SLPI is one of the two members of the ALP superfamily of proteinase inhibitors (the other being ESI/elafin; see below). The gene encoding SLPI [2.65 kilobases (kb) in length [1]] seems to be a relatively nonpolymorphic, stable gene that can be modulated at both the transcriptional and translational levels [1]. Transfection studies with fusion elements composed of fragments of up to 1.2 kb of the 5' flanking region of the SLPI gene demonstrated a high promoter activity in a 131 base pair (bp) fragment (- 115 to + 16) relative to the transcription start site. Interestingly, Kikuchi et al [4] have recently delineated within this region a proximal 41-bp region that confers lung specificity for SLPI expression. It is a 11.7 kDa protein consisting of 107 amino acid residues [1] and comprising two domains. It contains 16 cysteine residues that form eight disulfide bridges [1]. SLPI has been shown to inhibit human neutrophil elastase, cathepsin G, trypsin, chymotrypsin and chymase. Its major target is thought to be the human neutrophil elastase in view of its high affinity and kinetic constants (inhibition constant in the nanomolar range and k_ass in the micromolar range) [5].

SLPI has been purified from different sources, including parotid, cervical, seminal and lung secretions [1].

In lung, SLPI is produced in vitro by tracheal, bronchial, bronchiolar and type II alveolar cells, and by monocytes, alveolar macrophages and neutrophils [1,6]. It has also been shown to be produced in vivo by tracheal serous glands and bronchiolar Clara cells, and to be closely associated with elastin fibers in the alveolar interstitium [1]. Its roles in inflammatory cells such as macrophages or neutrophils are uncertain but antibacterial or anti-inflammatory actions have been proposed (see below). Outside the lung, it is secreted in a variety of mucosal sites (leading to its alternative name, mucosal proteinase inhibitor [1]).

The neutrophil elastase inhibitor ESI/elafin (the other member of the ALP superfamily of proteinase inhibitors; reviewed in [2]) was first identified as a `non-SLPI' low-molecular-mass anti-elastase by Hochstrasser et al [7] and Kramps and Klasen [8], and further characterized by us in bronchial secretions [9] and by Wiedow et al and Molhuizen et al in the skin [10,11]. The sequence of the gene for ESI [12] showed that it is approx. 2.3 kb long, and is composed of three exons and two introns and containing typical 5' TATA and CAAT boxes, as well as 5' regulatory sequences such as activator protein-1 and nuclear factor-κB sites [2]. Zhang et al demonstrated that a positive regulatory cis-element present in the region between -505 and -368 bp is responsible for the upregulation of the elafin gene in normal breast epithelial cells (reviewed in [2]). Further characterization will be needed to determine whether this region is tissue-specific or is also important for expression in lung cells. The molecule is composed of 117 amino acid residues, including a hydrophobic signal peptide of 22 residues. Elafin is part of a `four-disulfide' core protein family that has recently been termed the trappin family [2]. Elafin can be divided into two domains, the carboxy-terminal domain containing the antiproteinase active site and the amino-terminal domain containing characteristic VKGQ sequences (in single-letter codes for amino acids). These sequences allow the elafin molecule to glue itself into polymers and bind other interstitial molecules through transglutamination [2]. This feature could make elafin maximally effective as a tissue-bound inhibitor as opposed to A1-Pi, which is present in large amounts in the circulation. SLPI has also been suggested to have a locally protective role against neutrophilic damage, presumably because of its small size and negative charge. The elafin molecule shows a 40% homology with the SLPI molecule and the active sites of both inhibitors are very similar.

Like SLPI, elafin has a high content of cysteine residues[2], which are arranged in four disulfide bonds in the C-terminal proteinase-inhibiting region [2]; a partial crystal structure, not containing the VKGQ sequences (a major feature of the molecule) has recently been determined [2]. Elafin has been shown to be more specific in its spectrum of inhibition than SLPI: it inhibits pig pancreatic elastase, human neutrophil elastase and proteinase-3 [2].

As mentioned above, elafin was first demonstrated in the skin and in lung secretions but is also present at mucosal sites in many tissues. Its purification from sputum, its presence in tracheal biopsies and bronchoalveolar lavage from both normal subjects and patients [13], and its synthesis by Clara cells and type II cells indicate, as for SLPI, a tracheo-bronchioalveolar origin. It is present in highest concentrations in sputum (at approximately 10% of the concentration of SLPI). We have recently investigated its presence in the peripheral lung and have shown by immunohistochemistry that macrophages are a primary source (J-M Sallenave, unpublished data).

