 |
Description  |
|
|
BACKGROUND OF THE INVENTION
The epithelium is the first line of defense of the airways against invading
pathogens. Many of the non-specific defenses against such invaders arise
from respiratory epithelial cells. Epithelial cells produce low molecular
weight antimicrobial peptides, antibacterial enzymes, and antiproteases.
In addition, secretory immunoglobulin A, a non-specific immunoglobulin
defense, is trafficked into the airway via a specialized receptor, the
polymeric immunoglobulin receptor (pIgR), that is expressed only on
epithelial cells.
These epithelial defenses are breached early in the life of patients with
cystic fibrosis (CF). Once live bacteria reach their surface, the
epithelial cells direct the initial inflammatory response by releasing
interleukin-8 (IL-8) and interleukin-6 (IL-6) as well as reducing
expression of interleukin-10 (IL-10). The chemoattractants, combined with
increased expression of adhesin molecules for neutrophils, enhance
inflammatory cell migration into the airways. Once there, the neutrophils,
in an attempt to clear the bacteria, release lytic enzymes in the process.
If the neutrophils remain adherent to the epithelium, these enzymes are
released right at the epithelial surface. Both mechanical disruption of
cells and even low concentrations of neutrophil elastase (NE) result in
the greater release of pro-inflammatory mediators from the respiratory
epithelium. Thus, the inflammatory response is further enhanced.
Several strategies to interrupt this cycle have been proposed. Augmenting
the antibacterial defenses of the airway at the epithelial surface may be
useful. Prevention of the escalation of the inflammatory responses
engendered by the neutrophils migrating into the airway could be
accomplished by preventing the action of elastase at the airway cell
surface. Both antibiotics and antiproteases are available for clinical
use. Unfortunately, the results of clinical studies examining the use of
the antiprotease in patients with CF have been disappointing. The systemic
administration of alpha.sub.1 -antitrypsin (A.sub.1 AT) is inefficient,
and the levels achieved by the intravenous administration of the
antiprotease are insufficient to inhibit the overwhelming amount NE in the
lung of patients with CF. Aerosolized A.sub.1 AT should permit the direct
delivery to the airways, but the antiprotease delivered by nebulization
has been uneven and deposits the drug atop the mucus blanket rather than
the critical site at the surface of the cell.
Thus there is a need in the art for methods to circumvent these
difficulties and protect the respiratory epithelial cell surface.
SUMMARY OF THE INVENTION
It is an object of the invention to provide a fusion-protein useful for
protein delivery.
It is another object of the invention to provide a method of delivering a
therapeutic protein to an epithelial cell.
It is yet another object of the invention to provide a nucleic acid
molecule which encodes a fusion protein useful for protein delivery.
These and other objects of the invention are provided by one or more
embodiments as described below. In one embodiment a fusion protein is
provided. The fusion protein comprises a single chain Fv molecule directed
against a human transcytotic receptor covalently linked to a therapeutic
protein. The therapeutic protein may be, for example,
.alpha.,-antitrypsin, a cytokine, such as interleukin-2 or interleukin-10,
or a peptide antibiotic. Suitable peptide antibiotics include aerosporin,
amphomycin, aspartocin, bacitracins, caperomycins, colistins,
dactinomycins, glumamycins, gramicidin D, gramicidin S, mikamycin B,
polymixins, pristinamycin, siomycin, staphylomycin S, thiostrepton,
tyrocidines, tyrothricin, valinomycin, vancomycin, veramycin B. Any
therapeutic protein which one wants delivered to epithelial cells may be
used. The fusion protein may further comprise a linker region of less than
50, 40, 30, 20, or 10 amino acid residues. The linker can be covalently
linked to and between the single chain Fv molecule and the therapeutic
protein.
Also provided according to another aspect of the invention is a method of
delivering a therapeutic protein to an epithelial cell. The method
comprises: administering a fusion protein as described above to a patient,
whereby the therapeutic protein is delivered to an epithelial cell. The
eptithelial cell may be an airway epithelial cell or an intestinal lumen
cell, for example. The liver may also be targeted. The administration mode
may be any known in the art. However, intravenous administration has been
found to be both convenient and efficient.
Nucleic acid molecules are also provided by the present invention. These
encode a fusion protein comprising a single chain Fv molecule directed
against a transcytotic receptor covalently linked to a therapeutic
protein. The therapeutic protein may be, for example, .alpha..sub.1
-antitrypsin, a cytokine, such as interleukin-2 or interleukin-10, or a
peptide antibiotic. Any therapeutic protein which one wants delivered to
epithelial cells may be used. The fusion protein may further comprise a
linker region of less than 50, 40, 30, 20, or 10 amino acid residues. The
linker can be covalently linked to and between the single chain Fv
molecule and the therapeutic protein. Host cells and vectors for
replicating the nucleic acid molecules and for expressing the encoded
fusion proteins are also provided. Any vectors or host cells may be used,
whether prokaryotic or eukaryotic. Many vectors and host cells are known
in the art for such purposes. It is well within the skill of the art to
select an appropriate set for the desired application.
