Biochemistry, Protein Targeting and I Cell Diseases

Fundamentals

Certain receptors recognize specific sequences in newly-produced proteins to bring the proteins to the necessary area. After the protein has been delivered, the receptors will be recycled for further use. Proteins made in the RER are sent to the Golgi through gaps in the RER membrane known as endoplasmic reticulum exit sites (ERES). After quality control at Golgi, misfolded proteins are sent back to the RER to go through a degradation pathway ending at the 26S proteasome. The adequately configured cargo is encircled at the ERES by COPII-coated vesicles. The Sec16 gene maintains the ERES and the COPII vesicular transport. These vesicles are then taken to the RER-Golgi-intermediate compartment (ERGIC) to begin protein targeting.[4]

The Golgi apparatus is a transit for approximately one-third of newly-formed proteins before reaching their final destination.[5] This organelle is stationary in animal cells, whereas it moves at speeds of up to 4 micrometers/sec in plant cells. The Golgi is composed of parallel stacks of maturing cisternae. It is subdivided into cis (the side facing the RER), medial, and trans (the side facing the plasma membrane and opposite the cis side) compartments. The cis cisternae mature into the trans cisternae. The trans cisterna is continuous with the trans-Golgi network (TGN).[6][5] 

Described as a separate organelle, the TGN is responsible for routing proteins to their appropriate destinations.[7] Also known as the trans-Golgi reticulum, the TGN contains enzymes and compounds such as sialyltransferases, P-galactosyltransferase, and sialylated glycoconjugates.[8] Proteins that reach the TGN have their N-linked oligosaccharide chains transformed into sialylated complex carbohydrates due to these enzymes.[9]

Proteins enter the cis-Golgi, pass sequentially through all the cisternae, and leave through the TGN, which is the organelle's sorting component.[10][8] The TGN sends proteins to the plasma membrane, endosomes, or lysosomes. Pakdel and Blume concluded that sorting at the TGN occurs through a CaATPase SPCA1 channel which brings calcium into the lumen to be recognized by Cab45, a calcium-binding protein. Upon calcium binding, oligomerization and binding to specific proteins occur, differentiating cargo that remains in the Golgi from cargo that goes to other cellular locations.[11] Clathrin-coated vesicles from the trans end of the Golgi carry the proteins to the endosomal system, which eventually end up in lysosomes.[8] 

Endosomes sequentially mature into lysosomes by first becoming early endosomes (sorting endosomes) and then late endosomes. The late endosomes become lysosomes by fusing with existing lysosomes.[10] Between the fusion of late endosomes to lysosomes, an endolysosome forms, which gets converted into the lysosome proper. These transformations are accompanied by pH changes due to the activity of the V-ATPase, which pumps protons into the endosomes/lysosomes, acidifying the pH. The pH values gradually decrease from approximately 6.2, 5.5, and 5.0 as the endosomes mature from the early and late endosomes to the lysosome, respectively. The acidic pH uncouples the receptor from its ligand and allows optimal lysosomal enzyme activity. The early endosomes separate molecules that need to be reused by the plasma membranes from the components that need to be targeted to the lysosomes.[12]   

Cargo delivery to the plasma membrane is dependent on GTP requiring ARF (ADP-ribosylation factor), GTPases (guanine triphosphatases), or Rab (Ras-associated binding) GTPases. The activity of either microtubules or actin directs the vesicles to the plasma membrane. Kinesins play a role in this route. The vesicles and granules fuse with the plasma membrane via the octameric exocyst complex composed of Sec3, Sec5, Sec6, Sec8, Sec10, Sec15, Exo70, and Exo84.[13]    

Cellular Level

The process of protein targeting begins with the synthesis of proteins and ends with their delivery. Depending on where the protein is produced and where it needs to be delivered determines the method of transport. Three protein-coated vesicles have been described in intracellular transport: COPI, COPII, and clathrin.[14][15] They were termed 'coat proteins' due to a proteinaceous deposit in the cytoplasmic face on some membrane buds and vesicles.[16] 

Apart from playing roles in vesicle formation, these three proteins decide which protein and lipid cargo are sent in each package.[17] SNARE proteins perform a fusion of membranes. The stepwise fusion events require the correct pairing of SNARE (SNAP receptor) proteins on the donor and acceptor membranes. After creating a trans-SNARE complex by combining SNAREs on the surfaces of the vesicle (v-SNARE) and target site (t-SNARE), the energy barrier posed by the lipid bilayers is overcome, and the trans-SNARE complex is converted to a cis-SNARE complex on the target membrane. 

Membrane fusion is completed by the SNARE complex in a zipper-like fashion by zippering from the distal N-terminal to the proximal C-terminal region. Before the fusion, the vesicle's coat must be slightly restructured, and this may happen reversibly by phosphorylation and ubiquitylation, but this needs further exploration. The cis-SNARE complex binds to the alpha-SNAP (soluble NSF-attachment protein) and the NSF (N-ethylmaleimide-sensitive fusion) protein to create a 20S complex. The ATPase of the NSF breaks down ATP. This breaks up the union of the v-SNARE and t-SNARE, allowing them to be used again for membrane fusion.[18]  

COPI

COPI-coated vesicles participate in retrograde transport (cis-Golgi/ERGIC to RER). It is suggested that they also play a role in anterograde transport from the RER. Additionally, they have also been found in the post-Golgi compartment.[19] COPI-coated vesicles are approximately 75-100 nm in diameter. They are mobile, most likely by way of dynein. An actin pathway leading to COPI assembly has been identified.[20] These vesicles have an inner coat and an outer shell layer composed of seven subunits. The inner coat is formed by beta-COP, gamma-COP, delta-COP, and zeta-COP. The outer shell is composed of alpha-COP, beta'-COP, and epsilon-COP.[14] 

The membrane-anchored and myristoylated ARF1 GTPase brings the COPI components to the Golgi upon activation from ARF-GEFs (guanine nucleotide exchange factors). The loaded COPI vesicles proceed to the RER via signals from the cytoplasmic or endoplasmic located multisubunit DSL1 tethering complex and SNARE proteins (syntaxin-18, Sec20, Sec22B, and Slt1). Long coiled-coil tethers are also involved in tethering. While multisubunit tethering complexes control vesicle targeting to membranes over shorter distances, long coiled-coil tethers target membranes over longer distances.