In contrast to the A1-Pi deficiency in patients with the genetic form of emphysema, no polymorphisms have been reported so far for SLPI or elafin, although such studies are scarce. There is therefore no definite indication of whether a deficit in either SLPI or elafin could be responsible for the development of COPD in patients that are otherwise A1-Pi sufficient. When the levels of SLPI were measured in relevant controls and patients with COPD, SLPI levels were found in significant amounts in bronchoalveolar lavage and sputum [21,22]. In contrast with the situation in genetic emphysema, in which A1-Pi levels are drastically decreased, we and others have found antigenic levels of SLPI to be increased in COPD and cystic fibrosis [6]. Compared with SLPI, elafin levels seem to be down-regulated in these pathologies, suggesting that SLPI and elafin might be differently regulated in chronic situations.

In ARDS, both SLPI and elafin are markedly increased in bronchoalveolar lavage fluid [13] compared with control subjects. Interestingly, SLPI BAL levels were positively correlated with the multiple organ failure score (MOFS, see Fig. 2), suggesting that antiproteases such as SLPI might be released in this pathology as an aborted and unsuccessful attempt to control injury. Furthermore, in patients with pneumonia, SLPI levels were found to be increased in serum [23], which is consistent with the hypothesis that these inhibitors are `alarm inhibitors' whose role might be relevant at the onset of the inflammatory process.

Although still a contentious issue, emerging evidence suggests that serine proteases might be important in the pathophysiology of asthma. Mast cell and leukocyte serine protease levels are elevated in the airways of asthmatic patients and one study has shown that patients with A1-Pi deficiency develop asthma more readily [24]. Although studies measuring SLPI and elafin in asthmatic patients are scarce, animal models of allergic diseases have shown SLPI to be of benefit in inhibiting both early and late phase events [25].

A1-Pi protein replacement therapy is actively being pursued in patients with severe A1-Pi deficiency and COPD. Although definite results from these studies are eagerly awaited, low-molecular-mass protease inhibitors such as SLPI and elafin have the advantages of size, diffusibility, and possible anchorage to the interstitium (see above), which is where the action is. Although no studies in patients have been reported with the use of recombinant elafin, the use of SLPI in normal subjects and in cystic fibrosis patients has been investigated. Unfortunately, these studies have been hampered by a short half-life and poor accessibility of the recombinant product to diseased areas in cystic fibrosis patients (reviewed in [1]).

Although still in its infancy compared with pharmacological methods of drug delivery, gene therapy might be advantageous, owing to its less transient nature. Because of its natural tropism for the lung, adenovirus could be the vector of choice for gene therapy applications. Historically, most of these applications have been directed in the lung at chronic pathologies such as cystic fibrosis or cancer. This is by no means the best case, given the chronic nature of these pathologies in which a repeated administration of vectors would be necessary to achieve adequate levels of therapeutic transgenes. Unfortunately, the immune response directed against the currently available `first and second generation' adenovirus vectors currently precludes their use in such settings. However, the generation of `gutless' vectors (which express almost no viral genes) has shown to drastically improve the duration of transgene expression in rodent models [26]. This type of vector is an important feature of adenovirus gene therapy techniques, but its generation remains technically challenging at present.

However, the constructs currently available are already useful for the study of experimental models in which immune responses are less of a concern and that mimic a clinical setting in which a therapeutic opportunity might exist between the presentation of patients and the development of irreversible acute inflammation. Such situations could include lung infections leading to pneumonia, ARDS or sepsis.

Indeed, success has been achieved in animal models of lung infections and endotoxemia in rodents, by using adenovirus vectors expressing molecules such as IL-12 and TNF (lung infections) and IL-10 (endotoxemia) [27,28,29]. An important and useful feature of adenovirus administration is the compartmentalization of transgenes administered in the lung, which thereby prevents unwanted systemic effects of potentially toxic molecules such as TNF. Because of the previously mentioned characteristics of SLPI and elafin as `alarm molecules' involved in the regulation of early events in the inflammatory process, the overexpression of these inhibitors would be of benefit in combating bacterial infections and their inflammatory sequelae. Our own studies have shown that the overexpression of elafin with an adenovirus vector (AdmCMV-elafin, in which CMV stands for cytomegalovirus) [30] significantly improved acute lung injury induced by Pseudomonas aeruginosa in C57/Bl6 mice (AJ Simpson and J-M Sallenave, unpublished data). The exact mechanism of the protection conferred by the elafin molecule is unclear at present and is still under investigation, but it might involve a dual mode of action, by killing the pathogens while dampening down the unwanted effects of neutrophil histotoxic molecules such as elastase.