The present invention thus provides an efficient means of delivering
therapeutic proteins to body parts which are often inaccessible or
difficult to reliably access.
BRIEF DESCRIPTION OF THE DRAWING
FIG. 1. Schematic diagrams showing the transport of fusion proteins from
the systemic circulation to the epithelial surface (FIG. 1A). The fusion
protein is bound to pIgR at the basolateral surface and is trafficked to
the apical membrane (FIG. 1B). Once it reaches the this surface, the
fusion is released into the airway lumen, attached to secretory component
of the polymeric immunoglobin receptor (SC), where the antiprotease
component binds and neutralizes elastase (FIG. 1C1). FIG. 2. Clearance and
hepatic uptake of the anti-rat SC Fab. Blood (FIG. 1A) and bile (FIG. 1B)
were collected every 10 minutes after injection with 50 .mu.g of
irrelevant (n=5) or anti-rat SC Fab (n=5), and analyzed for rabbit-derived
antibodies using an ELISA. The Fab levels are represented as the
percentage of peak serum concentrations. FIG. 3. Schematic diagram of
antibodies, Fab fragments, and single chain Fv fragments, showing the
variable (V.sub.L and V.sub.H and constant (C.sub.L and C.sub.H) regions.
The antibody can also be resurfaced to replace the murine-specific
framework regions in the Fv portion of the antibody with human sequences,
which should further reduce its potential immunogenicity.
FIG. 4. Binding of anti-human SC monoclonal antibodies. (FIG. 4A) Relative
binding of antibodies to purified human SC, as measured by ELISA.
Monoclonal antibody 4121 is indicated by solid column. (FIG. 4B) Relative
binding of antibodies to sIgA. (FIG. 4C) Additional characterization of
anti-human SC antibody 4121. Note the difference in affinity of the
antibody to human and rat SC.
FIG. 5. (FIG. 5a) Construction of the anti-hSC Fv protein by PCR. Total
cellular RNA was extracted from the antibody-producing cells and treated
with Moloney Murine Leukemia Virus reverse transcriptase using random
hexamers as primers. The resultant cDNA were screened for the V.sub.L and
V.sub.H domains using different oligonucleotide primers, and these
sequences were amplified by the PCR. The V.sub.L and V.sub.H domains were
then amplified to include linker sequences that permitted splicing using a
PCR technique called overlap extension to produce the full-length gene
encoding the single chain Fv. (FIG. 5b) Schematic diagram of the structure
of the anti-human SC Fv/human A.sub.1,AT chimeric genes. (FIG. 5c)
Restriction endonuclease digestion of plasmids containing the anti-human
SC Fv/human A.sub.1,AT chimeric gene. One microgram of plasmid DNA was
digested with Cla I/Hind III (lane 1), Cla I/Xba I (lane 2), and Hind III
(lane 3). Molecular weight markers are indicated in the right lane. FIG.
6. In vitro transcription and translation of the anti-human SC Fv, human
A.sub.1,AT, and anti-human SC Fv/human A.sub.1,AT fusion proteins.
Messenger RNA was translated using reticulate lysates, and [.sup.35
S]-labeled methionine was incorporated in the synthesized proteins.
Analysis of the proteins by electrophoresis in SDS-polyacrylamide gels
showed the presence of anti-human SC Fv, 26 kDa; human A.sub.1 AT, 52 kDa;
and anti-human SC Fv/human A.sub.1 AT, 78 kDa.
FIG. 7. Expression of anti-human SC Fv/human A.sub.1 AT and anti-D8
Fv/human A.sub.1 AT in prokaryotes. Protein extracts from bacterial clones
obtained after transformation were purified by nickel-chelate affinity
chromatography, subjected to electrophoresis in SDS-polyacrylamide gels,
and transferred onto nitrocellulose membrane filters using established
methods. The polyhistidine, single chain Fv, and human A.sub.1 AT
components were each identified by Western blot hybridization. The
following samples were examined: anti-D8 Fv/human A.sub.1 AT (IR), and
anti-human SC Fv/human A.sub.1 AT (SC). Arrow shows the expected molecular
weight of the fusion. A lower molecular weight band was also present, most
likely representing truncated fusion protein. p FIG. 8. Recognition of
human SC by the anti-human SC Fv/human A.sub.1 AT, as measured by ELISA.
The anti-human SC Fv-based fusion proteins (SC) bound to SC from human
milk, indicating that the Fv portion is functional. Fusion proteins
containing the irrelevant, anti-D8 Fv (IR) did not bind to human SC.
