Biochem. J. (1991) 273, 497-502
`
`(Printed in Great Britain)
`
`497
`
`A Fischer rat substrain deficient in dipeptidyl peptidase IV activity
`makes normal steady-state RNA levels and an altered protein
`Use as a liver-cell transplantation model
`
`Nancy L. THOMPSON,*§ Douglas C. HIXSON,* Helen CALLANAN,* Marilyn PANZICA,*
`Donna FLANAGAN,* Ronald A. FARIS,* Wanjin HONG,tII Sabine HARTEL-SCHENK$T
`and Darrell DOYLEt
`*Department of Medical Oncology, Rhode Island Hospital/Brown University, Providence, RI 02903, U.S.A.,
`tDepartment of Biological Sciences, State University of New York at Buffalo, Buffalo, NY 14260, U.S.A.,
`and tDepartment of Molecular Biology and Biochemistry, Freie Universitat Berlin, D-1000 Berlin 33, Germany
`
`Dipeptidyl peptidase IV (DPPIV) is a serine exoproteinase expressed at high levels in epithelial cells of kidney, liver and
`small intestine. Recently Watanabe, Kohima & Fujimoto [(1987) Experientia 43, 400-401] and Gossrau et al. [(1990)
`Histochem. J. 22, 172-173] reported that Fischer 344 rats are deficient in this enzyme. We have examined DPPIV
`expression in Fischer 344 rats available from U.S. and German suppliers and find that livers of the U.S. Fischer rats, in
`contrast with their German counterparts, express active DPPIV (D +). Northern analysis of liver RNA showed
`comparable levels of 3.4 kb and 5.6 kb DPPIV transcripts in both D + rats from the U.S. and German (D -) rats.
`Monoclonal antibody (MAb) 236.3 to DPPIV immunoprecipitated a 150 kDa enzymically active (105 kDa, denatured)
`protein from surface-labelled D + hepatocytes and reacted with canalicular and sinusoidal membranes (as shown by
`immunofluorescence microscopy). MAb 236.3 failed to immunoprecipitate a labelled peptide from D - cell extract or to
`stain D - liver sections. Polyclonal antibody (PAb) specific for DPPIV immunoprecipitated an enzymically active peptide
`from D + hepatocyte extracts and a smaller, inactive peptide from D - hepatocyte extracts. Peptide maps of DPPIV
`immunoprecipitated from D + extracts with MAb 236.3 and PAb were identical, but differed from that of the D -
`hepatocyte component recognized by PAb. The molecular basis of the DPPIV deficiency in the D - rats thus appears to
`be the translation of an enzymically inactive protein missing the epitope recognized by MAb 236.3. We have exploited
`these D - rats as hosts for syngeneic transplantation of liver cells from D + Fischer rats. DPPIV expression is stable in
`the transplanted cells and allows them to be readily distinguished from the surrounding D - tissue.
`
`INTRODUCTION
`Dipeptidyl peptidase IV (DPPIV) (EC 3.4.14.5) is a serine
`exoproteinase which cleaves dipeptides in vitro from the N-
`terminus of oligopeptides with a penultimate proline or alanine
`residue (Hopsu-Havu & Glenner, 1966). Physiologically reactive
`substrates in vivo are not known, but potential substrates
`identified by sequence include substance P, kentsin and inter-
`leukin 2 (Hartel et al., 1988a). The localization of DPPIV in the
`extracellular matrix of liver as well as on the hepatocyte cell
`surface (Walborg et al., 1985), and its reported binding to
`collagen (Bauvois, 1988; Hanski et al., 1988) and fibronectin
`(Piazza et al., 1989), have suggested that one of its functions
`in vivo may be in cell-matrix interactions. However, high tissue
`concentrations of DPPIV are found in kidney and small-intestine
`brush-border membranes, hepatocyte bile-canalicular mem-
`branes and on other cell types (Hopsu-Havu & Ekfors, 1969;
`Gossrau, 1981) that are not in contact with extracellular matrix,
`suggesting that additional functions exist for this enzyme.
`Watanabe et al. (1987) reported that the activity of membrane-
`bound DPPIV was markedly reduced in kidney membranes from
`Fischer 344 rats compared with Wistar strain rats, whereas the
`other peptidase activities examined were equivalent between the
`
`strains. Tiruppathi et al. (1990a,b) recently showed that a
`deficiency of renal DPPIV is characteristic of Fischer rats from
`Japan, but not those obtained from three different U.S. suppliers.
