Sorting behavior of a transgenic erythropoietin-growth hormone fusion protein in murine salivary glands

Abstract

Salivary glands are useful gene transfer target sites for the production of therapeutic proteins, and can secrete proteins into both saliva and the bloodstream. The mechanisms involved in this differential protein sorting are not well understood, although it is believed, at least in part, to be based on the amino acid sequence of the encoded protein. We hypothesized that a transgenic protein, human erythropoietin (hEpo), normally sorted from murine salivary glands into the bloodstream, could be redirected into saliva by fusing it with human growth hormone (hGH). After transfection, the hEpo-hGH fusion protein was expressed and glycosylated in both HEK 293 and A5 cells. When packaged in an adenovirus serotype 5 vector and delivered to murine submandibular cells in vivo via retroductal cannulation, the hEpo-hGH fusion protein was also expressed, albeit at approximately 26% of the levels of hEpo expression. Importantly, in multiple experiments with different cohorts of mice, the hEpo-hGH fusion protein was sorted more frequently into saliva, versus the bloodstream, than was the hEpo protein (p < 0.001). These studies show it is possible to redirect the secretion of a transgenic constitutive pathway protein from salivary gland cells after gene transfer in vivo, a finding that may facilitate developing novel treatments for certain upper gastrointestinal tract disorders.

FIG. 1

Immunoreactive hEpo and hGH detected in media and lysates of transfected HEK 293 cells. Forty-eight hours after transfection either with pAC-CMV-hGH, pAC-CMV-hEpo, or pAC-CMV-hEpo-hGH, media were collected and subjected to (A) hEpo ELISA or (B) hGH ELISA. Media from hEpo- and hGH-transfected cells were immunoreactive only in their respective ELISA. Media from hEpo–hGH-transfected cells reacted positively with both hEpo and hGH ELISAs. Data are presented as means ± SEM (n = 3). Western blots (C) from lysates of cells transfected with plasmids encoding hEpo–hGH reacted with both anti-hEpo (left) and anti-hGH (right) antibodies. hGH and hEpo have apparent molecular masses of ~22 and ~34 kDa, respectively. The predicted size of the chimeric hEpo–hGH fusion protein is ~56 kDa. The arrows to the right indicate the molecular masses of simultaneously electrophoresed protein standards. See Materials and Methods for additional details.

FIG. 2

Western blot analyses of the glycosylation state of hEpo and hEpo–hGH protein. Forty-eight hours after transduction with AdCMVhEpo or AdCMVhEpo-hGH lysates were collected from transduced HEK 293 cells and either treated with N-glycosidase F (+) or left untreated (−). Proteins were then separated by gel electrophoresis, blotted, and reacted with anti-hEpo antibody. On the basis of the shift in electrophoretic mobility after incubation with N-glycosidase F, hEpo, the fusion protein hEpo–hGH, and the likely dimer of hEpo–hGH all were present in their glycosylated form with apparent molecular masses of ~34, ~56, and ~110 kDa, respectively. See Materials and Methods for additional details.

FIG. 3

Immunoreactive hEpo in aqueous extracts of mouse submandibular glands. Left: Forty-eight hours after administration of AdCMVhEpo-GH or AdCMVhEpo at 10¹⁰ viral particles/gland, glands were harvested and frozen at −80°C. Thereafter, glands were homogenized and levels of immunoreactive hEpo were determined in gland extracts by ELISA. Glands transduced with AdCMVhEpo had a median protein level of 2.99 mU/ng protein (range, 1.18–1292.4 mU/ng protein), whereas glands transduced with AdCMVhEpo-hGH showed a median protein level of 0.79 mU/ng protein (range, 001–255.6 mU/ng protein). The horizontal bar represent the median value (n = 20, p < 0.001). Right: Extracts from glands were electrophoresed as described in Materials and Methods, blotted, and then reacted with anti-hEpo antibody. Mouse glands transduced with AdCMVhEpo-hGH or AdCMVhEpo expressed immunoreactive hEpo–hGH and hEpo protein bands with apparent molecular sizes of ~56–58 and ~32–34 kDa, respectively.

FIG. 4

Ratios of immunoreactive hEpo products in saliva and serum from mice transduced with either AdCMVhEpo-hGH or AdCMVhEpo at 10¹⁰ particles/gland. Mice were administered either vector via the submandibular glands. After 2 days serum and saliva were collected and hEpo-immunoreactive products were measured as described in Materials and Methods. Animals transduced with AdCMVhEpo-hGH had significantly higher saliva:serum ratios of immunoreactive hEpo products compared with those transduced with AdCMVhEpo (p < 0.001), that is, relatively higher levels of immunoreactive hEpo were being secreted into saliva. For these analyses, the number of mice tested was n = 32 for hEpo–hGH and n = 14 for hEpo.