FIG. 9. Expression of the human pIgR in transfected MDCK cells from the
initial sort after selection for neomycin resistance and after serial
sortings by fluorescent activated cell sorter analysis. Clones with the
highest level of the pIgR expression were selected and used in subsequent
experiments.
FIG. 10. (a) Schematic diagram of the cell model system, showing the
transport of fusion proteins or antibodies across polarized MDCK cells
expressing the pIgR in the basolateral-to-apical direction. (b) Transport
of the anti-human SC antibodies across the MDCK cell monolayer that
express the pIgR. Apical media was collected over six hours after addition
of the antibodies to the basolateral media, and the concentration (ng/ml)
of the mouse-derived antibody was determined by ELISA.
The anti-human SC antibodies (4121 and 4214) were effectively transported
from the basolateral surface to the apical media, whereas an irrelevant
antibody (D8) did not.
None of the antibodies were transported in the apical-to-basolateral
direction.
FIG. 11. Effect of nocodazole and leupeptin on transport of antibodies
across transduced MDCK cells. Apical media was collected at different
times after addition of the monoclonal antibody (4121) to the basolateral
media, and the concentration (ng/ml) of the mouse-derived antibody was
determined by ELISA. Both leupeptin and nocodazole reduced the amount of
antibody detected in the apical medium in a dose-dependent fashion.
Moreover, no transport of the antibody occurred in nontransfected MDCK
cells or transduced cells in the apical-to-basolateral direction.
FIG. 12. Transport of anti-human SC-based protein conjugates across a cell
monolayer. Twenty-four hours after addition of 1 .mu.g of the conjugate to
the basolateral or apical media, media was collected from the apical or
basolateral compartments, respectively. The amount of immunoreactive
A.sub.1 AT (ng) transported was determined by ELISA. The anti-human
SC/human A.sub.1 AT conjugate was effectively transported in the
basolateral-to-apical direction (solid columns) across the MDCK cells that
express the pIgR. Virtually no transcytosis of the fusion proteins
occurred in the opposite, apical-to-basolateral direction (open columns).
Nontransfected MDCK cells did not transport either the bona fide or
irrelevant fusion protein (in each group, n=5).
FIG. 13. Transport of fusion proteins across a cell monolayer. Twenty-four
hours after addition of 2 .mu.g of the fusions to the basolateral or
apical media, media was collected from the apical or basolateral
compartments, respectively, and the amount of immunoreactive A.sub.1 AT
(ng) transported was measured by ELISA. The anti-human SC Fv/human A.sub.1
AT fusion was effectively transported in the basolateral-to-apical
direction (solid columns) across the MDCK cells that express the pIgR.
Considerably less of the fusion protein (approximately 25%) was
transcytosed in the opposite, apical-to-basolateral direction (open
columns). Purified human A.sub.1 AT was not transported in either
direction. Nontransfected MDCK cells did not transport either human
A.sub.1 AT or the fusion protein (in each group, n=5).
DETAILED DESCRIPTION OF THE INVENTION
The inventors have discovered that one can efficiently deliver functional
proteins to relatively inaccessible sites by targeting the respiratory
epithelium via the polymeric immunoglobin receptor (pIgR). Once
synthesized, the pIgR is trafficked to the basolateral surface of
epithelial cells where it is specifically adapted for the internalization
and nondegradative transfer of polymeric antibodies (21), i.e., dimeric
immunoglobulin A (dIgA) and pentameric immunoglobulin M (pIgM). Once
internalized, the receptor-ligand complex is transported across the cell
to the apical surface, where the receptor is cleaved, releasing dIgA bound
to the ectoplasmic domain of the receptor, or secretory component (SC),
into the lumen FIG. 1). The receptor does not require the natural ligand
for endocytosis, and antibodies (or Fab fragments) directed against human
SC also undergo efficient transcytosis (22). In humans, the receptor is
expressed in airway epithelial cells which reach the luminal surface and
in cells of the submucosal glands, especially serous cells (23). Thus, the
pIgR in humans is well-suited for the delivery of fusion proteins to
bronchi and bronchioles. Indeed, this receptor permits the delivery of the
therapeutic proteins, such as antiprotease, to the apical surface of the
respiratory epithelium.
The Fv portion of an antibody is a 26 kDa heterodimer consisting of the
amino-terminal variable domains of the heavy and light chains, and is the
smallest fragment to bear the antigen binding site (26). Genetically
engineered single chain Fv (Fv) peptides have been synthesized by
attaching the carboxyl terminus of one variable domain to the amino
terminus of the other with a peptide linker (27-29). These Fv fragments
have been shown to bind specific antigens, like the transferrin receptor
(30), have been used to localize fusion proteins to targeted cells.
