino V, Dell'Orco D, Manara E, Baglivo M, Fiorentino A, Maltese P E, Pinelli M, Herbst K L, Dautaj A, Bertelli M., Aldo-Keto Reductase 1C1 (AKR1C1) as the First Mutated Gene in a Family with Nonsyndromic Primary Lipedema. Int J Mol Sci. 2020 Aug. 29; 21(17):6264. doi: 10.3390/ijms21176264. PMID: 32872468; PMCID: PMC7503355.
From structural analysis and molecular dynamics it was found that (
The NADP(H)-binding residues are highly conserved and include Thr23, Asp50, Ser166, Asn167, Gln190, Tyr216, Leu219, Ser221, Arg270, Ser271, Phe272, Arg276, Glu279 and Asn280, which contribute toward the binding affinity and specificity of the cofactor (see
To describe the interaction of the enzyme with cofactor and substrate in energetic terms, thus to furnish an energy landscape of binding, molecular dynamics simulations were run on the AKR1C1/steroid/NADP(H) ternary complex, and binding energy was calculated as well by use of the MMPBSA (Genheden and Ryde, 2015) method and GROMACS molecular dynamics software (Abraham et al., 2015). The overall energies of binding for the two are (Table 3):
The MMPBSA method also allowed for quantification of the contribution to binding of each amino acid, allowing the impact of an amino acidic missense substitution to be evaluated as follows (see also Table 4). MMPSA profile of STR binding shows three amino acids account for 50% of the binding energy: Tyr24, Leu54, and Trp227; another significant contribution is given by Asp50, Tyr55, Trp86, Val128, Ile129, Leu306, showing an overall hydrophobic nature of the binding (see
MMPBSA profile of NADP(H) binding is dominated by charge pairs giving prominent repulsion/attraction peaks between charged amino acids of the protein and the phosphate groups of the cofactor. The four most prominent negative binding energy peaks derive from Lys33, His222+Arg223, Lys270, Arg276, all neighboring the phosphate group on the 2′ position of the ribose ring that carries the adenine moiety (see
Multiple alignments of protein sequences produce a matrix of aminoacids; by elaborating the columns as vectors, entropy of aminoacidic positions can be calculated according to Shannon, describing the amount of variability through a column in the alignment. The lower the value, the lower the variability accepted by the position. The inventors aligned 120 sequences from the AKR1C family to derive Shannon entropy (Strait & Dewey, 1996) of each position; values for each missense mutation from Table 2 (AKR1C1 selected variants) are reported in Table 4.
In silico mutagenesis of AKR1C1 and molecular dynamics simulations, entropy evaluation, binding energy for cofactor and for substrate allowed for the determination of the structural impact of variants, thus the structural consequence prediction on AKR1C1 that are conducive of loss function for many of the selected mutations in Table 2. Mutations are reported alongside their predicted effect in Table 4.
‡references: (Penning et al., 2019; Hara et al., 1996; Matsuura et al., 1997).
1These variants have been created in a mutagenesis experiment described by Couture et al.
2The enzyme activity parameters described by Couture et al. were used to calculate those of the first variant identified by Michelini et al. in a family with lipedema, p.(Leu213Gln).
3This variant is not described in the above database, the respective DNA and protein sequences are reflected by SEQ ID NO 5 and 6, respectively.
In the following the Applicant reports a detailed descriptions of structural consequences of variants found in lipedema families.
In the inventors' patients, four missense mutations were found, namely Leu54Val, Leu54Phe, Asn280Lys, and Leu213Gln. The effects of such mutations on enzyme folding, stability, and biological activity have been studied with structural biology, and molecular dynamics approach to evaluate their involvement in lipedema development.
Starting with Leu54Val and Leu54Phe, the role of Leu54 in substrate selectivity has been already elucidated (Penning et al., 2019; Hara et al., 1996; Matsuura et al., 1997), and can be summarized as follows (see also
Similarly, the interaction between the steroid and the enzyme is disrupted by the replacement of the same Leu54 by phenylalanine, as shown by the molecular dynamics simulation. In the wildtype, Leu54 and Trp227 play a significant role in binding the steroid by interacting with opposite faces of the polycyclic ring of the ligand and contribute as much as 33% of the overall binding energy. Mutation of Leu54 to Phe, although enhancing the hydrophobic nature of the interaction, introduce a second large, aromatic sidechain in place hampering the ligand entrance in the site and conducive of binding disruption (see
Interestingly, phenylalanine is present at position 54 in the wildtype, non-human AKR1C8P, but here the steric hindrance with the opposite amino acid 227 is compensated by the presence of the smaller asparagine. At the same time, the cumbersome tryptophan is ‘shifted’ to position 228. Nonetheless, 1C8 preserve the same 20α-HSD activity of 1C1. As previously mentioned, this may indicate coevolution between positions 54 and 227.