Examples of multisubunit tethering complexes include DSL1, COG, GARP, HOPS, CORVET, EXOCYST, and TRAPP (transport protein particle). The DSL1 complex is comprised of Dsl1, Tip20, and Dsl3/Sec39. TRAPP I, TRAPP II, and TRAPP III collectively participate in anterograde cargo transport from the RER to Golgi and are also involved in autophagy. Examples of long coiled-coil tethers include p115, Giantin, CASP, GM130, Golgin-97, Golgin-245, GCC185, and EEA1. Giantin, GM130, and p115 are involved in transport from the RER to the Golgi.[18] 

COPI is also uniquely involved in endosomal and lipid droplet functioning, mRNA transport, and the nuclear envelope's breakdown. The final step of COPI vesicle formation is the disassembly of the COPI complex by the activity of ARF GAPs (GTPase activating proteins).[21] ARF GAP2 and ARF GAP3 play more of a role in coat disassembly than ARF GAP1.[20] 

COPI proteins are also involved with intra-Golgi trafficking and maintaining the structural integrity of the mammalian Golgi complex interface. COPI-coated vesicles perform most of the trafficking from the Golgi to the RER. However, it may not be the sole protein involved in this path since delivery of specific cargo like Shiga toxin, Shiga-like toxin, or certain glycosylation enzymes still occurred after blocking the recycling of ERD2 receptors or the VTC (vesicular-tubular clusters) markers.[22] COPI scission occurs at the neck since the membrane curvature is negative. These vesicles are rich in phosphatidylcholines. Liquid lipid cargo such as sphingomyelin and cholesterol are excluded from COPI-coated vesicles due to the ability of membrane lipids to affect the construction of vesicles at various stages.[17] This also explains why the RER has less sphingomyelin and cholesterol when compared to the cis-Golgi.[20]

COPII

COPII-coated vesicles participate in anterograde transport (RER to the Golgi). COPII may also be active beyond the Golgi since COPII proteins, Sec23 and Sec24, have been found on endosomes.[19] COPII complexes comprise an inner layer composed of a Sar1-Sec23-Sec24 combination with the outer cage formed of Sec13-Sec31. Assembly of the COPII complex is initiated by the Sar1 signal and mediated by the Sec12/PREB guanine nucleotide exchange factor.[23] COPII-coated vesicles move in a saltatory fashion on paths delineated by microtubules. After proceeding to the Golgi, these vesicles return to the ERES.[22] 

These vesicles are approximately 60-100 nm in size. In-vitro knockdown studies have shown that the ability of COPII to accommodate large cargo, such as procollagen fibrils composed of a triple-helix matrix extending up to 400 nm, is due to the action of the transmembrane protein TANGO1. Another hypothesis suggests that a portion of the ERGIC is recruited and combined with the ERES, creating tubular uncoated buds capable of carrying large cargo. RAB GTPases, tethering factors, and SNARE proteins mediate vesicular movement and fusion. In mammals, tethering events are coordinated by RAB1, GM130, GRASP65, p115, and the TRAPP1 complex. COPII vesicles are formed by the COPII coat-mediated deformation of the RER.[24] COPII directs misfolded proteins to degradation via the ubiquitin-proteasome system.[25] Sar1, a GTPase, plays a role in COPII-coated vesicle scission.[17]

Clathrin

Clathrin-coated vesicles are essential for lysosome biogenesis, endocytosis, retrograde transport (endosome-to-TGN), and anterograde transport (TGN-to-endosomes).[19][26] Their sizes vary from 50 to 80 nm. The structure of clathrin is described as a 'triskelion' due to the combination of three heavy chains that each associate with a light chain being approximately 190 kDa and 25 kDa, respectively. Having a pH below 7.0 promotes clathrin cage formation.[27]

Clathrin-coated vesicles contain high amounts of MPR (mannose-6-phosphate receptors). The TGN is almost entirely populated by clathrin-coated vesicles, with COPI vesicles also present to a lesser extent. Clathrin adaptor proteins can be either monomeric or tetrameric. Clathrin adaptor protein 1 (AP1), a heterotetrameric protein, functions to form vesicles at the TGN, sort lysosomal hydrolases, and retrieve cargo like SNAREs, MPRs, sortilin, and furin.[10] AP2, which forms at the cell membrane, mediates the clathrin-dependent endocytosis. AP3 may directly transport proteins from the TGN to lysosomes, but the function of AP4 has not been clearly defined.[28] 

The externalized LDL (low-density lipoprotein) receptor takes LDL into the cell via clathrin-mediated endocytosis. Once LDL binds to its receptor, a clathrin-coated pit forms, which moves the LDL receptor-ligand complex intracellularly. Clathrin dissociates from these vesicles and then goes back to the plasma membrane for further use.[29][30] Dynamin, a GTPase, plays a role in the scission of clathrin-coated vesicles.[17]

Molecular Level

Transport of Lysosomal Proteins (Enzymes)

Lysosomal hydrolases are made in the RER and then carried through several organelles before reaching their target.[31] N-acetylglucosaminyl (GlcNAc)-1-phosphotransferase identifies and marks over 70 lysosomal enzymes with the mannose-6-phosphate (M6P) targeting signal for successful transport to the lysosomes.[32][33] GlcNAc-1-phosphotransferase is believed to be located in the cis-Golgi along with N-acetylglucosaminyltransferase.[34] The activity of G1cNAc-1-phosphotransferase decreases from the cis-Golgi to the trans-Golgi.[35] 