Investigators have used such chimeras the deliver fusion proteins
containing recombinant toxins (e.g., Pseudomonas exotoxin) and selectively
kill cells in vitro and in vivo that express the appropriate receptor
(30,31). There is considerable experience with expressing such fusions and
retaining function of both components.
Different Fv fragments can be employed to target different receptors,
permitting the targeting of alternative cells. For example, cancer cells
have receptors which can be used to target toxins to cancer cells. One
example of such a receptor is EGFRvIII. See U.S. Pat. U.S. 5,212,290,
which discloses antibodies to such a cancer cell specific receptor.
Another useful receptor which can be targeted is the serpin-enzyme complex
(sec) receptor. This receptor is found on macrophages. Targeting the sec
receptor would allow the delivery, e.g., of anti-tuberculous antibiotics
into macrophages, where tubercle bacilli reside. Preferably the receptor
will be a transcytotic receptor, which can deliver the therapeutic fusion
protein to the other side of a cell. Such receptors include the
immunoglobulin transporting receptors in the gut of infants, and the
immunoglobulin transporting receptors in the placenta.
A variety of functional proteins can be preferentially delivered to the
respiratory epithelial surface. These include A.sub.1 AT and SLPI. For
example, if Pseudomonas aeruginosa interacts with respiratory epithelial
cells to stimulate the production of IL-8 and other pro-inflammatory
mediators, then it may be crucial that antibacterial protection occurs
right at the epithelial surface. Recombinant defensins or protegrins,
endogenous antibacterial peptides, could also be delivered to the
periciliary space using such a bifunctional protein. Indeed, such proteins
have been identified in human airway epithelial cells (8,9). The function
of certain defensins against Pseudomonas aeruginosa may be hindered by the
altered electrolyte composition of ELF in the CF lung (32). Thus,
salt-insensitive forms of these antibacterial peptides may be used if the
sodium chloride concentration of the ELF is abnormal. Another intriguing
strategy is coupling Colistin, an agent already in use to treat pulmonary
infections in patients with CF, to anti-human SC Fv (33). Killing
Pseudomonas at the epithelial surface may be of value if it is the
interaction of the bacteria and epithelial cell that incites the
inflammatory process. Anti-inflammatory cytokines can also be transported
to the epithelial surface in this manner (34) and pulmonary inflammation
can be blocked by the specific delivery of interleukin-10(IL-10), which
can prevent the influx of neutrophils in the airway.
The targeting of proteins by the pIgR in humans provides an additional
level of safety in vivo, since the fusion protein not delivered to the
lung will be transported to the intestinal lumen, through either the
enterocyte or in bile, where it will be excreted. Indeed, the trafficking
pattern of the anti-human SC Fv-based fusion protein could also
potentially be exploited to deliver relevant proteins to other tissues.
The bile ducts and intestine are relatively inaccessible from the luminal
surface and could also be targeted in this manner.
The expression and tissue distribution of the pIgR in rodents is different
from that observed in humans (35,36). In rodents, the expression of the
receptor is significantly greater in the liver than lung, based on the
production of SC. Moreover, fifty per cent of the radiolabeled dIgA
injected into the systemic circulation is rapidly transported from blood
to bile by rat hepatocytes, while less than two per cent was detected in
the lung after two hours (37). The clearance and tissue distribution of
rabbit-derived, anti-SC Fab antibodies in rats are similar to the natural
ligand. Yet, rodents may still serve as a model for the delivery of fusion
proteins to epithelia.
SLPI has potential advantages over A.sub.1 AT as the antiprotease component
of the bifunctional protein. SLPI is a potent antagonist of serine
proteases that accounts for the majority of elastase inhibitory capacity
of endobronchial secretions, and it also does not require glycosylation
for its function or stability in serum (11,12). An analogous fusion
protein can be produced using SLPI, replacing human A.sub.1 AT with SLPI.
The gene encoding a bifunctional protein containing anti-human SC Fv
linked to SLPI can be constructed using the same techniques described
above. The entire SLPI cDNA from human respiratory epithelial cells can be
amplified using primers for the antiprotease gene. Like the previous
constructs, specific sites of recognition for restriction endonucleases,
including a unique Cla I site, can be incorporated in human SLPI
oligonucleotide primers to permit cloning into an expression vector. The
gene encoding the SLPI can be inserted "in frame" into the cloning vector,
the plasmid pRc/CMV, downstream to anti-human secretory component single
chain Fv from monoclonal antibodies. The sequence of the chimeric gene can
be readily confirmed. Other proteins which may be used as therapeutic
components of the fusion protein include cytokines, interleukin-2,
interleukin-10, and peptide antibiotics.