`Moreover, Gossrau et al. (1990) demonstrated that tissues from
`Fischer 344 rats obtained in Germany also lack histochemically
`detectable DPPIV and fail to react in immunohistochemical
`assays with either polyclonal antisera or monoclonal antibodies
`to DPPIV. We have examined the expression of DPPIV in livers
`of Fischer 344 rats obtained from both U.S. and German
`suppliers to determine the molecular basis for the enzyme
`deficiency and have transplanted cells from the DPPIV-positive
`(D +) to DPPIV-negative (D -) rats to determine the stability of
`the phenotype and its utility as a syngeneic transplantation
`model.
`EXPERIMENTAL
`Animals and procurement of cells and tissues
`Fischer 344 adult male rats were obtained in the U.S.A. from
`Charles River, Wilmington, DE, U.S.A. German Fischer 344
`adult male and female rats were obtained from Charles River
`Wiga, Muinchen, Germany. All animals were maintained on
`normal rat chow ad libitum under standard conditions. All
`
`Abbreviations used: DPPIV, dipeptidyl peptidase IV; MAb, monoclonal antibody; PAb, polyclonal antibody; D+, Fischer 344 rats from U.S.
`suppliers with DPPIV enzymic activity; D-, Fischer 344 rats from German supplier deficient
`in DPPIV activity; AFC, amino-4-
`trifluoromethylcoumarin; MNA, 4-methoxy-,8-naphthylamide; PBS, phosphate-buffered saline; poly(A)+, polyadenylated; GAPDH, glyceraldehyde-
`3-phosphate dehydrogenase; RIP, radioimmunoprecipitation.
`§ To whom correspondence should be addressed.
`11 Present address: Institute of Molecular and Cell Biology, National University of Singapore, Singapore 0511.
`T Present address: Johns Hopkins University School of Medicine, Baltimore, MD 21205, U.S.A.
`
`Vol. 273
`
`SAXA-DEF-00017
`
`Page 1 of 6
`
`AstraZeneca Exhibit 2021
` Mylan v. AstraZeneca
` IPR2015-01340
`
`

`
`498
`
`procedures involving animals were carried out in accordance
`with the policies and guidelines for proper care and humane
`treatment of research animals of the Rhode Island Hospital/
`Brown University Animal Care and Use Committee. In exper-
`iments comparing cells and tissues from the U.S. and German
`rats, care was taken to use male animals of approximately the
`same age. Hepatocytes were obtained by a modification (Hixson
`et al., 1983) of the collagenase perfusion technique of Bonney
`et al. (1974). Viability of the isolated hepatocytes was greater than
`80 % by Trypan Blue dye exclusion. Tissues for RNA extraction
`and frozen sections were flash-frozen in liquid N2 immediately
`after dissection and stored at -70 'C.
`RNA extraction and Northern hybridization
`Total RNA was prepared as described by Chirgwin et al.
`(1979). Polyadenylated [Poly(A)+] RNA was prepared from total
`RNA by oligo(dT)-cellulose chromatography (Davis et al., 1986).
`Aliquots [10,ug of total or 4 csg of poly(A)+] of RNA were
`separated by agarose/formaldehyde-gel electrophoresis, trans-
`ferred to Nytran (Schleicher and Schuell, Keene, NH, U.S.A.)
`filters and hybridized as described previously (Thompson et al.,
`1989). The 2.0 kb endonuclease-EcoRI fragment of the Ac2
`cDNA clone of rat DPPIV (Hong & Doyle, 1987) and the
`1.25 kb PstI fragment of glyceraldehyde-3-phosphate dehydro-
`genase (GAPDH) (Fort et al., 1985) were labelled for use as
`probes with [32P]dCTP (New England Nuclear, Boston, MA,
`U.S.A.; 3000 Ci/mmol) using a random primer labelling kit
`(Boehringer Mannheim Biochemicals, Indianapolis, IN, US.A.).
`Antibodies and immunochemical protocols
`Monoclonal antibody (MAb) 236.3 specific for rat DPPIV has
`been described (Hixson et al., 1984; Walborg et al., 1985). The
`IgG fraction of a rabbit antibody raised against purified rat liver
`DPPIV (Hartel et al., 1987) was used for immunoprecipitation
`analysis.