G1cNAc-1-phosphotransferase is a hexameric enzyme complex composed of two alpha and two beta subunits encoded by the GNPTAB gene and two gamma subunits encoded by GNPTG. The alpha and beta subunits are synthesized as alpha/beta precursors. The enzyme subunits are assembled in the RER and are then sent to the Golgi apparatus. The alpha/beta precursor is then disassembled into individual alpha and beta subunits by a site-1 protease. The gamma-subunit enhances the M6P post-translational modification activity of G1cNAc-1-phosphotransferase by binding to the alpha-subunit.[32] 

The mannose residue on newly formed hydrolase enzymes acquires a phosphomannosyl targeting signal in the cis-Golgi by transferring G1cNAc phosphate from UDP-N-acetylglucosamine. To expose the M6P in the TGN, the outer phosphodiester-linked N-acetylglucosamine is removed by N-acetylglucosamine-1-phosphodiester alpha-N-acetylglucosaminidase, which is an uncovering enzyme.[36][33] These newly modified enzymes then bind to the MPR, mainly present in the TGN, allowing for transport from endosomes to lysosomes.[8]

There are two types of MPRs named according to their molecular mass. MPR46 (cation dependent/CD) and MPR300 (cation independent/CI) are 46 and 300 kDa, respectively. MPR300 is identical to the human insulin-like growth factor (IGF) II receptor, and therefore it is also referred to as the M6P/IGF II receptor. MPRs is a type 1 integral membrane proteins. Initially, the receptor-enzyme complex is sent to a prelysosomal compartment and then to the lysosome. Upon reaching the lysosome, the acidic intralysosomal pH assists in dissociating the enzymes from the attached MPRs. The detached MPRs are then recycled back to the Golgi for later use.[34] 

Retromer, Rab9, and TIP47 mediate the recycling process.[5] A small fraction of these functional lysosomal enzymes are exported out of the cell but retrieved by MPRs (mainly MPR300) at the cell membrane and brought back into the cell.[32][35] The cell surface MPRs increase when treated with M6P, IGF I, IGF II, and tumor-promoting phorbol esters since they activate protein kinase C. The quantity of the same receptors decreases when treated with weak bases.[37] 

In the study by DiCioccio et al., which focused on pseudo-Hurler polydystrophy and ICD, there were normal intracellular and intralysosomal concentrations of specific lysosomal enzymes. This led to the investigation of alternate trafficking pathways, including the M6P-independent path. Both diseases are due to decreased levels of GlcNAc-1-phosphotransferase. DiCioccio et al. found that most enzymes in ICD and pseudo-Hurler polydystrophy cells, which did not have the M6P signal, were still localized to the lysosome.[38] Cells like hepatocytes, Kupffer cells, leukocytes, and organs such as the liver, kidney, and brain have near-normal levels of lysosomal enzymes in ICD. Acid phosphatase and beta-glucocerebrosidase do not use the M6P-dependent pathway in any cell type.[39] 

Lysosomal integral membrane protein type 2 (LIMP-2/SCARB2) is the MPR-independent trafficking receptor responsible for beta-glucocerebrosidase transport.[40] The pH-dependent cargo (beta-glucocerebrosidase) is bound to LIMP-2 in the RER. The complex travels across the cis-Golgi to the trans-Golgi and is taken to the lysosome, where the acidic pH disconnects the complex.[40] 

LIMP types I, III,10, 12, and 39 also cross the Golgi in an anterograde fashion. The lack of involvement of beta-glucocerebrosidase in the M6P-dependent trafficking pathway accounts for its normal serum levels in ICD. The identification that LIMPs can transport to lysosomes in the presence of tunicamycin, a drug that blocks the traditional transport of hydrolases, further adds evidence that LIMP transport is not coupled with that of soluble hydrolases and to the fact of the existence of an MPR-independent pathway.[41] 

Cathepsin D, alpha-glucosidase, and acid phosphatase are other enzymes using this lesser-used delivery system.[42] Sortilin is another receptor that can bind several targets, such as neurotensin and receptor-associated protein, via an M6P-independent route. It can perform Golgi to lysosome transport of prosaposin, GM2 activator protein, acid sphingomyelinase, cathepsin D, and cathepsin H.[31] Though some are detailed, the exact mechanisms of how each lysosomal enzyme is delivered via the M6P-independent trafficking pathway are yet to be fully explored.[39] 

Protein Targeting and the Signal Recognition Particle

The signal recognition particle (SRP) is a cytoplasmic protein-targeting device that targets the hydrophobic signal sequences of newly forming proteins from ribosomes and brings them to the RER.[43] SRP is composed of a 7S RNA and a signal sequence recognizing protein; therefore, it is deemed a ribonucleoprotein.[44] SRP is regulated by GTPases present within itself and its receptor. After the SRP binds to the signal peptide of the translating protein from the ribosome, a complex cumulatively termed ribosome nascent chain complex (RNC), the RNC then interacts with the SRP receptor on the RER membrane. The binding of SRP with the RNC results in the halting of further peptide elongation.[45] GTP hydrolysis dissociates the SRP-receptor complex after the RNC interacts with the translocation channel. After the translation of the protein is complete, a signal peptidase separates the signal peptide leaving the completely translated protein in the RER lumen.[46]