Referring now to Asn280Lys, it can be said that asparagine 280 takes part in cofactor binding; together with Gln279 it is responsible for adenine group binding through a hydrogen bond to the amine group (see
Although such variant involved the replacement of a small side chain with a bulky one, molecular modelling showed how hydrophobic moiety of lysine can be easily accommodated by displacement of water molecules. MD simulation confirmed a small effect is exerted on the protein structure, while the missing H-bond acceptor capability of Lys led to the loss of interaction with the adenine ring, resulting in the aromatic ring flipping away from its position, also because of the attraction of Lys280 to the phosphate group. The optimal binding geometry is then disrupted rather than folding.
AKR1C1 is a member of the AKR1C family of enzymes that share a high percentage of amino acid sequence identity (from 84 to 98%). This family catalyzes NADPH dependent oxydoreductions either for the biosynthesis or inactivation of steroid hormones, bile acids and neurosteroids. All AKR1C enzyme catalyze a sequential ordered Bi-Bi substrate enzyme reaction. In particular, AKR1C1 in involved in the “alternative pathway” of androgen biosynthesis inactivating the most potent androgen 5alpha-dihydrotestosterone (5alpha-DHT) to 5alpha-androstane-3beta,17beta-diol, a potent agonist of ERbeta which exerts anti-proliferative effect. Androgens play an important role in regulation of body fat distribution in humans. They exert direct effects on adipocyte differentiation in a depot-specific manner, via the androgen receptor (AR), leading to modulation of adipocyte size and fat compartment expansion. AKR1C1 can also regulate the cellular concentration of allopregnanolone by preventing its formation from progesterone and by catalyzing its inactivation. Indeed, AKR1C1 catalyzes progesterone reaction to form the less potent progestogen 20alpha-hydroxy-4-pregnen-3-one, reduce 5alpha-pregnane-3,20-dione (5alpha-DHP) to form 20alpha-hydroxy-5alpha-pregnan-3-one or 3alpha-hydroxy-5alpha-pregnan-20-one (allopregnanolone) to a less neuroactive 5alpha-pregnane-3alpha,20alpha-diol. AKR1C1 therefore is involved in the inactivation of allopregnanolone, that acts in the central nervous system as positive allosteric modulator of gamma aminobutyric acid receptor A (GABAA). As other enzyme of the family can reduce also 20alpha-hydroxy-5alpha-pregnan-3-one to 5alpha-pregnane-3alpha,20alpha-diol. Progesterone has lipogenic action on adipose tissue by upregulating adipocyte determination and differentiation through 1/sterol regulatory element-binding protein 1c (ADD1/SREBP1c) expression in primary cultured preadipocyte from rat parametrial adipose tissue (Lacasa et al., 2001). ADD1/SREBP1c promotes adipocyte differentiation and gene expression linked to fatty acid metabolism (Kim and Spiegelman, 1996). The levels of progesterone and 5alpha-dihydrotestosterone can be detected in body fluids. Levels of progesterone ranges during normal menstrual cycles from 0 ng/ml (follicular phase) to 28 ng/ml (central luteal phase), values range from 11 to 422 ng/ml during pregnancy, while in post menopause or in males, levels of progesterone are less than 1.2 ng/ml. Levels of 5alpha-DHT range from 250-990 pg/ml in males, from 24-368 in pre-menopause females and from 10-181 in post menopause females.
A recent study revealed that the best combination to diagnose polycystic ovary syndrome (PCOS), including up to four steroids, was a ratio comprising androstanediol, estriol, 20βDHcortisone and cortisol accordingly to the following formula: (androstanediol1.5×20β-DH-cortisone)/(20β-DH-cortisone+[cortisol×log(estriol)]. This ratio was significantly increased in PCOS compared to controls at a threshold value of ≥435 (Dhayat et al., 2018). Considering the activity of the AKR1C1 enzyme, this ratio reasonably has diagnostic value in lipedema.