`labelling
`of hepatocytes
`1251
`Na'25I;
`Surface
`with
`(as
`Amersham, Arlington Heights, IL, U.S.A.; 500mCi/ml) was
`performed by the lactoperoxidase/glucose oxidase procedure of
`Keski-Oja et al. (1977). Procedures for the immunoprecipitation
`of radiolabelled antigens using Staphylococcus aureus (IgGsorb;
`Enzyme Center, Boston, MA, U.S.A.) have been described
`(Hixson et al., 1983). One-dimensional SDS/PAGE was per-
`formed as described by Laemmli (1970). One-dimensional peptide
`mapping was performed by digesting gel slices containing labelled
`bands with S. aureus V8 proteinase (Miles Scientific, Naperville,
`IL, U.S.A.) (Cleveland, 1983) as described previously (Hixson
`et al., 1985). Apparent molecular masses were calculated from
`prestained protein standards (myosin, 200 kDa; fl-galactosidase,
`116 kDa; phosphorylase b, 97.4 kDa; BSA, 66.2 kDa; oval-
`bumin, 42.7 kDa; Diversified Biotech, Hyde Park, MA, U.S.A.)
`run concurrently with radioactively labelled samples.
`Comparison of the reactivity of two different antisera was
`determined by immunodepletion analysis (Hixson et al., 1985).
`extracts were sequentially
`aliquots of 1251I-labelled
`Briefly,
`immunoprecipitated with the first antibody until SDS/PAGE
`analysis indicated that all reactive components had been removed.
`The immunodepleted extracts were then immunoprecipitated
`with the second antibody and the immunoprecipitates analysed
`by SDS/PAGE. If the second antibody showed no reactivity
`with the depleted extract, this was interpreted as evidence that
`the first antibody recognized all components reactive with the
`second.
`Immunoadsorption experiments were performed by incubating
`5 x 101 hepatocytes in 0.5 ml of anti-DPPIV antisera for 2 h at
`4 'C. Hepatocytes with adsorbed antibody were then recovered
`by centrifugation at 100 g, washed three times in phosphate-
`
`N. L. Thompson and others
`
`buffered
`(PBS;
`saline
`140 mM-NaCl/4 mM-KCl/2 mm-
`KH2PO4/20 mM-Na2HPO4, pH 7.4) and finally resuspended in
`0.5 ml of0.1 M-citrate buffer, pH 2.2, for 10 min to elute adsorbed
`antibody. After neutralization to pH 7.4 with 1 M-Tris, the eluted
`antibodies were used for immunoprecipitation analysis.
`Immunofluorescence was performed on frozen sections of rat
`liver, 4-6,um thick, after fixation in cold (0 °C) acetone for
`10 min. Incubation with MAb 236.3 to DPPIV was followed by
`fluorescein-conjugated affinity-purified goat anti-mouse IgG as
`previously described (Hixson et al., 1983). Sections were ex-
`amined on a Nikon Microphot-FX fluorescence microscope
`equipped with an epifluorescence illuminator.
`
`Assay of DPPIV activity
`DPPIV was identified after SDS/PAGE in unfixed gels by its
`enzymic activity using an overlay technique (Smith, 1984) in
`which cellulose acetate strips impregnated with the fluorogenic
`substrate Ala-Pro-amino-4-trifluoromethylcoumarin (Ala-Pro-
`AFC) (Enzyme Systems Products, Livermore, CA, U.S.A.) were
`placed over the gels, incubated 30 min at 37 °C, and then observed
`and photographed under u.v. light (Piazza et al., 1989). DPPIV
`activity in acetone-fixed frozen liver sections was assayed by
`histochemical staining (Lojda et al.,
`1979), using Gly-Pro-
`methoxy-,J-naphthylamide (Gly-Pro-MNA) (Sigma, St. Louis,
`MO, U.S.A.) as substrate.
`Transplantation studies
`Freshly isolated U.S.-Fischer-rat hepatocytes were trans-
`planted into either the liver or pancreas of German Fischer rats.
`Briefly, an abdominal incision was made under metophane
`anaesthesia to expose the liver or splenic portion of the pancreas
`and 5 x 106 donor hepatocytes (in 0.5 ml of Hanks balanced salt
`solution) were slowly injected beneath the capsule of each organ
`with a 1 ml syringe and a 22-gauge needle. Leakage of cells was
`minimized by applying a Gelfoam sponge (Upjohn, Kalamazoo,
`MI, U.S.A.) to the site of injection. After transplantation of cells,
`the incision was closed with silk suture and stainless-steel clips
`and the animals were monitored carefully during recovery.
`Recipient rats were killed by CO2 asphyxiation at 7 and 17 days
`after surgery, and tissues were excised and rapidly frozen in
`hexane chilled by a solid-CO2/acetone bath. Transplanted donor
`cells were detected in frozen tissue sections by indirect immuno-
`fluorescence or immunohistochemical assays for DPPIV.
`
`RESULTS
`Localization of DPPIV in Fischer-rat liver
`DPPIV enzyme histochemistry on acetone-fixed frozen sections
`of liver from U.S.Fischer rats revealed dark-red reaction product
`abundant along the bile-canalicular membranes of adjacent
`hepatocytes, as well as weaker staining of the sinusoidal mem-
`branes and bile ducts (Fig.
