We have identified a mutant with PI caused by a deletion of the protease inhibitor Serpini2. This phenotype (pequeño) is inherited in an autosomal recessive pattern and is to our knowledge the first published mouse mutant to genetically model PI. The mutation, caused by a transgene insertion, led to an approximately 210-kb genomic deletion that completely removes the Serpini2 gene and a portion of the predicted gene E-62199.1/E-38943.1. Reintroduction of the Serpini2 gene alone by BAC complementation corrected the acinar cell defect and improved immune function, demonstrating that deletion of the Serpini2 gene is the primary genetic cause of the pequeño mouse phenotype.
was initially identified and implicated as a potential tumor suppressor gene because it was downregulated in pancreatic cancer cell lines when compared to normal pancreas [9
is also known as MEPI,
a gene expressed in normal breast myoepithelial cells, but not in malignant breast carcinoma cells [14
]. We did not observe any overt phenotype in the pequeño
mice with regard to mammary gland development/function. Female pq/pq
mice that had been maintained on a pancreatic enzyme diet supplementation were able to rear litters of 8–10 pups with no difficulty. It is possible that loss of expression in mammary tissue does not affect mice mammary gland function. Alternatively, there is the potential that another serpin family member may be compensating for the loss of Serpini2
activity, as has previously been observed for loss of Serpinb6
function resulting in upregulation of Serpinb1
SERPINI2 is a member of the serpin superfamily of proteins that have been implicated in a variety of functions including blood coagulation, angiogenesis, inflammation, and programmed cell death [16
]. In humans, 33 serpins have been identified [17
]. These proteins are further subdivided into two classes, inhibitory and non-inhibitory. The former group, which includes SERPINI2, is characterized by conformational structure changes that allow for formation of covalent acyl-enzyme intermediates with the proteins' target protease, and thus have been termed “suicide substrates” [18
]. The target substrates for SERPINI2 are unknown.
Further insight into the cellular location and function of SERPINI2 has come from work with the rat homolog, ZG-46p. SERPINI2/ZG-46p has been characterized in the pancreas, located predominately in the Golgi and in ZGs [19
], and was isolated from ZG membranes [20
]. ZGs in the acinar cells contain inactive proenzyme precursors for digestive enzymes including trypsin, amylase, chymotrypsin, and carboxypeptidase. Normal acinar cells produce large quantities of granules that are transported from the trans-Golgi network to the luminal side of the acinar cell and are then secreted to the ductal lumen upon hormonal and neural signaling mechanisms. The contents of the ZGs are transported via ducts to the duodenum (pp. 429–430 of [21
]), where these enzymes are required for proper digestion and absorption of food.
Given the subcellular location of SERPINI2 in the Golgi and ZGs, in addition to the predicted gene function as a serine protease inhibitor, we hypothesize that SERPINI2 may play a role in regulating the level of inactive zymogen precursors. The importance of tight regulation of active protease levels by protease inhibitors in the exocrine pancreas is well documented. This regulation is demonstrated by considering mutations in two genes that are mutated in human hereditary pancreatitis: the cationic trypsinogen gene PRSS1
and serine protease inhibitor Kazal type 1 (SPINK1)
. For a small number of individuals with chronic PI, mutations in either gene have been shown to alter this balance and lead to autodigestion of tissue with inflammation [22
]. However, genetic animal models for these two genes are lacking.
It is also important to note that differences in the severity of pancreatitis may be regulated by the type of cellular response to the inflammation, whether apoptosis or necrosis. In several models of PI, acinar cell apoptosis causes a less severe pancreatitis phenotype [24
], in contrast to the more severe pancreatitis that has been documented with acinar cell necrosis [25
]. It is interesting to speculate that other less severe alleles than the null allele present in the pequeño
mouse model may not have severe apoptosis and may demonstrate a pancreatitis phenotype similar to that observed in patients with SPINK1
Alternatively, SERPINI2 may function to allow proper sorting and/or transport of ZGs from the Golgi to the ductal lumen of the acinar cell. This hypothesis is consistent with the observations made of the acinar cells of 1-wk-old pq/pq mice that have yet to undergo apoptosis. In these acinar cells there is a dramatic loss of ZGs present in the cell and the ZGs appear smaller in size than those in wild-type and heterozygous littermates. Further analysis is required to differentiate these potential modes of action for SERPINI2.
It is clear that malnutrition is the primary cause of the observed pequeño
phenotypes. The inability of pq/pq
mice to adequately digest and absorb food from a young age adversely affects growth and immune system function. A large body of evidence argues that malnutrition, stress-induced hormones, and corticosteroids can adversely affect immune cell numbers and survival [28
]. In addition, leptin, a pleiotropic molecule that regulates food intake and metabolic endocrine functions, also plays a role in immune and inflammatory responses [29
]. Although the defects we observed here do not completely correlate with those associated with increased steroid exposure in mice, most studies of steroid exposure have examined short-term glucocorticoid exposure [30
]. However, over-expression of a glucocorticoid receptor in the thymus produced results similar to ours, with reduced cellularity observed at multiple stages of T cell development [31
]. Similar long-term studies have not examined B cell numbers or function. While we cannot rule out a direct effect of SERPINI2 on lymphocytes, the improvement of immune function by pancreatic diet supplementation strongly suggests that these defects are secondary to malnutrition, which could result in increased stress hormone and steroid production, decreased leptin levels, and/or a lack of nutrients.
It is interesting that growth retardation, PI, and immunodeficiency are present in both the pequeño
mouse and the human disorder SDS. The human disease is a rare and clinically heterogeneous disorder that manifests PI, short stature, lipidosis of the pancreas, neutropenia, pancytopenia, and predisposition to acute myelogenous leukemia. Although we have not observed neutropenia in the blood of pq/pq
mice (Figure S1
; Table S2
), it may be difficult to compare secondary immune defects in mouse and human, as effects of malnutrition and stress hormones may differ between the species. Nonetheless, the combination of PI and immune defects in this mouse mutant is intriguing, given that several recent studies indicate SDS may exhibit genetic heterogeneity [32
The most common mutation for SDS has been identified and is a gene conversion caused by recombination of the ubiquitously expressed SBDS
gene with the pseudogene SBDSP
located 5.8 Mb away on Chromosome 7 [32
]. The SBDS protein is predicted to be an RNA-processing enzyme based upon sequence homology to a yeast ortholog and cluster analysis data from microarray experiments [32
]. Given the overlap in phenotypes of the pequeño
mice and genetic heterogeneity predicted for patients with SDS, Serpini2
is a candidate gene for analysis in the subset of patient samples in which no mutation in the SBDS
gene can be identified.