Molecular Biology of Retinoids

12 May

INTRODUCTION

Acne  vulgaris is a multifactorial disease of the  skin  in   areas   rich   in   sebaceous follicles.  It is characterized by seborrhea, hypercornification of the  infundibulum (the neck of the sebaceous gland), and  the presence of comedones and  inflamma- tory lesions  such as papules and  pustules. The inflammatory pathway is mediated by antigenic and  inflammatory products of Propionibacterium acnes (1). Although acne  is  mostly   associated  with   puberty,  persistent or  late  onset   acne  may  be similar in  physiology to  pubertal acne,  with  hyperandrogenicity and  increased sebogenesis, being  the key factors  (2). The hyperproliferation of the infundibulum keratinocytes, characterized  by  the  expression of the  hyperproliferative marker proteins, Ki67 and  K6/16, and  the resulting immature stratum corneum does  not desquamate efficiently,  leading to clumps of squames that attach  to the hair follicle causing a blockage  to sebum flow. Hypercornification of the infrainfundibulum  is an early feature of the microcomedones, preceding the clinically observable inflam- matory process  (3).

Hypovitaminosis  A  has  been  proposed as  an  inducer of  sebogenesis, as retinoids suppress sebogenesis. Follicles with  a high  sebum output are thought to become  deficient in  vitamin A, which  will  disturb the  keratinization process  of the infundibulum of the gland, culminating in the initial  formation of microcome- dones in these  glands (4).

Recently, comedogenesis has been related to a subclinical microinflammatory state  in the sebaceous gland  (3). The increased interleukin-1 levels  may  influence the  production of other  growth factors  by  fibroblasts that  then  cause  the  ductal hyperproliferation and  faulty  differentiation (5).   Increased  transcription   factor signaling occurs  following binding of the  cytokines to  their  receptors, activator protein-1 (AP-1)  and  nuclear factor  kappa-b, thereby   activating  the   signaling pathways. As a result,  suppression of these  signaling pathways will be beneficial for  acne.  Retinoids can  improve epidermal differentiation, reduce sebogenesis, and  decrease AP-1 driven inflammatory pathways. It is not  too surprising, there- fore,  that  this  class  of agents improve the  acne  status. This  chapter reviews the molecular biology  and  biochemistry of this important class of molecules.

NUCLEAR HORMONE RECEPTORS AND THE RETINOID RECEPTORS

In the last several  years,  the structure and  function of nuclear receptors have  been determined (6). Based on the similarities among the sequences, it has been possible to classify  the receptor superfamilies. The endocrine receptors include the retinoic acid receptors (RARs) and  the thyroid hormone receptors, and the adopted orphan

FIGURE 1    Classification of nuclear  and  orphan receptors.  Abbreviations: LXR, liver-X-activated receptor;  PPAR,   peroxisomal  proliferators   activated  receptor;  RXR,  retinoid  X  receptor;  TR, thyroid hormone receptors, 9cRA, 9-cis-retinoic  acid.

receptors include the  retinoid X receptors (RXRs). These  retinoid receptors are active  as heterodimers (typical  receptors are illustrated in Fig. 1).

The two  classes  of nuclear  receptors  mediate  the   action   of  endogenous retinoids (7).  Both  types   of  receptors are  composed of  three   subtypes  (alpha, beta, and  gamma) (8). Each subtype consists  of six distinct domains referred to as A to F based  on homology with other members of the nuclear receptor superfamily (Fig. 2). Domains A and  B are at the amino  terminal and  contain isoform-specific ligand-independent  transactivation function-1 (AF-1). The DNA-binding domain (C or DBD) is highly  conserved and  contains two  zinc-binding motifs  responsible for the  recognition of the  retinoic  acid  response element (RARE), located   in   the promoter region   of  target   genes.  The  E  domain is  the  ligand-binding domain (LBD) and  is responsible for the  dimerization of the  receptors, ligand-dependent transcriptional AF-2, and  translocation to the nucleus (Fig. 2) (9).