Protein Targeting to Other Organelles

Macromolecules enter and leave the nucleus through gaps called nuclear pore complexes (NPC).[47] Through the nuclear localization signals, typically characterized by 6 to 10 basic amino acids, proteins are targeted to the nucleus.[48] A group of nuclear transport receptors, known as importins, execute the nuclear targeting of proteins. Importin alpha binds to the cargo in the cytoplasm and links it to importin beta. Through interactions with the NPCs, importin beta translocates the cargo into the nucleus. The receptor-ligand complex is cleaved by RanGTP, releasing the importin alpha cargo. RanGTP then recycles importin alpha by movement through the pores.[47] 

The peroxisomal proteins are trafficked post-translationally from the cytosol since the peroxisome has no DNA.[49] Proteins destined for the peroxisome have a PTS1 (peroxisomal targeting signal) or PTS2 marker. These proteins are then carried by cytoplasmic receptors Pex5 or Pex7 after they recognize the signal. These receptors bind specifically to a serine-lysine-leucine peptide.[50] Catalase is an enzyme that uses the PTS1 receptor Pex5 for peroxisomal transport. Pex14 functions as the translocon on the peroxisomal membrane.[51] 

PTS1 and PTS2 proteins are then carried inside the peroxisomes. The PTS and Pex complex dissociates, leading to the Pex receptor being recycled back to the cytoplasm for later use. To leave the peroxisome, ubiquitin is first added to Pex5 by E3 ligase.[49] This is followed by the export of the receptor by the AAA ATPase complex. Mutations of alanine glyoxylate aminotransferase, an enzyme recognized by Pex5, can lead to primary hyperoxaluria type 1.[52]

Cytoplasmic ribosomes translate mRNA into proteins which then target the mitochondria. These proteins need to have an amino-terminal targeting sequence to target the mitochondria. These protein sequences then allow for the localization of the protein to either the outer membrane, intermembrane space, or across the inner membrane into the mitochondrial matrix.[53] 

A TOM complex, present in the outer membrane, is a translocase responsible for the translocation of proteins into the mitochondria by recognizing the amino-terminal targeting sequence by the TOM receptors (Tom20 and Tom22). The TIM23 complex, present in the inner membrane, handles the import of proteins that localize to and cross the inner membrane.[53] Here the targeting sequences are removed by mitochondrial processing peptidase.[54] 

Tim21 is a subunit of TIM23 that regulates TIM23 and is involved with tethering the TOM and TIM23 complexes. Tim21 is not necessary for the early stages of tethering; therefore, TIM23 complexes without Tim21 function with subunits of the PAM complex instead. Components of the PAM complex include Hsp70 and Pam18. Pam18, through interactions with the Tim17 subunit, links the PAM and TIM23 complexes. Once the protein has entered the mitochondrial matrix, Tim21 replaces the PAM complex. The PAM complex does not completely dissociate from the inner mitochondrial membrane, and it is assumed to tightly binds alternate tight-sites on Tim17 with Tim21 binding in between.[53]

Testing

Prenatal Diagnosis

The autosomal recessive nature of ICD means a higher risk of inheritance among consanguineous families. In a case series performed by Okada et al., eight out of the 18 families had consanguinity. Five families were first cousins, two were second cousins, and one family had grandmothers who were both first cousins.[55][56]  

Manifestations of ICD are recognized at birth, with death occurring within the first decade of life at an average of five years.[57][58] The short lifespan warrants prenatal diagnosis. The decision to undergo the procedures accompanying prenatal diagnosis is associated with their complications. Hence, they are usually performed on a family with a previous history of having babies with ICD.[58] 

Aula et al. was the first group to comment on prenatal diagnosis in ICD. They suggested prenatal diagnosis be made by checking the levels of lysosomal hydrolases in extracellular fluids. They assayed enzymes N-acetyl-beta-glucosaminidase, N-acetyl-beta-galactosaminidase, beta-glucosidase, beta-galactosidase, alpha-mannosidase, and N-aspartyl-beta-glucosaminidase from samples of amniotic fluid, amniotic cells, fetal skin fibroblasts, and placental fibroblasts. When the amniotic fluid levels of these enzymes were compared with the enzyme levels from a normal pregnancy, there were several-fold increases in the ICD-affected pregnancy. They recorded a 5-fold increase in the levels of beta-galactosidase, a 7-fold increase in the levels of beta-glucosidase and alpha-mannosidase, a 6-to-10 fold increase in the levels of N-acetyl-beta-glucosaminidase, and an 8-to-20 fold increase in the levels of N-acetyl-beta-galactosaminidase. However, the enzyme levels of beta-glucosidase were decreased in the amniotic cells and fetal fibroblasts. The other enzymes showed similar changes. The characteristic cytoplasmic inclusions were also present in amniotic cell cultures and fetal fibroblasts. The amniotic cell cultures had a smaller amount of inclusions when compared to the fetal fibroblasts.[59] 

Hexosaminidase is another enzyme that may be elevated in maternal serum. As a diagnostic marker, Hug et al. suggested that mothers with a more than 15-fold increase in the serum levels of hexosaminidase, compared with enzyme levels in normal maternal serum, were likely to bear a child with ICD. They believed the proximity between the maternal serum and the trophoblastic cells played a role in the elevation of the serum level of the enzyme.[60] Uptake of [35S]Sulfate-labeled glycosaminoglycans by cultured amniotic cells is another diagnostic method.[61][62]

Parvathy et al. devised a trimester-based prenatal diagnostic plan. First-trimester prenatal diagnosis can be performed by demonstrating the intracellular deficiency of G1cNAc-1-phosphotransferase from a chorionic villus specimen, with the second-trimester diagnosis being established by looking for the enzyme deficiency within amniotic fluid cells. The trimester-based diagnosis was reinforced by looking for the intracellular lack and extracellular elevation of several lysosomal enzymes in fetal cell cultures and enzyme elevations in maternal serum. Confirmatory tests should be performed when the initial testing is indefinite or when the sample is insufficient. Prenatal diagnosis based on enzyme elevation can be similarly applied to Hunter syndrome and Tay-Sachs disease.[63]