`la). Indirect immunofluorescence
`microscopy with MAb 236.3 on duplicate sections of U.S.-
`Fischer-rat liver gave a staining pattern identical with the
`histochemical result (Fig.
`lc). By contrast, no histochemical
`reaction product was detected in sections of liver from German
`Fischer rats (Fig. lb), confirming the work of Gossrau et al.
`(1990). In addition, MAb 236.3 failed to react in indirect
`immunofluorescence microscopy with this tissue (Fig. ld). Sec-
`tions of German-Fischer-rat liver stained without fixation also
`gave a negative result (not shown), indicating that sensitivity of
`the protein to fixation was not responsible for the absence of
`staining and suggesting that either the D - German Fischer rat
`does not have the enzyme or it produces an inactive form of
`DPPIV that is not recognized by MAb 236.3.
`
`1991
`
`SAXA-DEF-00018
`
`Page 2 of 6
`
`

`
`Molecular mechanism for dipeptidyl peptidase IV deficiency
`
`499
`
`(a) Liver
`
`(b) Tissue comparison
`
`1 2 3 4 5 6
`p
`
`DPPIV
`5.6 kb--
`3.4 kb-AStm lo
`
`to
`
`GAPDH
`1.4kb-
`
`?*
`
`1
`
`DPPIV
`
`1 2 3 4 5 6 7 8
`
`3.4 k
`
`,:A
`
`EthBr
`
`f
`j
`
`-28 S
`-18 S
`1
`
`Fig. 2. Northern-blot analysis of DPPIV RNA expression in tissues from
`U.S. and German Fischer rats
`A portion (10 /sg) of total RNA [lanes 1, 2, 4 and 5 (a) and lanes 1-8
`(b)] or 4 ,ug of polyadenylated RNA (pA +, lanes 3 and 6, a) was
`electrophoresed on agarose/formaldehyde gels containing ethidium
`bromide (EthBr), blotted to Nytran, and hybridized to 32P-labelled
`cDNA probe for DPPIV. The blot shown in (a) (top) was stripped
`of DPPIV probe and rehybridized to GAPDH probe (bottom left).
`The blot shown in (b) was photographed under u.v. light (bottom
`right) before hybridization with DPPIV probe in order to reveal the
`RNA. The positions of the 28 S and 18 S ribosomal RNAs are
`indicated. RNAs extracted from tissues of U.S. Fischer rats are
`indicated by '+', whereas those from German Fischer rats are
`indicated by ' -'. Tissue sources are as follows: (a) lanes 1-6, liver;
`(b) lanes 1 and 2, kidney; lanes 3 and 4, small intestine; lanes 5 and
`6, liver; lanes 7 and 8 heart. (a) Samples 1 and 2 are duplicate RNA
`preparations from different D + animals, whereas samples 4 and 5
`are duplicate preparations from different D- animals.
`
`exposures of total RNA, particularly from kidney (result not
`shown).
`Radioimmunoprecipitation of DPPIV from surface-labelled
`hepatocytes and S. aureus V8-proteinase peptide maps
`To examine further the expression of DPPIV on D + and D -
`hepatocytes, isolated viable cells from both substrains were
`surface labelled with 125I and used to prepare detergent extracts
`for immunoprecipitation analysis with monoclonal and poly-
`clonal (PAb) anti-DPPIV antibodies. As shown in Fig. 3a, both
`MAb 236.3 and PAb immunoprecipitated a 105 kDa band from
`extracts of 1251I-labelled D + hepatocytes (Fig. 3a, lanes 1 and 2),
`but PAb showed no reactivity when extracts were first immuno-
`depleted with MAb 236.3 (lane 3), thus confirming the identity of
`the component recognized by PAb as DPPIV. If immuno-
`precipitates from D + hepatocyte extracts were separated by
`PAGE without boiling, primarily the 150 kDa monomeric from
`of DPPIV was seen (lane 4). As expected from its lack of
`reactivity on frozen sections of D- liver, MAb 236.3 failed to
`immunoprecipitate any labelled component from extracts of D -
`hepatocytes (lane 5).
`To determine if D - rats expressed an enzymically inactive
`form of DPPIV, a two-step analysis was used. The first step
`involved determining whether D - hepatocytes were able to bind
`specifically anti-DPPIV antibodies from anti-DPPIV antiserum
`which could immunoprecipitate enzymically active DPPIV from
`detergent extracts of surface-labelled D + hepatocytes. As shown
`in Fig. 3(b), antibodies eluted from the surface of both D +
`(lanes 1 and 5) and D - (lanes 2 and 6) hepatocytes were capable
`of immunoprecipitating a component identical in size with
`DPPIV immunoprecipitated with MAb 236.3 (lanes 3 and 4).