AKR1C1 is also involved in catalyzing the synthesis of prostaglandins in humans (Dozier et al., 2008). It has been shown that prostaglandin 2 alpha (PGF2α) inhibited adipogenesis by activating at its specific receptor on preadipocytes (Lepak and Serrero, 1995; Taketani et al., 2014). In mice, a decrease in intra-adipose tissue PGF2α levels following Akr1b7 ablation leads to increased adiposity, a phenotype that is reversed by the chronic administration of Cloprostenol, a PGF2α agonist (Volat el al., 2012). PGF2α and its derivatives can therefore be used as molecular diagnostic/prognostic markers and therapeutic agents also in lipedema. PGF2α can be reliably quantified by measurement of 15-keto-13,14-dihydro-PGF2α, the major metabolite of PGF2α in plasma (Helmersson et al., 2005). The isoprostane 8-iso-Prostaglandin F2α (8-iso-PGF2α), a prostaglandin-like molecule, is a quantitative ROS biomarker used to measure oxidative stress in vivo which correlates positively with BMI, intra-abdominal fat and waist circumference (Milne et al., 2015; Jia et al., 2019). Both molecules can be easily quantified in different body fluids such as plasma, serum or urine.
A list of AKR1C1 metabolites for use in diagnostics is reported in table 5.
In the literature, a number of natural and synthetic compounds are known to exert a modulatory action on the key human progesterone-metabolizing enzyme, AKR1C1.
A list of compounds for treatments for lipedema comprising the use of natural molecules or chemicals that modulate the catalytic activity of the AKR1C1 enzyme are shown in table 6.
Passiflora coerulea
arietinum) and in other legumes
PGE2 and PGF2α and its analogue (viprostol, latanoprost, isopropyl unoprostone, bimatoprost) can exhibit antiadipogenic properties. Some active constituents from Chinese herbs as ricinoleic acid, acteoside, amentoflavone, quercetin-3-O-rutinoside and hinokiflavone were predicted to be prostaglandin D2 synthase (PTGDS) inhibitors (Fong et al., 2015). Inversely, other natural supplements such as chlorella and green tea are proposed be used to decrease PGE2 and PGF2α levels (Koeberle et al., 2009; Haidari et al., 2018).
A list of compounds for treatments of lipedema comprising the use of natural molecules or chemicals that modulate prostaglandins are shown in table 7.
Biota orientalis
Platycladus orientalis
Platycladus orientalis
Ricinus communis
Cassia species
Natural and synthetic compounds that modulate AKR1C1 and listed in Table 6 were submitted to molecular docking procedure by using Autodock Vina 1.2 with the following parameters: AKR1C1 and NADPH coordinate from PDB entry 1MRQ; amino acids Tyr24, Leu54 and Trp227 set as flexible sidechain: docking box set centered at x=4.29 y=33.9 z=17.06 with size x=17.39 y=11.16 z=12.67, vina scoring function. Results are reported as binding affinity in Kcal/mol (the lowest, the better) in Table 8. Taking 2 Kcal/mol as the common threshold for binding energy significance, we have 11 top compounds (in bold); significantly 6 out of 11 are simple flavones/flavonones (in bold and italics). The double stacking interaction of B ring with Tyr24 phenol and Trp227 indole rings is the driving force of the interaction.
Flavanone
Flavone
3-Hydroxyflavone
5-Hydroxyflavone
Equilin
Diazepam
20α-Hydroxydydrogesterone
Coumarin
Glycyrrhetinic Acid
7-Hydroxyflavone
3,7-Dihydroxyflavone
The analysis has been repeated for AKR1C1 mutant Leu54Val by using the same parameters (Table 9). Such mutation is known to convert enzymatic activity of AKR1C1 into that of AKR1C2, which might eliminate androgen inhibitory effects on adipogenesis favouring progression of adipogenesis (Kiani et al., 2021), thus selective targeting of such mutation would modulate its possible effect on fat deposition. Again, flavones are among the favorites, but with lower affinity and competing with natural steroids like equilin or with large pentacyclic molecules glycyrrhetinic acid; this is due to the lower steric hindrance of valine vs. leucine resulting in less selective active site.
AKR1C1 Leu54Phe mutant is the other variant affecting substrate binding site accessibility presently analyzed. Oppositely but coherently with Leu54Val flavones are the tighter binders due to the incremented steric hindrance of phenylalanine which is able to stacking interact with A/C rings of the binder (Table 10).