Owada et al. suggest that relying on enzyme elevations, detection of inclusion bodies, and measuring uptake of radioactive material may not be as reliable or realistic to perform due to their time-consuming natures. Elevations of lysosomal enzymes in normal amniotic fluid may be present depending on the gestational age. Their study emphasizes the value of finding altered characteristics of enzymes rather than altered levels of enzymes. Concerning alpha-mannosidase, its Km in high-risk pregnancies had values ranging from 1.20 to 1.46 x 10^-3 mol/l, whereas the value in non-high-risk pregnancies had values of 2.63 x 10^-3 mol/l and 3.03 x 10^-3 mol/l. The alpha-mannosidase from high-risk pregnancies also had a relative resistance to heat as two of their samples functioned normally while being heated to 56 degrees Celsius for 20 minutes. Further, the enzyme showed an optimal pH of 3.5 compared with the control's pH of 5.[61]

Examination of the placenta may also provide clues pointing to ICD. The macroscopic examination may not show distinguishing features, but the syncytiotrophoblast layer will show characteristic inclusions. The cytotrophoblast layer may be normal.[64][65] The chorionic villi may also be swollen and will show more variations in size and shape when compared to that of normal placentas. The gross vacuolization of the placental cells may disturb the maternal-fetal interface leading to impaired transfer of nutrients and gases.[64] The vacuolization may also be responsible for the disturbed transplacental transfer of calcium, leading to the bony abnormalities characterizing ICD.[65]

Pregnancy may be complicated by polyhydramnios or end poorly with stillbirth and infant death.[58] Babies may have meconium aspiration syndrome.[56] Imaging modalities such as ultrasound may establish the diagnosis in a pregnant woman with a previous history of delivering a fetus with ICD.[58] Findings of a short femur, bowing of the femur, and periosteal cloaking of the femur and humerus suggest ICD.[66][58] 

With such non-invasive and safer methods, patients may not need invasive prenatal testing. However, these ultrasound findings are not pathognomic as several other aneuploidies, such as Turner syndrome and Down syndrome, may also show the same features. Achondroplasia, Russell-Silver syndrome, and Shprintzen syndrome also add to the differentials based on imaging findings alone.[66] 

In a mother without a previous history of having a fetus with an ICD, ultrasound identification of such features cannot guarantee a diagnosis. Further adding to the uncertainty behind this method, clinical suspicion of ICD should warrant additional ultrasound examinations as the routine 20-week ultrasound assessment may not identify these features.[58] 

In the report by Lees et al., their patient had a routine 20-week ultrasound, with the 30-week ultrasound showing femurs below the 5th percentile and an increased echogenicity surrounding the periosteum of the humerus and femur. They decided to perform a 30-week ultrasound due to the mother's previous history of delivering a baby with ICD and because the patient developed polyhydramnios in the presenting pregnancy.[58] 

Yuksel et al. suggest that a prenatal diagnosis of ICD should be considered even in mothers without a family history of ICD. Chromosomal analysis may help rule out other differentials. They ascertain a definite diagnosis by measuring the G1cNAc-1-phosphotransferase activity in chorionic villi or cultured amniocytes.[66] 

Diagnosis

History of a mother having a baby with ICD in the past or having a baby with dysmorphic features matching those of ICD may warrant subsequent investigation. The diagnosis is usually made between 6 months to 12 months of age.[67] Diagnosis can be established by measuring lysosomal enzyme activity in the plasma, dried blood, or cultured fibroblasts. This should be followed up with a genetic analysis of the GNPTAB and GNPTG genes to differentiate subtypes of ML (mucolipidosis) III from ML II, which is an ICD. An additional test that can be done is a Western blot test using a single-chain antibody fragment that detects M6P-containing proteins from the patient's fibroblasts.[32] Lymphocytes from a peripheral blood smear will show vacuolization.[33] 

The clinical and lab reporting similarities between ICD and rickets often lead to late recognition. The similarities may be attributable to the exact chromosome location (chromosome 12) of the genes encoding the GNPTAB and the vitamin D receptor. The significance of the proximity has to be explored further. Though similar PTH (parathyroid hormone) and calcidiol levels can be seen in both ICD and rickets, an exaggerated decrease in calcidiol and increased PTH point more towards the latter.[68]

Pathophysiology

In 1972, Hickman and Neufeld [69] first provided a hypothesis for the pathophysiology of ICD. They used culture mediums to identify that I-cells could endocytose acid hydrolases put out by normal fibroblasts, but normal fibroblasts could not endocytose enzymes from I-cells. This finding led them to concentrate their theory not on the defective membranes that allowed enzymes to escape but on defective lysosomal enzymes that failed to reach their target destination.[69] Dysfunctional protein targeting may lead to the accumulation of proteins and enzymes in abnormal locations. ICD is a disorder of protein targeting. ICD, also known as Leroy I-cell disease, is an autosomal recessive lysosomal storage disorder due to a mutation of G1cNAc-1-phosphotransferase.

The term 'mucolipidosis' covers diseases belonging to both mucopolysaccharidoses and sphingolipidoses.[66] ICD has features of both disease categories and is classified as such. There are four types of mucolipidoses: type I (sialidosis), type II (ICD), type III (pseudo-Hurler polydystrophy), and type IV (mucolipidosis). All of these diseases are classified according to the deficient or mutated enzyme.[33] Due to sharing similar clinical features, differentiating Hurler syndrome from ICD is a difficult task. A unique clinical feature exclusive to Hurler syndrome is a temporary acceleration of skeletal growth at one year of age.[70] Leroy et al. first distinguished between mucopolysaccharidoses and ICD by identifying the cytoplasmic inclusions and variations in beta-glucuronidase and acid phosphatase levels. 