`When the immunoprecipitates were boiled before electrophoresis,
`the characteristic 105 kDa denatured band was seen (Fig. 3b,
`lanes 1-3), whereas primarily the 150 kDa monomer was seen in
`
`Fig. 1. Antibody and histochemical localization of DPPIV in livers of U.S.
`and German Fischer rats
`(a) and (b) Indirect immunofluorescence microscopy of acetone-
`fixed frozen liver sections with MAb 236.3 localizes DPPIV to the
`hepatocyte bile-canalicular membrane (long arrow) with lighter
`sinusoidal membrane domain staining (short arrow) in the U.S.
`Fischer rat (a). No antibody reactivity is seen in the German Fischer
`rats (b). Magnification x 150. (c) and (d) DPPIV histochemical
`staining of acetone-fixed frozen liver sections shows abundant
`dipeptidase reaction product localized in the bile-canalicular region
`(long arrow) and slight reactivity along the sinusoids (short arrow)
`of the U.S. Fischer rats (c). No enzyme activity is seen in the
`German Fischer rats (d). Magnification x 160. The bar represents
`100 ,um.
`
`DPPIV RNA expression
`Northern-blot hybridization of total RNA isolated from livers
`of D + and D - Fischer rats showed that the 3.4 kb DPPIV
`mRNA was expressed by both and that steady-state levels
`were approximately equal between them (Fig. 2a). Isolation of
`poly(A)+ RNA from these livers demonstrated that the 3.4 kb
`species as well as a novel 5.6 kb RNA species were polyadenylated
`in both strains (Fig. 2a), suggesting a non-transcriptional mech-
`anism for the loss of DPPIV activity in the D - strain. Hybrid-
`ization of the same blots to probe for the glycolytic enzyme
`GAPDH was used to standardize DPPIV expression by densito-
`metry ratios of the resulting autoradiographs, since the former
`'housekeeping' gene should be expressed at equivalent levels in
`both strains. When this calculation was made, steady-state levels
`of DPPIV mRNA in the D- livers were found to be 1.1-1.5
`times those of D+ livers. In addition to liver, DPPIV RNA
`levels were compared in several other tissues from these rats (Fig.
`2b). Confirming previous observations in Sprague-Dawley-rat
`tissues (Hong et al., 1989 a, b), the highest levels of the 3.4 kb
`transcript were found in kidney, followed by small intestine and
`liver, with only barely detectable levels found in heart (Fig. 2b).
`The 5.6 kb RNA was also apparent in long autoradiographic
`
`Vol. 273
`
`SAXA-DEF-00019
`
`Page 3 of 6
`
`

`
`500
`
`N. L. Thompson and others
`
`Ab ... P M P/MM M
`RIP..
`+ + +
`-
`
`1 50 kDa
`105 kDa
`
`(a)
`
`1
`
`2
`
`:3 4 5
`
`Eluted... + - C C + -
`-+ -+
`RIP... +
`+ +
`
`P P M P/M
`Ab...
`RIP... + - --
`
`150 kDa
`105 kDa
`
`4
`
`105 kDa
`95 kDa
`
`(b)
`
`1
`
`2
`
`3 4 5 6
`
`(d)
`
`1 2 3 4
`
`El uted . . . C ±
`RIP... + + t
`
`150
`
`U
`
`(c)
`
`1 2 3
`
`P M P/M M
`Ab ...
`P
`Band (kDa) ... 105 95105 95 150
`
`43 kDa
`
`31 kDa
`
`22 kDa
`
`(e )S....
