2.1. Genomic Organization, Structure and Natural Forms of TFF Peptides
In humans, all three TFF genes are clustered on chromosome 21q22.3 in a head-to-tail arrangement in the order telomer-TFF1-TFF2-TFF3-centromer (review:
[24]). In the mouse, a syntenic region is located on chromosome 17q. Of note, all TFF domains are encoded by separate exons, as is typical of shuffled modules.
As a hallmark, the TFF domain contains six conserved cysteine residues, creating a disulfide-linked, three-looped structure (
Figure 1)
[3][5]. TFF1 and TFF3 show high similarity, as both consist of a TFF domain and an additional seventh cysteine residue near the C-terminal (Cys
VII). A major difference between TFF1 and TFF3 is probably the nucleophilicity of Cys
VII, which seems to be enhanced in TFF1 due to steric exposure by neighboring proline residues (
Figure 1). Furthermore, the pKa of Cys
VII is probably different in TFF1 and TFF3 because of the multiple flanking acidic residues, particularly in TFF1 (
Figure 1). In contrast, TFF2 consists of two TFF domains arranged in tandem; the N- and the C-terminals are linked via two additional cysteine residues (
Figure 1). This gives rise to a circular structure, which probably can adopt different conformations
[25][26]. Notably, human TFF2 is N-glycosylated, and the glycosylation probably differs in the stomach when compared with the duodenum; gastric TFF2 contains an unusual fucosylated LacdiNAc oligosaccharide (
Figure 1)
[27]. In contrast, murine and porcine TFF2 lack N-glycosylation sites.
Figure 1. Schematic structures of human TFF1, TFF2, and TFF3 and their major natural forms (minor forms are mentioned in parentheses). Cysteine residues (C; numbering in Roman numerals) and disulfide bridges are shown in green. Proline residues (P) at the C-terminal outside the TFF domains are indicated. Acidic residues in proximity to the C-terminal cysteine residues are shown in blue. Additionally, the N-linked carbohydrate moiety in gastric TFF2 is illustrated.
The secondary and tertiary structures of recombinant TFF peptides (TFF1 and TFF3 homodimers) and purified porcine pancreatic TFF2 were studied by the use of X-ray crystallography as well as two-dimensional
1H NMR spectroscopy (review:
[9]). Generally, the TFF domains show very similar folds. The 3D structures are available from the Protein Data Bank (PDB;
https://rcsb.org/, accessed on 7 September 2021). There is a pronounced groove between the ends of the second and third loops, and this has been discussed as a potential binding site for either an oligosaccharide or an aromatic amino acid side chain. Conserved residues are placed around this groove.
Of special note, TFF1 and TFF3 contain an odd number of cysteine residues, which is unusual for secretory proteins because disulfide formation is enforced in the endoplasmic reticulum (ER). Normally, such free thiol groups act as a three-way switch, mediating assembly, retention, or degradation in the ER, but they are not secreted.
Only long-lasting systematic studies have revealed that TFF peptides appear in unexpected different molecular forms (
Figure 1), indicative of diverse molecular functions (see
Section 3.2; reviews:
[12][14]).
Surprisingly, TFF1 occurs predominantly in a monomeric form, probably with an unusual free thiol group (
Figure 1). This has been demonstrated for
X. laevis (ortholog xP1), mouse and human
[26][28][29]. The reason is probably the cluster of acidic residues as well as the proline residues before (
Figure 1), which would allow TFF1 to escape ER retention, as has been similarly described for Ig light chains
[26][28][29]. Minor forms of TFF1 are disulfide-linked heterodimers with IgG Fc binding protein (FCGBP), a yet-unknown protein X with a Mr of 60 k (TFF1-X), and gastrokine 2 (GKN2), as well as a homodimer
[12][26][29][30][31].
Gastric TFF2 mainly exists in a high molecular mass form associated with the mucin MUC6 (TFF2/MUC6 complex;
Figure 1)
[12][25][31][32][33][34], whereas porcine pancreatic TFF2 mainly forms a non-covalently linked homodimer
[32].
TFF3 mainly forms a disulfide-linked heterodimer with FCGBP, as shown for the intestine and the saliva (
Figure 1)
[35][36]. Minor forms are a monomer and a homodimer. There are indications that the homodimer spontaneously forms only after secretion.
In the frog
X. laevis, the TFF1 ortholog is termed xP1, whereas xP4.1 and xP4.2 are the
X. laevis orthologs of mammalian TFF2
[28][37][38]. As a special feature, xP4.1 and xP4.2 contain four TFF domains arranged in tandem. In xP4.1, exon shuffling of TFF modules has also been documented
[39]). Thus, the TFF domain is a unique shuffled module encoded by a single exon belonging to class 1-1 (review:
[24]).
2.5. Pathological Expression of TFF Peptides: Links to Inflammation and Cancer
Based upon histopathological evaluation (immunofluorescence, immunohistochemistry, in situ hybridization), the ectopic expression of TFF peptides was detected in pathological conditions, particularly during chronic inflammation (for reviews, see
[7][10][13][14][21]). Typical examples are gastro-esophageal reflux disease, Barrett’s esophagus, gastric and duodenal ulcers, diverticulitis, inflammatory bowel disease, pancreatitis, hepatolithiasis, cholecystitis, salpingitis, inflammatory nasal polypi, and chronic obstructive lung disease (COPD). In many cases, a glandular structure termed “ulcer-associated cell lineage” (UACL, also known as pyloric or pseudo-pyloric metaplasia) was the prominent site of TFF peptide synthesis
[46]. After mucosal injury/ulceration, TFF2 was characterized as an early response gene, and TFF3 as a late response gene
[47].
TFF expression is also dysregulated in various metaplasias
[48][49], as well as in tumors (for compilations, see:
[7][10][21][24]). The potential roles of TFF peptides for tumor progression have not been clarified conclusively
[6][50][51][52][53][54].
Regulation of TFF gene expression is complex, and the three TFF genes share several cis-acting elements in their promoter regions
[55]. Regulatory signals reported include estrogen, transforming growth factor α (TGFα), fibroblast growth factors (FGFs), gastrin, TFF peptides (interregulation), pro- and anti-inflammatory cytokines, prostaglandins, arachidonic acid, indomethacin, aspirin, omeprazole, butyrate, hydrogen peroxide, osmotic stress, hypoxia, X-ray irradiation, and pathogens (reviews:
[21][51][52][56][57]). Stimulation by omeprazole and FGFs could be secondary effects due to gastrin.
Numerous studies have analyzed the regulation of TFF expression during inflammation in particular, and the three TFF genes have been shown to often respond differently to inflammatory mediators (for review, see
[13]). Generally, the links between TFF peptides and inflammation are rather complex. On the one hand, inflammatory signals regulate TFF expression. On the other hand, TFF peptides influence inflammatory processes, which also enables complex feedback loops with inflammatory signals.