Of the two genes encoding the functional components of G1cNAc-1-phosphotransferase, GNPTAB has been connected with the mutations causing ICD. Mutations like frameshift, nonsense, splicing defects, missense, and deletions/duplications/insertions lead to the production of a stop codon 80% of the time. The stop codon forms an incomplete and dysfunctional enzyme or may lead to truncated enzyme subunits.[57] A majority of the mutations are frameshift mutations (39%). Kudo et al. found the specific mutations of GNPTAB to be of FS211X (type 1), FS288X, FS546X, FS588X, FS737X, FS1081X, and FS1172X. In particular, FS1085X (type 1) had a frameshift mutation due to inappropriate splicing, removing exon 17. K4Q and S15Y mutations in the alpha-subunit can lead to decreased G1cNAc-1-phosphotransferase activity in the Golgi complex.[71]

Although the mutation variants are evenly spread throughout GNPTAB, approximately 25% of the mutations are in the 1112-bp exon 13. Clinical severity correlates with the level of enzyme activity. Absent or severely reduced GlcNAc-1-phosphotransferase activity due to strong alterations of GNPTAB leads to the severe ICD phenotype. Having residual enzyme activity (approximately 10%) due to at least one GNPTAB allele being active is associated with the less severe ML III alpha/beta disease.[32][56] 

Patients with ML III can survive into their adulthood, whereas patients with ML II often die within the first decade of life. ML III is divided into ML IIIA (pseudo-Hurler polydystrophy) and ML IIIC (variant pseudo-Hurler polydystrophy). In ML IIIA, the G1cNAc-1-phosphotransferase activity is reduced. In ML IIIC, G1cNAc-1-phosphate transfer to alpha-methylmannoside is normal, but the transfer to lysosomal substrates is reduced.[57] Measuring G1cNAc-1-phosphotransferase levels can help differentiate ML II from other mucopolysaccharidoses.

The lysosomal enzymes, lacking the M6P signal, cannot be sensed by the receptors in the TGN, so further targeting of the endosome/lysosome system cannot be carried out.[72] The progression of protein targeting is due to the cargo-loaded and ligand-bound carrier vesicles. The receptors remain to wait for the signaling ligand to continue the trafficking process. Since the ligand never arrives, the unbound receptors are carried by a group of vesicles, different from the carrier vesicles, to undergo exocytosis.[34]  

Braulke et al. found that I-cell fibroblasts have twice as many MPRs and a higher affinity for M6P and IGF II than normal fibroblasts. The increase in receptors is due to increased mRNA expression of MPRs in I-cells. The receptor count in I-cells remained relatively stable under the influence of substances that increased the surface receptor count.[37] The MPRs are found mainly in the Golgi apparatus with a reduced or absent concentration found on the path to and in the endosomes/lysosomes.[34] 

In ICD, lysosomal enzymes like glycosidases and sulfatases, among several others, are improperly sorted and targeted, leading to their extracellular accumulation. Lysosomes become dysfunctional and enlarged due to nondegradable macromolecules such as cholesterol, phospholipids, and glycosaminoglycans. This disrupts cellular functioning and homeostasis.[33] Exogenous hydrolase uptake by lysosomes may also be affected due to altered M6P carbohydrate chain or residue recognition markers on the hydrolase molecule.[67][36] 

The inclusions of ICD have been described extensively from microscopic studies of cultured fibroblasts.[73] Two types of cytoplasmic inclusions have been described: Type 1 and Type 2. Type 1 inclusions are predominantly clear, membrane-bound vesicles of 0.5 to 2.0 micrometers in diameter. Type 2 inclusions contain abundant polymorphic material and concentric globules, membrane-bound and similar in size.[59] 

Both types of inclusions are often found within the same cell. In the study by Aula et al., fibroblasts, primitive mesenchymal cells like macrophages, and endothelial and perithelial cells contained the inclusions, whereas epithelial and glandular cells did not. A peculiarity was found by identifying numerous type 1 inclusions and some type 2 inclusions in the glomerular podocytes of the kidney, which are differentiated epithelial cells.[59] In the 2019 study by Yokoi et al., three patients with ICD only had inclusions in B cells without any inclusions present in CD4 T-cells, CD8 T-cells, natural killer cells, monocytes, or neutrophils. The B cell inclusions contained HLA class II molecules. This finding suggests that G1cNAc-1-phosphotransferase has an immune function.[74] 

Apoptosis

Apoptosis is programmed cell death that can occur via two different pathways (mitochondrial and death receptor pathways). Lysosomes play a role in the progression of apoptosis, particularly with the release of proteases such as cathepsin D.[75] As a result of the misdirection of various lysosomal enzymes in ICD, apoptosis still occurs but is delayed. This is related to the lower activities of cathepsins D, B, and L in I-cells compared to normal fibroblasts, as found by Terman et al. The same group performed a series of experiments that assessed I-cells' ability to undergo apoptosis compared to normal fibroblasts. Their group found that apoptotic inducers such as MSDH, staurosporine, and naphthazarin had a decreased effect on I-cells. The self-destruction process was reduced when using apoptotic inhibitors leupeptin and pepstatin A on staurosporine-treated normal fibroblasts. No significant change was noted when the same experiment was performed on I-cells.[72]

Clinical Significance

Epidemiology

A high concentration of cases is present in the Saguenay Lac-Saint-Jean region of Quebec, Canada, with a prevalence of 1/6,000 live births.[32] The introduction of disease in this region may date back to the 17th century due to the migration of immigrants from France and Scotland.[32] 

Worldwide, however, the prevalence is estimated to be between 1/100,000 to 400,000.[55] The low prevalence accounts for gaps in the natural clinical history and management considerations for patients with ICD.[55] Inconsistencies of some clinical features may be attributed to the limited cases discussed in the literature.    

