RHGH Synthesis
RHGH Synthesis
RHGH Synthesis
1. Introduction
Michael]. Ross, Kenneth C. Oison, Michael D. Geier,John V. O'Connor, and Andrew ].S.Jones •
Genentech, Inc., South San Francisco, California 94080.
241
Human Pituitary
~
m RNA
Re verse I Trans crlptase
DNA • Polymerase
ds DNA
l
III I 1IIIII jill i III i II i III I III11111 i I
Haelll Re stric tion Nuclease
i II iii I I I I I iI I II II iii Ii II i i Ii i i
551 bp FRAGMENT
Terminal l Transferase
ccccc JillIIDiiCllllIIDillll!IJillJlllIIDilCllllIIDIiClIIIJ:lilJlllIIIJIiCII cccccc
TAILED FRAGMENT
DNA l Ligase
MIXTURE OF
LIGATED PLASM IDS
Figure 1 Cloning of the human growth hormone (hGH) message. Size fractionated messenger
RNA isolated from human pituitaries was copied using reverse transcriptase to form a heteroduplex
molecule. The RNA strand was then displaced with a second DNA strand with DNA polymerase I
Klenow fragment. The resulting mixtures of double-stranded DNA molecules were cut with
restriction endonuclease Hae III and fractionated on an agarose electrophoresis gel. The hGH-
specific cDNA of 551 base pairs (bp) in length was "tailed" with deoxycytosine using deoxynucleotide
terminal transferase and ligated into dG tailed cloning vectors using DNA ligase. The mixture of
ligated plasm ids was used to transform E. coli K-12. Clones containing the hGH cDNA fragment
were isolated and the amplified 551-bp fragment purified.
-0
Eco RI Haelll Endonucleases
IIIIII1IIIII11111 III1IIIIIIII
DNA 1 Ligase
Figure 3 Expression of met-hGH. The Hae III-Eco Rl fragment of the hGH eDNA contains the
coding sequence for amino acids 23-191 of hGH: an 84-base fragment of DNA which coded for a
methionine start codon and amino acids 1-22 of hGH. These two fragments were then ligated
together into an expression plasmid derived from pBR322. The expression plasmid contained
promoters, selection markers, and ribosome binding site to allow efficient expression of hGH.
mones that are found in the methionyl hGH preparation are very similar in
their isoelectric pattern to the isohormones found in pituitary-derived hGH.(5)
The major difference observable using isoelectric focusing is that there are
fewer isohormones present in met-hGH.
1.2.2 . Folding
The three-dimensional structure of the molecule has been shown to be
indistinguishable from pituitary hGH by circular dichroism studies.(6) Immu-
nochemical studies comparing met-hGH and pit-hGH in sera from both
hyperimmunized rabbits and patients have as yet found no new determinants
in the methionyl compound, indicating that the amino terminus of the molecule
is probably not accessible on the molecule's surface.
1.2.3. Bioassay
Finally, bioassay data using rat weight gain and rat tibia length models(4.7)
as well as radioreceptor and receptor modulation assays show met-hGH to be
equally potent with pit-hGH.(8)
The first step in the process of manufacturing met-hGH for clinical use is
fermentation, a highly reproducible and easily scalable process which has been
244 III • Biosynthesis of Human Growth Hormone
11 J.£
2/10,000"
Single Bacterium
8 Flask
/
o
00
00
000
000
10..l to 1,000,000..l
Figure 4 Controlled fermentation processes can be carried out at scales ranging from 10 liters to
106 liters. Each batch is derived from the single colony of bacteria which is grown up to yield a
master culture. This master culture is exhaustively characterized and subdivided into identical vials
and stored frozen. For each new batch, a single vial of this master culture is inoculated into a flask
of nutrient broth where it is grown and used to inoculate a fermenter, which can then be used to
inoculate larger fermenters.
Figure 6 An SDS-PAGE analysis (Coomassie blue stain) of column fractions from the initial ion
exchange column used to purify Protropin.
18 • Recombinant DNA Synthesis of Human Growth Hormone 247
1 2 4 3 4
Figure 8 Met hGH (Process Code GII002) was analyzed by SDS-PAGE using silver staining and
immunoblotting: 20 IJ.g, 10 IJ.g, 2 IJ.g, and 1 IJ.g of met-hGH (denatured in SDS/BME) were loaded
onto three identical sections of the same gel in lanes 1, 2, 3, and 4, respectively. The proteins in
the gel section in panel A were transferred to nitrocellulose and blotted with rabbit anti pituitary
hGH antibody. Antibodies bound to the blot were visualized by 125I-labeled protein A followed by
autoradiography. The major monomer band and the dimmer band can clearly be identified as
hGH as well as some minor higher- and lower-molecular-weight species. Panel B is a section of gel
stained with silverY7) A contaminant between hGH monomer and dimer can be easily seen as some
higher-molecular-weight species (larger than hGH dimer). Panel C is a section of gel which was
transferred to nitrocellulose and blotted with anti-E. coli protein antibody. The contaminants can
be clearly identified as being from E. coli origin.