The fusion protein may comprise other polyamino acid sequences, in addition
to the single chain antibody and the therapeutic protein. Linker regions
may be desirable to space the two portions of the protein from each other
and to provide flexibility between them. Typically these will be less than
30 amino acid residues, and consist of predominantly neutral residues.
Other moieties may also be included, as desired. These may include a
binding region, such as avidin or an epitope, which may be useful for
purification and processing of the fusion protein. In addition, detectable
markers may be attached to the fusion protein, so that the traffic of the
fusion protein through a body or cell may be monitored conveniently. Such
markers may include radionuclides, enzymes, fluors, etc.
The fusion proteins of the invention are particularly useful for
administration to epithelial cells. Thus airway epithelial cells and
intestinal lumen cells are particularly good targets for the proteins.
While any mode of administration to these organs will work, such as
targeted or localized administration modes, the fusion proteins of the
present invention can be administered systemically. The single chain
antibody portion of the protein provides an excellent means of targeting,
thus alleviating the need for targeted means of administration.
Suitable dosages for administration can be readily determined, and will
depend somewhat on the therapeutic protein being delivered. However,
typical dosage ranges will be between 0.05 mg and 500 mg, preferably
between 0.5 mg and 50 mg, and more preferably between 1 mg and 10 mg. Due
to the targeted nature of the fusion protein, lower dosages can be used of
the therapeutic protein than are required when administering the
therapeutic protein alone. In the case of the delivery of antiproteases to
the cystic fibrosis lung, approximately 10-100 mg, preferably 70 mg, would
need to be adminstered intravenously to achieve protection against
neutrophil elastase.
The following examples provide specific modes of carrying out the
invention. However, the invention is not limited or defined by the scope
of these examples.
EXAMPLE
Example 1
Targeting the Polymeric Immunoglobulin Receptor In Vitro
We have shown that a complex consisting of the Fab portion of
rabbit-derived, polyclonal antibody raised against human SC covalently
linked to poly (L-lysine) will bind and condense plasmid DNA (Ferkol, et
al. (1993) Gene transfer into respiratory epithelial cells by targeting
the polymeric immunoglobulin receptor. J. Clin. Invest. 93: 2394-2400.)
The complexes effectively delivered foreign genes to human tracheal
epithelial cells in culture which were induced to express pIgR (38).
Delivery was specific for cells in culture that express the receptor,
since human tracheal epithelial cells grown on plastic, a condition that
down-regulates the expression of the receptor, fail to express the
reporter gene whereas cells from the same trachea maintained on collagen
gels can be transfected. Delivery of DNA is inhibited by excess human SC
in the medium, which presumably occupies the recognition site on the Fab
fragment, preventing its interaction with the receptor.
However, competition for the pIgR with dIgA in a four-fold molar excess
failed to block the delivery of the complex, perhaps indicating that the
binding site(s) on the pIgR for dIgA and antibody do not overlap.
Alternatively, the natural ligand may not compete effectively with the
anti-human SC for the receptor, or the receptor may be present in excess.
Uptake is not due to a non-specific increase in pinocytosis secondary to
the presence of the Fab fragment in the culture medium, since the use of
complexes with Fab fragments from irrelevant antibodies did not permit
expression of reporter genes.
A variable percentage of human tracheal epithelial cells in primary culture
were transfected through the pIgR. We have shown that differences in
receptor expression in the cultured cells accounts for much of the
observed variation. The proportion of human tracheal epithelial cells in
culture which express pIgR which is detectable by immunofluorescence
ranged from eight to thirty-five per cent, compared to five to sixty-six
per cent of the cells which express the reporter gene delivered by the
conjugate. The expression of the reporter gene co-localized to cells that
expressed the receptor, as identified by immunohistochemical means. Thus,
conjugates containing Fab fragments directed against human SC mediated the
in vitro uptake of macromolecules into cells that expressed pIgR (38).
Example 2
Targeting the Polymeric Immunoglobulin Receptor in Vitro
We have examined the pattern of transport of the anti-rat SC Fab fragments
in adult rats to determine if these antibodies have the same vascular
distribution and clearance as dIgA. Fifty micrograms of the anti-rat SC
antibodies were injected into the systemic circulation, and serial samples
of bile and blood were collected every ten minutes and examined for the
rabbit-derived antibody by enzyme-linked immunosorbent assay (ELISA). The
anti-SC Fab was rapidly cleared from the blood, and the antibody appeared
in the bile twenty minutes after infusion (FIG. 2). No uptake in the bile
was noted after injection with pre-immune rabbit-derived Fab antibody
(FIG. 2). The rat anti-SC Fab, however, was not detected in BAL fluid
obtained two hours after injection, which may be related to the fifty-fold
dilution of ELF. In addition, BAL preferentially samples the alveolar
space, where the receptor is not expressed.