Clinical Features

ICD is characterized by a multisystemic involvement with prominent bony abnormalities. Table 1 shows the various manifestations and clinical features of ICD. The presence of corneal clouding is variable. When ICD was first considered in 1967 by Leroy et al., two patients did not have corneal clouding or excessive urinary secretion of acid mucopolysaccharides. The absence of these features helped distinguish Hurler syndrome from ICD. Other studies report that the corneas may be clear.[76] In their case series, Okada et al. found that 19 out of the 21 patients had clear corneas.

Patients with ICD often have a developmental delay, of which motor delay is more prominent than cognitive delay.[33] Paton et al. developed a novel mouse model to explore symptomatology in humans fully. This process found that patients with ICD can have psychomotor retardation and an ataxic gait because of progressive neurodegeneration in the cerebellum with Purkinje cell loss.[77] Due to a degree of psychomotor retardation, milestones such as self-feeding and toilet training are usually not attained.[76] 

In contrast, Okada et al. reported several cases to have exceptional levels of cognitive development where children were independent, toilet-trained, had good academic performance, were able to sing songs, and recognized foreign words. Autopsy studies on patients with ICD have not revealed any brain abnormalities.[76][78] This finding is further supported by studies on knock-in mice which have demonstrated that active Niemann-Pick C2 protein maintains cellular functioning and prevents neurodegeneration in ICD. Cathepsin D and B also remain unaffected in ICD since they use M6P-independent pathways to transport neuronal lysosomes.[33]

Taber et al. suggest that hepatomegaly is a feature mainly identified with Hurler syndrome. However, other reports have identified liver enlargement in patients with ICD.[68][78]

Table 1:Clinical Features of ICD 

Bony and Radiological Abnormalities

Dysostosis multiplex is the broad term that encompasses the skeletal abnormalities of ICD. General x-ray findings are demineralization, osteopenia, coarse trabecular bones, and periosteal cloaking (predominantly long bones).[65][76] X-rays of the upper limbs may show metacarpal pointing and bullet-shaped phalanges. Pelvic radiographs may show iliac flaring. A J-shaped sella turcica may also be visualized.[73]

Bone dysplasia with shortened and curved bones is considered a prenatal manifestation of ICD. Deformities of the vertebral bodies may show anterior beaking, leading to abnormal spinal curvatures.[33] The cervical spine vertebral bodies are normal, whereas the dorsal and lumbar spine vertebral bodies have a reduced anteroposterior diameter. In ICD, the bones of the upper limb are more affected than the bones of the lower limb. This is exemplified by the hypoplastic carpal bones and normal tarsal bones.[76] 

The cellular causes underlying the bony features of ICD may be due to the excessive formation of normal osteoclasts (osteoclastogenesis) and subsequent destruction of bone along with the subfunctional activity of osteoblasts. The increased osteoclasts are due to the osteoblasts producing the osteoclastogenic cytokine IL-6.[82] Periosteal cloaking, a finding of increased new bone formation, disappears by about ten months of age when the bony overgrowth merges with the cortex underneath. The skeletal deformities may be due to deviations from normal calcium metabolism. PTH and PTHrP (parathyroid hormone-related peptide) regulate fetal calcium homeostasis. Unger et al. presented three cases of ICD with prenatal onset of hyperparathyroidism.[65]

These patients had normal extracellular calcium levels with an elevated PTH, indicating that the parathyroid glands compensated to overcome the prenatal calcium deficiency.[65] This means that the fetal mechanism for calcium regulation was intact since the low calcium, detected by the CASR (calcium-sensing receptor), was met with an appropriate elevation of PTH. The initial decrease in calcium may be related to the impaired transplacental transfer of calcium due to the extensive inclusions.[65]

Using a mouse model, Unger et al. suggested that a receptor other than the CASR is responsible for recognizing and coupling the signal of fetal PTHrP to the transplacental transfer of calcium.[65] The transplacental transport cannot occur due to extensive vacuolization of the syncytiotrophoblast (the layer actively regulating transplacental transport). This vacuolization interferes with the action of PTHrP, thereby causing the PTH to break down bones to maintain calcium homeostasis. In 1993, Shohat et al. presented a case of skeletal dysplasia having periosteal cloaking with histopathological findings of numerous osteoclasts with many marrow regions lacking trabecular bone.[83] They later named this condition "Pacman dysplasia." These researchers later recognized Pacman dysplasia, not as a separate entity but as ICD with secondary hyperparathyroidism due to the radiological features aligning with ICD and the histopathological features being consistent with hyperparathyroidism.[65]

Differential Diagnosis

Mucopolysaccharidoses, ML III, rickets, and osteogenesis imperfecta are other conditions to consider before diagnosing a patient with ICD.[33][68][65] ICD can be differentiated from other conditions with the utilization of clinical history and examination, basic metabolic panel, and urinary tests for glycosaminoglycans.    

Treatment

Treatment for ICD is mainly supportive and symptomatic, as there is no definitive treatment.[33] Further reporting is required to fully elucidate which management strategies are safe and beneficial to use in ICD. 