`
`1
`
`2
`
`3
`
`4
`
`5
`
`Fig. 3. Radioimmunoprecipitation (RIP) analysis of DPPIV
`(a) Immunoprecipitation of DPPIV by monoclonal and polyclonal antibodies from extracts of surface-labelled hepatocytes. Hepatocytes were
`isolated by collagenase perfusion, surface-labelled with 125I and used to prepare detergent extracts for RIP analysis with either polyclonal (P) or
`monoclonal (M) antibody specific for DPPIV. Both antibodies immunoprecipitate a 105 kDa band from surface-labelled U.S.-Fischer (+)-rat
`hepatocytes when samples are boiled before electrophoresis (lanes 1 and 2). No DPPIV could be immunoprecipitated with PAb from D + extracts
`immunodepleted by MAb (P/M) (lane 3). Unboiled samples show primarily the 150 kDa monomeric form of DPPIV (lane 4). MAb fails to
`immunoprecipitate a band from extracts of German-Fischer (-)-rat hepatocytes (lane 5). (b) RIP of DPPIV from extracts of U.S.-Fischer-rat
`hepatocytes by PAbs bound by, and eluted from, the surface of U.S.- and German-Fischer-rat hepatocytes. Hepatocytes (5 x 106) were incubated
`with 0.5 ml of PAb for 2 h at 4 °C, washed three times in PBS, and bound antibody was eluted in 0.1 M-citrate, pH 2.2. Eluted antibody was used
`for immunoprecipitation of DPPIV from extracts of U.S.-Fischer-rat hepatocytes. Lanes 1 and 5, RIP with antibody eluted from U.S.-(+)-
`Fischer-rat hepatocytes; lanes 2 and 6, RIP with antibody eluted from German-(-)-Fischer-rat hepatocytes. Lanes 3 and 4, control (C) RIPs with
`MAb 236.3. Samples in lanes 1-3 were boiled, where those in lanes 4-6 were unboiled. The arrow denotes the position of approx. 300 kDa DPPIV
`dimer visible in lane 6. (c) Dipeptidase activity of immunoprecipitates by substrate/gel-overlay technique. Lanes 4-6 (unboiled RIPs) of the gel
`shown autoradiographed in (b) were overlaid with a sheet of cellulose acetate impregnated with the DPPIV fluorogenic substrate Ala-Pro-AFC.
`Enzyme activity is shown by the white bands visible on the cellulose acetate under u.v. light corresponding to the 150 kDa components
`immunoprecipitated from hepatocyte extracts. Lane 1, control RIP with MAb 236.3; lane 2, RIP with PAb eluted from U.S.-(+)-Fischer-rat
`hepatocytes; lane 3, RIP and PAb eluted from German-Fischer-(-)-rat hepatocytes. (d) RIP of a 95 kDa band from extracts of German-(-)-
`Fischer-rat hepatocytes by PAb. In contrast with the 105 kDa band immunoprecipitated from extracts of U.S.-(+)-Fischer-rat hepatocytes (lane
`1), PAb immunoprecipitated a 95 kDa band from extract of German-(-)-Fischer-rat hepatocytes (lane 2). MAb failed to immunoprecipitate a
`labelled band from D- extracts (lane 3), and D- extracts immunodepleted with MAb still contained the 95 kDa component (lane 4). (e) V8-
`proteinase maps of immunoprecipitated DPPIV and 95 kDa component. Bands immunoprecipitated by DPPIV antibody in (b) and (d) were sliced
`from dried gels, digested with V8 proteinase, and separated by electrophoresis on 12.5 % gels, which were then dried and autoradiographed to
`produce the peptide maps shown. Lanes 1, 2 and 4 are peptide maps of the labelled bands immunoprecipitated with PAb in lanes 1, 2 and 4
`respectively of (d). Lanes 3 and 5 correspond to V8 digestions of the 105 kDa and 150 kDa bands of DPPIV immunoprecipitated with MAb 236.3
`in (b), lanes 3 and 4 respectively. The position of molecular-mass markers are shown on the left.
`
`unboiled immunoprecipitates (Fig. 3b, lanes 4-6), with some
`higher-molecular-mass material, probably representing dimer,
`observed as well (lane 6, arrowhead). The identity of these
`150 kDa bands as DPPIV was confirmed by substrate overlay of
`the gel containing the unboiled samples (Fig. 3c). DPPIV activity
`was associated with the 150 kDa bands immunoprecipitated by
`the bound and eluted PAb (Fig. 3c, lanes 2 and 3) as well as the
`MAb 236.3 control (Fig. 3c, lane 1). Having thus established the
`presence of a cross-reacting component on D- hepatocytes,
`direct immunoprecipitation analysis of 125I-labelled D - hepato-
`cyte extracts with PAb was done (Fig. 3d). In contrast to the
`105 kDa band immunoprecipitated from D + hepatocyte extracts
`(Fig. 3d, lane 1), PAb immunoprecipitated a 95 kDa component
`from D - hepatocytes (Fig. 3d, lane 2). This component was not
`removed by prior depletion with MAb 236.3 (Fig. 3d, lanes 3 and
`
`4), indicating that the reactive epitope was either not present or
`not exposed on the 95 kDa protein.
`To explore further the relationship of the 95 kDa band to the
`105 and 150 kDa forms of DPPIV, S. aureus-V8-proteinase
`peptide maps were generated from each labelled band and
`compared (Fig. 3e). DPPIV immunoprecipitated from U.S.