The molecules of tables 9 and 10 have been analyzed considering the interaction with two specific variants, both on nucleotide 54, of particular interest is the Leu54Val variant. For every substance indicated in table 8, it is possible to identify through a study determining the affinity to AKR1C1 the most efficient one for a patient with a specific variant, as done for a patient with a variant on nucleotide 54.
Turning now to the surprising discovery of the inventors that AKR1C2 overexpression is a frequent feature of lipedema, a cohort of 19 lipedema patients and 2 affected family members from the family previously described in [Michelini et al., 2020] were enrolled in the study (Table 11).
The L213Q family tree of the inventors' previous study [Michelini et al., 2020] is represented in
qPCR analysis was performed on blood RNA extracted from a pool of 21 patients with lipedema (Table 12,
For the qPCR analysis total RNA was extracted from blood using the Tempus™ Spin RNA Isolation Kit following manufacturer protocols. The SuperScript VILO cDNA Synthesis Kit was used to generate first strand cDNA. Quantitative real-time polymerase chain reaction (qPCR) was performed by using the PowerUp™ SYBR™ Green Master Mix (Thermofisher) on a QuantStudio 3 Real-Time PCR Systems. The primers used in the qPCR experiments were previously described and are the following:
In two additional patients, two other variants were found (Table 12).
One included a variant in the regulatory region downstream AKR1C1, which may affect the relative expression levels between AKR1 C1 and AKR1C2; on the other hand, the other variant consisted in the c-term removal of AKR1C2, which is reported to affect DHT reduction rate in pig AKR1C1. The C-terminal region significantly contributes to the NADPH-dependent reductase activity for the 5α-DHT reduction [Son et al., 2015], an activity reserved for AKR1C2 in humans. We performed molecular dynamics study to evaluate the effect of this deletion. The results show that DHT adapts a more stable conformation in the truncated protein, where it is properly sandwiched between Val54 and Trp227, with Trp227 interacting with the β-face of the steroid; an interaction which is disrupted in the wildtype (
Additionally, the effect of variants located in regulatory regions, namely rs28571848 (chr10:5019979) and rs34477787 (chr10:5071991) to overexpression of AKR1C2 and fat accumulation is already reported in [Ostinelli et al., 2021].
Using the GWAS catalog database, the inventors gathered variants present in subjects affected by obesity. The majority of them were located upstream AKR1C2 (Table 13) in proximity to the regions reported in [Ostinelli et al., 2021], namely the binding sites of retinoid acid-related orphan receptor and the glucocorticoid receptor, two transcription factors important to the regulation of AKR1C2 expression in adipose tissue.
Since lipedema is still difficult to be diagnosed, and is often misdiagnosed as obesity, the inventors discovered that these variants yield important insight in developing the link between the relative expression of AKR1C genes (in particular AKR1C2 overexpression) and lipedema. At this regard, table 13 lists variants extracted from obesity patients in the GWAS catalog database.
All the variants reported in Table 13 are located in regulatory regions of AKR1C1, AKR1C2, and AKR1C3 enzymes.
The following paragraphs deal with therapeutics for lipedema individuated by the inventors. A comprehensive list of polyphenols was submitted to molecular docking by using Autodock Vina 1.2.2 [Eberhardt et al., 2021], with AKR1C1/AKR1C2 and NADPH coordinates from PDB entry 1MRQ [Couture et al., 2003]. From molecular dynamics simulations, W227, L54, Y24 contributed to more than 50% of the overall binding energy, implying their importance in substrate specificity; thus, they were set as flexible residues during docking. The box was located at the active site, centered at x=4.29, y=33.9, z=17.06 with dimensions x=17.39, y=11.16, z=12.67 enclosing the residues that are responsible for the binding of steroids. The docking results are reported as binding affinity in Kcal/mol (the lower, the better).
Two databases were used to select potential inhibitors, namely off-label molecules (db1) and a wide range of polyphenols and other molecules (db2). From the list of the inhibitors triazolam and alprazolam seem to be the most interesting due to their specific binding capabilities to AKR1C2 and not AKR1C1. Both scored higher than the natural substrate of AKR1C2 (DHT). However, among the natural molecules, Flavanones are among the best binders to AKR1C2, although still scoring less than its natural substrate (DHT), while flavones are the best binders to AKR1C1. Flavones and flavanones differ from which other in that the latter have the C ring saturated, and unlike flavones, the double bond between positions 2 and 3 is saturated.