The DNA response element of nuclear receptors comprises of two hexameric motifs  with  the  nucleotide structure AGGTCA.  The organization of these  repeat palindromes and  the length  of the nucleotide spacing between the two  hexamers determine the  binding specificity   of  a  particular  receptor. The  simplest is  the direct   repeat-1 of  the  RXR – RXR homodimer, RARE.  However, the  polarity  of binding to the  response element is reversed in these   receptors  with   the   RXRs occupying the 50  half-site  when coupled to the RAR (9).

The pleotropic effects of retinoids are  due  to the  existence  of multiple RAR isoforms and  as a result  of the different combinations of RAR – RXR heterodimers. RARs and  RXRs mainly act as heterodimers on binding to the RAREs. The RARs can  be  activated  by  binding  all-trans-retinoic acid   (all-trans-RA)   or  9-cis-RA; however, of the  different retinoids, RXRs can  only  be   activated   by   9-cis-RA. The RXRs predominate in human skin,  especially RXR alpha. Of the  RARs, 87%

FIGURE 2    Schematic  overview   of  the   retinoic   acid   receptor/retinoid  X  receptor,  receptor dimerization, and  binding to the  retinoic  acid  response element in the  promoter  region  of a gene. Abbreviations: AF, activation  function;  DBD, DNA-binding  domain;  LBD, ligand  binding  domain; RA, retinoic acid; RAR, retinoic acid receptor RXR, retinoid X receptor.

is gamma and  13% alpha. Only  small  amounts of RAR beta  are found in dermal cells   and   melanocytes.  Human   sebocytes  in   vitro   also   express  mRNA    or RAR  alpha and  gamma together with  RXR alpha. Although both  all-trans-RA and  9-cis-RA bind  RAR in  vivo,  RAR gamma preferentially binds  all-trans-RA. Both 9-cis-RA and  13-cis-RA can be isomerized to all-trans-RA.  The isomerization of 9-cis-RA has been reported to occur  in keratinocytes (8).

The  ligand-binding pockets of nuclear receptors have  been  determined by X ray  crystallography (10). This domain consists  of a series  of alpha helices  that give  rise to a novel  antiparallel alpha-helical sandwich. Side chains  of the  amino acid,  which  constitute this  structure, also  include the  homo-  and  heterodimeric interfaces and  the  surfaces that  provide binding sites  for the  nuclear corepressor and  coactivator molecules (Fig. 2). Binding  of ligands to the LBD induces confor- mational changes in the receptors (11). Although this is true  of all the ligand acti- vated nuclear receptors, it is particularly true  of the  RXRs. The  ligand-binding pocket  is a deep  hydrophobic pocket  within the LBD, and  it is different in its archi- tecture, which  plays a significant role in contributing to the different ligand-binding properties. For  instance, in  the  peroxisomal proliferators activated receptor, the ligand-binding pocket  is relatively large  (1200 – 1500 nm3)  and  when ligands bind to  these  receptors they  occupy   only  a  small   fraction   of  the  available volume usually 15% to 25%, the  rest  being  occupied by water molecules (12). The  large

binding pocket  and the limited number of ligand-receptor contact  points result  in a relatively low  affinity  of binding of ligands to these  receptors and  because  of the large  pocket,  the receptor is promiscuous and  binds  a diverse range  of molecules. In  contrast to  these  low  affinity  receptors, the  other   nuclear receptors have  a much   higher affinity  for  their  ligands. In  general, the  ligand-binding pocket  is much  smaller (400 – 500 nm3),  and  the ligand occupies a high  fraction  of the avail- able volume; for example, 9-cis-RA occupies 75% of the  ligand-binding pocket  of RXR alpha while  all-trans-RA  occupies 60% (13). The  ligand-binding pocket  of the  RARs  can  accommodate both  the  all-trans-RA  and  9-cis-RA, but  due  to  the

bulkier  side   chains   of  the   aminoacid  residues  of  the   RXRs,  it  only   permits binding of 9-cis-RA.