`Fischer rats by both the polyclonal (Fig. 3e, lane 1) and
`monoclonal (Fig. 3e, lanes 3 and 5) antibodies gave identical
`peptide maps no matter whether boiled (Fig. 3e, lanes 1 and 3)
`or unboiled (Fig. 3e, lane 5) immunoprecipitates were used.
`However, the V8 peptide map obtained for the 95 kDa com-
`ponent from the D - hepatocytes (Fig. 3e, lanes 2 and 4) differed
`in that it lacked a major labelled peptide of approx. 17 kDa, did
`not display labelled peptides larger than approx. 32 kDa, and
`contained a doublet of labelled bands at about 32 kDa, somewhat
`
`1991
`
`SAXA-DEF-00020
`
`Page 4 of 6
`
`

`
`Molecular mechanism for dipeptidyl peptidase IV deficiency
`
`501
`
`(Fig. 4a), indicating that DPPIV expression was stable in the
`transplanted cells. Similarly, the D- phenotype of the host
`appeared to be stable, since there was no indication ofenzymically
`active DPPIV in the surrounding tissue. No infiltration of
`lymphoid cells was seen in the area, suggesting that an immune
`response was not raised against the transplanted cells. Hepato-
`cytes from the D + rat could also be detected after transplantation
`of the liver of D - hosts by immunofluorescent staining with
`MAb 236.3 (Fig. 4b). Colonies of cells positive for this antibody
`and with characteristic canalicular staining were clearly dis-
`tinguishable in livers ofhost animals 1 week after transplantation.
`
`DISCUSSION
`We have investigated the molecular basis for the lack of
`DPPIV activity in livers of German Fischer rats. As reported
`recently (Tiruppathi et al., 1990a, b), renal membranes of Fischer
`344 rats from Japanese suppliers are deficient in DPPIV activity,
`whereas renal membranes from Fischer rats available from U.S.
`suppliers possess normal levels ofthis enzyme. Similarly, Gossrau
`et al. (1990) found that kidney, as well as other tissues of Fischer
`rats from Germany, also lack this enzyme, as demonstrated by
`histochemical and antibody reactivity, whereas levels of other
`membrane-bound enzymes examined were normal. We have
`confirmed and extended these findings to show that, despite the
`lack of functional enzyme, German Fischer rats have comparable
`steady-state levels of the 3.4 kb DPPIV transcript found in U.S.
`Fischer rats, as well as a previously unreported 5.6 kb RNA
`species hybridizing with DPPIV cDNA. The presence of the
`5.6 kb transcript for DPPIV was not reported by Hong et al.
`(1989a,b), who originally cloned DPPIV as gpl 10 from a rat
`kidney expression library, most likely because the RNA prepar-
`ations used for Northern hybridizations in those studies were
`total RNA and, in our experience, the larger transcript is much
`less abundant in total RNA as against poly(A)+-selected RNA.
`The relationship between the two transcripts is not presently
`known. DPPIV has been reported to be a single-copy gene (Hong
`et al., 1989a). Independent cDNA clones have been obtained
`from a rat liver library by Ogata et al. (1989), which vary from
`the kidney clones obtained by Hong et al. (1989a,b) in five amino
`acid residues in the coding region and which, owing to a
`frameshift caused by one of these residues, limits the open
`reading frame from 792 residues in gpl 10 to 767 in DPPIV. In
`addition, Ogata et al. (1989) reported the presence of more than
`one poly(A) signal, which may imply polymorphism in the 3'
`non-coding region of the mRNA. Alternatively, the mRNAs
`may initiate at distinct promoters on a single gene, as was
`recently found for another rat plasma-membrane ectoenzyme,
`y-glutamyl transpeptidase (Chobert et al., 1990).
`Lack of active DPPIV in hepatocyte membranes was demon-
`strated by histochemical assay in frozen sections of German-
`Fischer-rat liver using specific substrate as well as by immuno-
`precipitation with antibodies to DPPIV and substrate-overlay
`assay of the undenatured immunoprecipitated material separated
`by SDS/PAGE. We found that surface-labelled hepatic mem-
`branes from German Fischer rats lack the 105 kDa band,
`corresponding to DPPIV monomer, that is found in membranes
`of U.S. Fischer rats. Tirrupathi et al. (1990a) reported similar
`findings for renal membrane proteins of Japanese Fischer rats
`separated by SDS/PAGE. All other membrane proteins com-
`pared between hepatocytes of the German and U.S. Fischer rats
`were present in equal abundance, indicating that the DPPIV
`deficiency is not due to a generalized aberrant processing and
`transport of cell-surface glycoproteins. Deficiencies of another
`proteinase associated with kidney brush-border membranes,
`merpin, as well as the production of an immunologically related
`
`Fig. 4. Transplantation system
`(a) Colonies of U.S.-Fischer-rat hepatocytes detected by DPPIV
`histochemical staining 17 days after transplantation to the pancreas
`of a German Fischer rat. Histochemical staining of acetone-fixed
`frozen sections shows that the transplanted U.S.-Fischer-rat hepato-
`cytes visible as colonies in the pancreas are positive for DPPIV
`activity (white histochemical reaction product), whereas the sur-
`rounding pancreatic tissue of the German Fischer rat is negative.