Example 3
The pIgR was exploited for gene delivery into rats in vivo (Ferkol, et al.,
(1995) Gene transfer into airways in animals by targeting the polymeric
immunoglobulin receptor. J. Clin. Invest. 95: 493-502.).- Because the
receptor is asymmetrically distributed, predominantly on the basolateral
surface of epithelial cells, the complex should best be delivered by the
systemic circulation. In our initial experiments, we tested the transfer
of reporter genes into the lungs and livers of rats (39). Two tissues that
do not express the pIgR, heart and spleen, were also tested as controls.
Three hundred micrograms of the expression plasmid pGL2, consisting the
SV40 viral promoter and enhancer ligated to the Photinus pyralis
luciferase gene inserted into the Escherichia coli pUC 19 vector,
complexed to the anti-SC Fab-polylysine conjugate into the caudal vena
cava of rats. Luciferase expression was found in the homogenates from the
liver and lungs, but not spleen and heart. Animals treated with the
complexes containing either an irrelevant plasmid or the bonafide
expression plasmid bound to a carrier based on an irrelevant Fab fragment
resulted in no significant luciferase activity in any tissue examined.
Thus, only tissues that contain cells bearing pIgR are transfected, and
transduction cannot be attributed to nonspecific uptake.
Specifically, transgene expression in the rats was greatest in the lung and
less in liver, despite the recovery of the anti-rat SC Fab preferentially
in bile. Thus, the airways are clearly accessed by the transfection
complex in substantial amounts. Although, when we examined cellular
distribution of the transgene, only seventeen per cent of the tracheal
epithelial cells were positive for bacterial .beta.-galactosidase. A more
sensitive reporter (i.e., interleukin-2 receptor gene), however, showed
that the majority of tracheal epithelial cells expressed the transgene.
These data demonstrate the ability of transfection complexes directed only
by the anti-SC Fab to access airway epithelial cells after intravenous
administration (39).
Example 4
Because of this serologic response, the molecular conjugates were
subsequently modified to reduce their immunogenicity. Specifically,
anti-SC Fv fragments have been synthesized as ligands (FIG. 3). This
approach requires the construction of monoclonal antibodies to replace the
polyclonal antibodies used in the studies described above, and then
prepare a single chain Fv. This strategy removes the species-specific
constant regions of the Fab, leaving only the framework regions in the
variable domain that are still murine-specific. Finally, since the peptide
binding domain for the polymeric immunoglobulin receptor has not been
mapped and probably is a three-dimensional region composed of sequences
from both IgA molecules and connecting J chain (41), there are no peptide
alternatives that could be used to target this receptor.
Example 5
Production of Monoclonal Antibodies Against the Human Polymeric
Immunoglobulin Receptor
Balb/c mice were hyperimmunized with purified human SC, which was isolated
from human colostrum. The mice underwent intraperitoneal injections with
human SC, subcutaneous injection with Freund's adjuvant, and bled weekly.
An ELISA was used to detect the production of antibodies directed against
human SC. Three of the five inoculated mice showed a substantial response
in serum against the antigen two weeks after the second immunization.
Spleens were harvested from two mice and used in fusion experiments with
the SP2/0 mouse myeloma cell line using a standard approach for the
generation of hybridomas (42).
The hybridoma cells were then placed in selective media containing
hypoxanthine and thymidine, which eliminates myeloma cells that have not
fused, and their supernatants were analyzed for the production of
anti-human SC antibodies. Positive hybridomas were cloned twice by
limiting dilution, and several subclones continued to produce of
anti-human SC antibodies, as detected by ELISA. Monoclonal antibodies
against human SC were generated, all of which cross-react with purified
human SC (FIG. 4) and sIgA (FIG. 4). Therefore these antibodies most
likely recognize epitopes other than the binding site of the natural
ligand. Based on analysis using a fluorescence activated cell sorter
(FACS), the monoclonal antibodies 4121 and 4214 were selected for the
initial experiments because they best recognized the receptor expressed on
the surface ofMDCK cells transduced with the cDNA encoding the human pIgR.