One  important region  of the  LBD is the  dimerization interface (14). For the RARs, symmetric assembly of the  dimerization interface to form  homodimers is energetically unfavorable. Asymmetric interactions between the  interfaces of the RXR give rise to an extended area  of intermolecular contact  that  stabilizes hetero- dimer formation. The RXR is unique among the nuclear receptors, having a dimer- ization  interface that  is stable  in  both  the  symmetrical configuration (RXR/RXR homodimers) and  the  asymmetrical confirmation (RXR/RAR heterodimers). The RXR homodimer (but not the heterodimer) generates a new dimerization interface that allows  tetramer formation, and  a large fraction  of the unliganded RXRs is also found in this formation. Binding  of ligands to the apo-RXR tetramer induces confor- mational changes that  destabilizes the  tetramer conformation to form  homo-    or heterodimers. In  the  tetramer form,  however, even  though the  homodimers  of RXR do  not  bind  all-trans-RA,  the  distorted ligand-binding pocket  binds  it but does  not  induce dissociation of the  tetramer format.  In fact, all-trans-RA  can  act as a competitive antagonist of ligand-induced RXR tetramer dissociation. In this tet- ramer state,  the RXRs cannot  bind  corepressor or coactivators and  are sequestered in a transcriptionally inactive  pool  called  “auto-silencing.” Thus,  one of the ways that RXR ligands can activate transcription is by increasing the pool of RXRs avail- able for heterodimerization with  RARs (10).

Liganded RXRs preferentially form  homo-  or  heterodimers. The  interface between the  receptors in heterodimers such  as RXR/RAR  is substantially larger than  the dimer interface (550 vs. 500 nm3),  resulting in a preferential formation of hetero-  versus homodimers. Binding  of 9-cis-RA to the RXR also results in confor- mational changes that generate an agonist confirmation, dependent on the binding of  coactivator  molecules.  The  initiation  of  a  “mouse  trap-like”  configuration induces structural changes in the protein in such  a way  that  it not only closes the lid of the ligand-binding pocket,  but  also simultaneously generates a high  affinity coactivator-binding site. In the presence of partial agonists like oleic acid, the struc- tural  changes cause  transrepression rather than  activation (10 – 14).

RXR ligands can alter  RXR activity  both  by altering the availability of RXRs for heterodimer formation and  by altering the  intrinsic transactivation activity  of heterodimeric complexes and receptor degradation pathways. Ligands can acceler- ate  the  degradation  of  the  receptor, particularly  through  the  26S  proteasome complex. Binding  of ligands to  the  different isoforms could  induce degradation of both  receptors or  the  selective  degradation of only  one  receptor. It  can  also depend on the subisoform in each receptor class.

Like  other   nuclear receptors, retinoid-induced  transactivation  of  genes  is mediated as a result  of alterations in chromatin packing and  structure (15). The chromatin is tightly  packed inside  the  nucleus and  is a  complex, consisting  of

FIGURE 3    Schematic overview of coactivator/corepressor and  histone deacetylase proteins  and their  functions   in  chromatin   condensation  and   gene activation.   Abbreviations:  HDAC,  histone deacetylase;  NcoR,   nuclear  receptor  corepressor;  SMRT,  silencing   mediator  for  retinoid  and thyroid; SRC,  steroid  receptor coactivator.

DNA, histones, and  nonhistone proteins. A nucleosome is the basic building block of chromatin, which  contains 147 base pairs  of DNA wrapped around a core of four histone partners—an H3 – H4 tetramer and  2 H2A – H2B dimers. When  condensed, this structure represses gene transcription. In the absence  of the ligand, the nuclear receptors recruit nuclear corepressor proteins, nuclear receptor corepressor or the silencing mediator for  retinoid and  thyroid hormone, and  Sin 3, which  in  turn forms  a complex with  histone deacetylase enzymes (HDAC)  resulting in transcri- ptional silencing of the genes (Fig. 3). This suppression occurs  because deacylation of the histone proteins creates  conformational changes in the chromatin structure, limiting the  access  and  binding of the  nuclear receptors and  RNA  polymerase to the related genes. At physiological concentrations of RA (1029 –1028 M), the nuclear cor- eppressors and HDAC are dissociated, which  in turn  results in recruitment of coactiva- tors  with  histone acetyltransferase activity  such  as the  steroid receptor coactivator-1. Acetylation of lysine  residues in the  N-terminal of histones opens  up  the  chromatin structure and allows gene transcription (16). Enhancement of retinoid activity  is antici- pated by combining HDAC  inhibitors with  retinoids (see later).

Random Posts

Comments are closed.