`Phase contrast; magnification x 185. (b) Immunofluorescent localiz-
`ation of U.S.-Fischer-rat hepatocytes 7 days after transplantation to
`the liver of a German Fischer rat. Indirect immunofluorescence on
`acetone-fixed frozen sections with MAb 236.3 reveals a colony of
`positive-staining cells with the characteristic canalicular DPPIV
`localization visible against the surrounding negative tissue. Magnifi-
`cation x 320. The bar represents 50 ,um.
`
`larger than a similar doublet seen in V8 maps of DPPIV from
`D + hepatocytes.
`
`Transplantation system
`The availability of(1) syngeneic rat substrains with and without
`active DPPIV and (2) histochemical and immunocytochemical
`assays to detect DPPIV suggested that these rats could be
`employed as a transplantation system to examine the stability of
`DPPIV expression as well as the fate of the small number of D +
`normal or preneoplastic cells transplanted to D - hosts. To
`investigate this possibility, hepatocytes isolated from D + rats
`were transplanted beneath the capsule of liver and pancreas of
`D - rats. Colonies of DPPIV-histochemically-positive hepato-
`cytes were found in the vicinity of the injection site in the
`pancreas of a German Fischer rat 17 days after transplantation
`
`Vol. 273
`
`SAXA-DEF-00021
`
`Page 5 of 6
`
`

`
`502
`
`altered protein, have been previously demonstrated in some
`inbred mouse strains (McKay et al., 1985).
`The 95 kDa band immunoprecipitable from extracts of surface-
`labelled D - hepatocytes by polyclonal antibody may represent
`an inactive truncated form of the enzyme. This band is not seen
`on autoradiographs of immunoprecipitates from D + extracts
`and thus cannot be attributed to another cross-reactive surface
`molecule. A DPPIV-immunoreactive species on the surface of
`D- hepatocytes is also indicated by the ability of these cells to
`bind antibodies which can then specifically immunoprecipitate
`enzymically active DPPIV from D+ hepatocyte extracts. The
`nature of this molecule is presently not clear. V8 peptide maps of
`the 95 kDa band show that it clearly differs from DPPIV
`immunoprecipitated from D+ hepatocytes. MAb 236.3 is not
`directed against the active site of the enzyme, since it does not
`decrease DPPIV activity in a quantitative assay (G. Piazza &
`D. Hixson, unpublished work), suggesting that loss ofthis epitope
`alone is not responsible for lack of enzymic activity.
`Although two forms of DPPIV have been purified from rat
`liver plasma membranes, a detergent-soluble membrane form
`and a soluble form released by papain which is approx. 4 kDa
`smaller, these forms are identical, except that the N-terminal 35
`amino acids, including the signal peptide which anchors the
`enzyme in the plasma membrane, are missing from the smaller,
`soluble, form (Ogata et al., 1989). A smaller, soluble, form has
`also been observed to account for much of the DPPIV activity in
`rat hepatoma cells (Hartel et al., 1988b), but is not likely to
`correspond to the 95 kDa band we observe in D - extracts.
`The availability of rats deficient in DPPIV has permitted
`functional studies of the role of this enzyme in various physio-
`logical and pathological processes, such as peptide hydrolysis
`and transport in renal brush-border membranes (Tiruppathi
`et al., 1990a,b) and in passive Heymann nephritis (Natori et al.,
`1989). We have chosen to exploit the D - rats as a syngeneic host
`for the transplantation of small numbers of hepatic cells from
`D+ animals in order to study their colonization and differen-
`tiation potential. Such a system obviates the need to manipulate
`the transplanted cells by introducing markers like recombinant
`DNA expression vectors or to use anti-allogeneic sera and F,
`animals in order to identify the transplanted cells in the same
`kind of tissue. It also overcomes the problem of immunological
`rejection or the use of nude mice. Since DPPIV is expressed
`normally in a variety of cell types, including hepatocytes and
`ductular cells of liver, and its expression begins at a defined stage
`of development (Hong et al., 1989 a,