Molecules that inhibit AKR1C2 are of great interest, especially after evidence that AKR1C2 overexpression may be a feature of lipedema. Nonetheless, achieving specificity in inhibiting AKR1C2 without affecting the activity of AKR1C1 will be difficult, due to the similarity of these enzymes' active sites, with the main driving force of the interaction being the stacking interaction of the B ring with Y24 and W227, while L54 and V54 in AKR1C1 and AKR1C2 respectively, are responsible for substrate specificity. In any case, it is a plausible approach to use molecules that are only specific for AKR1C2 but not for AKR1C1 in the lipedema therapy and/or that exhibit a high negative binding affinity that is more negative than the binding affinity of DHT.
Finally, while there are differences in preferences towards specific molecules, it can be seen that the whole class of polyphenols binds with similar affinities to the enzyme's natural substrate, acting as competitive inhibitors. This, once again, demonstrates the ability of polyphenols to act as natural remedies and to treat lipedema.
Tables 14 and 15 show the results for AKR1C2 docking and AKR1C1 docking, respectively.
The screening of inhibitors allowed to individuate sets of drugs and natural molecules that are specific to the inhibition of AKR1C2 but not AKR1C1. In addition, Random Accelerated Molecular Dynamics (RAMD) [Kokh et al. 2018] can be used to study the residence time of the ligands bound to the enzyme in order to understand variants in AKR1C2 that create an enzyme with increased binding capabilities to its natural substrate. At the same time, the results can be extended on the binding affinity of different natural inhibitors to AKR1C2, for accurate predictions on the molecules that inhibit AKR1C2.
The current data have opened the way to establish a more rigorous correlation between these genes and the pathology. Having said that, it is useful to have a method that screens variants in these genes that is disruptive to the protein's function. Therefore, the inventors developed a criterion to quickly predict disruptive missense variants found in AKR1C1 or AKR1C2 (for the latter, the interest goes to variants that are excluded by the inventors' criterion, since they expect an increased activity of AKR1C2 to lead to lipedema). The criterion is applied on a landscape of these variants derived from their properties as explained in the next paragraph (an example of this landscape for AKR1C2 can be found in Table 17).
The criterion developed by the inventors consisted in the evaluation of the following properties: the position's entropies, the predicted ΔΔG of the variant, and the positions' contribution to the overall substrate binding energy or to catalysis. The variants in AKR1C2 that concern residues with high contribution to the overall substrate binding energy with respect to the natural compounds (natural binding partners) or that contribute to catalysis are favored, and are excluded from the list of the pathogenic variants linked to lipedema (e.g. Leu54, Tyr24, and Trp227 which account for roughly 50% of the total binding energy of AKR1C1 to Progesterone, or Tyr55, Asp50, Lys84, His117 which form the catalytic tetrad); then the entropy (conservation of a position) and the predicted ΔΔG (predicted change of the protein's fold stability when the variant is introduced) are taken into account. If the variant is located in a position with relative entropy >5 OR predicted ΔΔG<−3, it is considered it as potentially disruptive, and subject it to further studies. This way, by including both criteria, the criterion accounts for the importance of the position itself (entropy), and for the changes in physicochemical properties caused by the substitution of one residue with another (predicted ΔΔG).
It must be restated that one of the leading factors to lipedema is hormone imbalance, which is highly sensitive to environmental factors, and genetic factors alone could not lead to the onset of this pathology. Having a selection of disruptive variants in genes that are correlated to lipedema, helps in the early detection of individual's predisposition to lipedema, as well as detect that lipedema is indeed more common than previously thought and could be present in higher frequencies in the general population (considering that it is often misdiagnosed or not diagnosed at all). The inventors ran these predictions on a set of AKR1C1 and AKR1C2 variants in the general population, extracted from the GnomAD database. In Table 16 an example of the disruptive variants in AKR1C1 filtered according to the inventors' criterion is presented, and in Table 17 the whole unfiltered predictions of AKR1C2 variants are presented.
This application is a Continuation-in-Part of application Ser. No. 17/734,708, filed on May 2, 2022, which claims priority under 35 U.S.C. § 119(e) to U.S. Provisional Application No. 63/183,313, filed on May 3, 2021, all of which are hereby expressly incorporated by reference into the present application.
| Number | Date | Country | |
|---|---|---|---|
| 63183313 | May 2021 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 17734708 | May 2022 | US |
| Child | 18516241 | US |