Example 6
Expression Plasmids Encoding Fv Fragments of Anti-human Secretory Component
The V.sub.L and V.sub.H portions of the anti-human SC antibodies were
cloned from the hybridoma cell lines. Total cellular RNA was extracted
from the antibody-producing cells, and the mRNA transcripts were treated
with Moloney Murine Leukemia Virus reverse transcriptase using random
hexamers as primers. The resultant cDNA were screened for the V.sub.L and
V.sub.H domains using different oligonucleotide primers, as described by
Nicholls and colleagues (43), and these sequences were amplified by the
polymerase chain reaction (PCR). The DNA sequences were spliced, separated
by an interdomain linker that encodes fourteen or fifteen amino acids
using a PCR technique called overlap extension (28,43). Molecular modeling
(44) and nuclear magnetic resonance analysis (45) have shown the optimal
size of the linker in a single chain Fv has been determined to be greater
than 12 amino acids. The use of glycine as the predominant amino acid in
the linker permits the greatest molecular flexibility, thus allowing the
two domains to fold properly and assume the proper orientation with each
other. The cDNA encoding the V.sub.L and V.sub.H domains of the anti-human
SC Fv were inserted into the expression vector downstream to the promoter
for T3 RNA polymerase. The domains of the single chain Fv was assembled in
the order V.sub.L -linker-V.sub.H, though either orientation has been used
to produce functional Fv fragments. The anti-human SC Fv cDNA was
sequenced by dideoxy chain termination.
Example 7
Chimeric Genes Encoding Anti-secretory Component Fv and Human
Alpha.sub.1,-antitrypsin
The CDNA sequences encoding the anti-human SC Fv from monoclonal antibody
4121 and irrelevant antibody directed against an anti-Pseudomonas
polysaccharide (D8) was amplified through 30 cycles, using the following
primers for the Fv: GG CCC AAG CTT GCC ACC ATG GAC ATT GTG CTG (SEQ ID
NO:1), a primer to detect the 5' region, and CCT AGT CTA GAC TTA CAT CGA
TGA GGA GAC TGT GAG AGT GGT GCC (SEQ ID NO: 2), an antisense primer. The
Kozak start sequence was placed immediately upstream of the Fv sequences
to permit the optimal translation efficiency of the fusion protein in
eukaryotes (46). Additional sequences encoding anti-human SC Fv isolated
from other hybridoma clones are currently being generated which will serve
as alternative ligands
The entire human A.sub.1 AT cDNA was amplified, using primers for the human
AIAT gene: GAG CCA TCG ATG CCG TCT TCT GTC TCG TGG (SEQ ID NO:3), a primer
to the 5' end, and CCT AGT CTA GAT AAG CTT TTA TTT TTG GGT GGG ATT CAC
(SEQ ID NO:4), an antisense primer that corresponded to the 3' end of the
gene. Specific sites of recognition for restriction endonucleases--Hind
III, Cla I, and Xba I sites in the Fv cDNA and the human A.sub.1 AT
primers were incorporated into both sets of primers to permit the excision
of the intact chimeric gene and their cloning into the expression vector
"in-frame."
An E. coli cloning vector (pRc/CMV) was used to construct the anti-human SC
Fv/human A.sub.1 AT chimeric gene. The vector pRc/CMV is designed for high
level, stable expression in eukaryotic cells, and contains a multiple
cloning site polylinker, cytomegalovirus promoter and enhancer, and bovine
growth hormone polyadenylation signal. The vector also contains a neomycin
resistance gene driven by the SV40 early promoter for the selection of
stable transformants. This plasmid was digested with Hind III and Xba I,
thus removing a segment of the multiple cloning site, and the amplified
cDNA encoding the anti-human SC Fv will be inserted. This construct was
digested with Cla I and Xha I, and the amplified cDNA encoding human
A.sub.1 AT was ligated into this site in the same transcriptional
orientation as the anti-SC Fv (FIG. 5). No episomal origin of replication
is present in the plasmid. The nucleotide sequences of the 5' and 3' ends
of the chimeric gene have been examined by dideoxy chain termination, and
no rearrangements have been noted. Moreover, the fidelity of the chimeric
gene was verified by restriction site analysis (FIG. 5)
Once the chimeric gene was constructed, it could be shuttled into
appropriate prokaryotic or eukaryotic expression vectors via Hind III
sites on both termini. For expression in prokaryotes, the anti-human SC
Fv/human A.sub.1 AT chimeric gene was excised by digestion with Hind III,
and ligated into the plasmid pQE-30 (Qiagen Inc., Chatsworth, Calif.). The
gene is driven by the E. coli phage T5 and two lac operon sequences to
eliminate expression prior to induction with
isopropyl-.beta.-D-thiogalactopyranoside (IPTG). This vector also contains
a ribosome binding site and ampicillin resistance gene. A sequence
encoding a polyhistadine (HHHHHH) label is located upstream in this vector
that permits the identification and purification of the translated
recombinant protein on a nickel-NTA resin column (47). Both prokaryotic
and eukaryotic expression plasmids containing the cDNA encoding the
anti-human SC and human A.sub.1 AT alone were constructed, and a fusion
protein containing an irrelevant anti-D8 Fv fragment ligated to the human
A.sub.1 AT was produced as control.
In vitro transcription and translation of the chimeric gene was performed
using a rabbit reticulocyte lysate system to determine if the chimeric
gene encoding anti-human SC Fv/human A.sub.1 AT could be expressed.