Medical Management

Bisphosphonate therapy has been used in those with profound bony involvement and decreased bone mineral density (z score <-2.5). Bisphosphonates work by reducing the activity of the numerous osteoclasts in ICD, reducing bone resorption. Bone marrow transplants affect both skeletal growth and cognitive development. A positive association between the duration of enzyme replacement therapy (ERT) and skeletal growth was seen in ICD. However, the benefits of ERT are limited to the viscera and are restricted in their effects on the bones and brain. The real challenge with ERT is creating multiple recombinant lysosomal enzymes already having the M6P residues.[33] 

Vitamin D supplementation may help with bone manifestations as well. Pazzaglia et al. reported two cases of ICD with radiological images of metabolic bone disease.[84] The case treated with vitamin D supplementation showed quicker resolution of the radiological features than the case that was not treated.[84] Vitamin D supplementation will also help normalize the lab profiles in those with coinciding ICD and vitamin D deficiency.[68] 

Genistein is a tyrosine kinase inhibitor of epidermal growth factor that reduces glycosaminoglycans like heparan sulfate. Genistein usage at 5 mg/kg/day showed an improved range of motion of joints and increased elasticity. Genistein's estrogen-like effect may play a role in its benefits to bone health.[85] The drawback with genistein is decreased cellular growth, hypothesized to be due to its interference with cellular growth factors.[33]

Surgical Management

In cases of acute airway obstruction, using a video laryngoscope may be more helpful to secure the airway, as laryngeal mask airways only provide temporal placement with questionable success.[55] Hip and knee replacements can be performed to reduce hip and knee pain, respectively. Valve replacement can be completed in cases of valvular dysfunction.[33] Performing synovectomy and resection of the palmar carpal ligament will help relieve pressure on the median nerves.[81]

Other Treatment Methods

Matos et al. devised a novel therapeutic strategy to reduce G1cNAc-1-phosphotransferase activity instead of absent enzyme activity by skipping exon 19 on GNPTAB.[86] Low-impact aqua therapy can be used to reduce stress on joints and tendons.[33]

Other Disorders of Protein Targeting

Based on previous literature, protein targeting has been implicated in neurodegeneration, several infections, and other metabolic disorders. LIMP-2 is concerned with the pathogenesis of hand, foot, and mouth disease by acting as a cell membrane receptor for entering enterovirus 71 and coxsackievirus 16.[40] LIMP-2 is also involved with transporting beta-glucocerebrosidase, the enzyme deficient in Gaucher disease, via an MPR-independent pathway.[5] 

Cholera toxin, Pseudomonas exotoxin, Shiga toxin, and ricin use the retrograde endosome-to-TGN pathway to avoid lysosomal degradation. These AB-type toxins use the B fragment to bind to the cell surface, allow the active A component to enter the cell, and use the endosome-to-TGN pathway. Manganese (specific to Shiga toxin) and Retro-1 or 2 (small molecular compounds) are protective against these toxins, as seen in cultured cells and animals.[10] Nef, a virulence factor for HIV-1, disrupts the trafficking of MHC-I, thereby preventing the destruction of virion-infected cells by cytotoxic T-lymphocytes.[5]

Clathrin-mediated endocytosis and COPI play roles in the infectivity of the Dengue virus. The acidic pH of the endosomes creates an optimal environment for integrating the virus with the endosomal membrane, leading to the release of the viral genetic material into the cytoplasm for further viral growth.[15]   

Neurons are susceptible to trafficking defects due to the long distances the cargo must travel and the requirement of exact localization. A unique feature of neurons is that membrane and secretory neuronal proteins often bypass the Golgi and go to the plasma membrane. Whether this feature increases the susceptibility for targeting defects remains to be known. Neuronal targeting defects cause diseases like epilepsy, hyperekplexia, and ataxia. Alpha-Synuclein, the predominant component of Lewy bodies in Parkinson's disease, is responsible for inhibiting RER-to-Golgi trafficking. This leads to its toxic cellular accumulation. Increases in RAB1, a modifier of alpha-Synuclein activity, have reduced neuron loss in animal models of Parkinson's disease. Patients with developmental disorders may have mutations of TRAPPC6A, TRAPPC6B, and TRAPPC9. Mutations of TRAPPC2L and TRAPPC12 can cause encephalopathy.[87]

Neurofibrillary tangles and amyloid plaques characterize Alzheimer's disease. The main component of these plaques is beta-amyloid. Beta-amyloid is formed by the beta-cleavage of amyloid precursor protein (APP) by beta-site APP-cleaving enzyme (BACE).[88] FRET (fluorescence resonance energy transfer) studies have revealed that APP and BACE, a secretase, localize mainly in the early endosome, meaning that the early endosome is the area of cleavage.[88][89] 

Other studies have shown that blocking proteolysis with protease inhibitors or removing gamma-secretase using a presenilin knock-out leads to APP collection within the lysosomes.[89] The studies done by Tam et al. showed that both gamma-secretase and beta-secretase work at an acidic pH in the lysosomes.[89] This explains a possible therapeutic benefit of using chloroquine, an antimalarial, in Alzheimer's disease. Chloroquine increases the pH in endosomes and lysosomes, disrupting APP's cleavage and inhibiting beta-amyloid formation. This leads to the accumulation of APP in lysosomes. Contrarily, with the Swedish mutation (APPsw), the rate of beta-cleavage is increased so much that marked fluorescent APP cannot be visualized in the lysosomes. Beta-amyloid plaques can disrupt synapses and membranes, leading to lysosomal rupture and causing cell death.[89] 

Sortilin is one of many trafficking proteins. One of its functions is to act as a sorting receptor for PCSK9 (proprotein convertase subtilisin/Kexin type 9) in lipid metabolism. PCSK9 degrades LDL receptors and is inhibited by medications like alirocumab and evolocumab to achieve cholesterol control. The sortilin-related receptor with A-type repeats (SorLA) is associated with Alzheimer's disease. Murine models lacking SorLA will have increased levels of beta-amyloid. Overexpression of SorLA can lead to the accumulation of APP in the Golgi, impeding its further cleavage.[5]

A mutation of SEC31A, a component of COPII, can lead to intrauterine growth retardation, developmental delay, and seizures. A receptor molecule involved with COPII functioning is cTAGE5. A proline-to-alanine replacement in cTAGE5, showing a variant known as P521A, is a risk factor for developing Fahr disease.[87] Combined factor V and VIII deficiency, which are mannose-containing factors, is thought to be due to a sorting defect in the RER instead of any coagulation factor abnormality.[19]