Messenger RNA encoding the chimeric genes was generated by transcribing
the expression plasmid with T7 RNA polymerase, which was then translated
in reticulate lysates using a coupled TNT system (Promega, Madison,
Wisconsin). The synthesized proteins were radiolabeled by adding [.sup.35
S] methionine to the translation reaction. Analysis of the lysates by
electrophoresis in SDS-polyacrylamide gels demonstrated the presence of
the appropriately sized proteins for the anti-human SC Fv, human A.sub.1
AT, and fusion protein (FIG. 6).
Example 8
Production of Anti-human Secretory Component Fv/human Alpha.sub.1
-antitrypsin Fusion Protein in Prokaryotes
Most single chain Fv constructs have been produced in bacteria because
prokaryotes can produce large quantities of the fusion protein. E. coli
strain M15[pREP4] was transduced with a plasmid containing the anti-human
SC Fv/human A.sub.1 AT chimeric gene driven by the E. coli phage T5, and
ampicillin resistant clones were selected in LB media containing 100 mg/ml
ampicillin. Protein extracted from isolated inclusion bodies of
transformed bacterial clones were purified by nickel-chelate affinity
chromatography, exploiting the polyhistadine tag, subjected to
electrophoresis in SDS-polyacrylamide gels, and transferred onto
nitrocellulose membrane filters using established methods. The protein
products were identified by Western blot analysis using antibodies
directed against the individual components, i.e., polyhistadine tag,
single chain Fv fragments, and human A.sub.1 AT, which demonstrated the
same intact, non-glycosylated fusion protein (FIG. 7). In order to recover
function since protein from cytoplasmic inclusion bodies are inactive, the
Fv-based fusion proteins was to be solubilized and renatured by diluting
resultant fusion proteins (48). We have tried several refolding
techniques, and dilution appears to be the most effective approach of
renaturing the fusion protein produced in prokaryotes, although the
efficiency of the process is uncertain.
Large-scale production of the fusion proteins will be critical for
experiments examining the transport through the pIgR in vitro and in vivo.
Although prokaryotes can produce large quantities of recombinant proteins,
Fv-based fusions produced by bacteria needs to be solubilized and
renatured in order to recover function since proteins purified from
cytoplasmic inclusion bodies are inactive due to the harsh conditions
necessary to solubilize them (48). Human A.sub.1 AT has been successfully
made by E. coli and it retains protease inhibition (49), but
non-glycosylated forms of the antiprotease has an extremely short
half-life in the circulation, decreasing the antiprotease half-life to
hours (50). The shortened lifespan of the recombinant human A.sub.1 AT may
not matter so much for our fusion proteins, since seventy-five percent of
the anti-SC antibody is cleared from the circulation in one hour. Thus, we
are confident that the single chain anti-human SC Fv/human A.sub.1 AT
fusion produced by prokaryotes will be adequate for experiments in vitro,
but may have limited usefulness in vivo.
Example 9
Recognition of the Human Secretory Component by Anti-human Secretory
Component Fv
ELISA was used to demonstrate that the single chain Fv portion of the
fusion protein recognizes human SC. Human SC was incubated in each well of
a 96 well microtiter plate, washed, and blocked. The fusion proteins
synthesized by bacteria were isolated and refolded using standard
techniques, then added to the wells. After washes, a rabbit-derived,
polyclonal anti-human A.sub.1 AT and goat-derived, anti-rabbit
(Fab').sub.2 conjugated with horseradish peroxidase were applied
sequentially and used to detect the antiprotease portion of the protein.
The fusion proteins containing the anti-human SC Fv effectively bound to
the sIgA, whereas fusions containing an irrelevant (anti-D8) Fv did not
(FIG. 8). Because the Fv portion of the molecule recognizes human SC, the
fusion protein used in later experiments could be purified by affinity
column chromatography.
Thus, transformed bacteria can yield fusion proteins that recognize human
SC, so the single chain Fv component is functional. This protein is also
recognized by antibodies directed against both parts of the bifunctional
protein. Specifically, a commercially-available anti-human A.sub.1 AT
antibody and an antibody we prepared against the framework regions covered
by the oligonucleotide primers and used to generate the single chain Fv
fragments. The polyhistadine tag from the expression vector was recognized
by the appropriate antibody, so all three immunologically recognizable
components of the fusion are present.
Example 10
Production of Anti-human Secretory Component Fv/human
Alpha.sub.1,--antitrypsin Fusion Protein in Eukcayotic Cells
On a theoretical basis, prokaryotes may not be the best expression system,
since they ail to process and glycosylate mammalian proteins
appropriately. For A.sub.1 AT, glycosylati | | |