(Fig. 8). Similar SDS gels were subjected to radioimmunological staining with
both anti-E. coli and anti-hGH antibodies to unambiguously assign the origin
of the newly detected bands. Two bands of E. coli origin were thus identified
and the remaining higher-molecular-weight species « 1% of the total protein
by Coomassie stain) were identified by anti-hGH antibody.
The purification process was then modified by monitoring SDS-PAGE
silver gels and by using data from experiments in which radio labeled E. coli
18 • Recombinant DNA Syntheris of Human Growth Hormone 249
I I
4. Final Product
4. Final Product
Figure 9 Radioactive spike process validation. Production organisms (E. coli missing the hGH
gene) are grown in media containing 14C_ or s5S-labeled amino acids. The labeled extract is added
to hGH containing cell paste from a normal production run. The purification is carried out as it
would be to produce Protropin except that the scale is reduced about 100X.
14C Label
35 8 Label
--
Protein
-
4
--
(mg.ml-',
cpm.ml- 1 1. Protein
(x10·3 ,
Img.ml·',
1.
--
3
14 cpm.mr 1
(x10·4 ,
12
10 2
•
6
4
102030 40 50 60
Fraction Number Fraction Numbar
Figure 10 Monitoring of an ion exchange purification step using radioactive spikes. Radioactively
labeled E. coli were added during the Protropin T. purification process as described in Fig. 9. The
separation of contaminants (radioactivity) from met-hGH and the final step of the GIl 008 process
when the met-hGH was 99.7% pure coming into this step are illustrated. Two separate experiments
are shown; the experimental designs were similar, but in one experiment the E. coli were grown
with 35S04 and in the other, grown with l4C-Iabeled glucose.
250 III • Biosynthesis of Human Growth Hormone
0.2
300
35S 200
cpm/tube
100
Figure 11 HPLC analysis of 35S impurity in Protropin process validation. Size exclusion HPLC
on a TSK-3000SW column run in 100 mM sodium phosphate 0.1% SDS, pH 7.0.
free radicals generated by the f3 particles and chlorine atoms which are the
result of the sulfur decay cause low-molecular weight 35S E. coli contaminants
such as glutathione to become covalently attached to met-hGH. The damaged
hGH could co purify with met-hGH and could have antigenic sites damaged.
Since this process would be an artifact of the use of radioisotopes in the
experiment, the similar met-hGH adducts are unlikely to exist in the clinical
product. Therefore, absolute quantitation of impurities is probably not accurate
if they represent less than -0.1 % of the major protein.
Phase II clinical trials were initiated in children with hGH deficiency. The
detailed results of this 18-month trial are presented by S. Kaplan et al. (this
volume, Chapter 20), but we will summarize the results very briefly: The met-
hGH proved highly efficacious in promoting linear growth and no drug-related
side effects were noted. However, a majority of the patients developed low-
titer antibodies to hGH. In an effort to eliminate this immunological side effect
of therapy further, process improvements were undertaken.
A c
1 2 3
Figure 12 Analysis of Process Code GII008 met-hGH by SDS-PAGE. This figure is analogous to
Fig. 8 in that 20 ILg. 10 ILg. 2 ILg. and 1 ILg of met-hGH (denatured in SDS/BME) were loaded on
three identical sections of the same gel in lanes 1.2.3. and 4. respectively. The proteins in the gel
section in panel A were transferred to nitrocellulose and blotted with anti-hGH antibody. Only
met-hGH monomer. a small amount of dimer. and a hint of trimer can be seen in the resulting
autoradiogram. Panel B is stained with silver. All of the visualized bands except a trace of a species
with a molecular weight lower than hGH monomer in the overloaded 20 ILg lane are identified in
Panel A as being of hGH origin. In panel C the proteins on the gel were transferred to nitrocellulose
and then blotted with anti-E. coli protein antibody. No bands can be seen on this autoradiogram.
The assay, however, was very process dependent. Only those antigens
which could contaminate the hGH prepared by this manufacturing process
(Code GllOl5) were selected during the preparation of those ECP's. The assay
is therefore specific for these contaminants. A further complication is that the
conditions (e.g., pH, buffer) for the extraction and handling of the ECP's have
the potential for changing the tertiary structure of the antigen molecules and,
hence, their antigenic determinants. This process with different first steps could
give falsely low values in this ECP assay. In fact, Process Code Gl1002 met-
hGH, which is prepared by a different process than GII015, gives an ECP
assay value of about 100-200 ppm although in fact we know the preparation
contains at least 10,000 ppm ECP contaminants.
This ECP assay was used to characterize the Gl1015 process for met-hGH
which products contain less than 30 ppm ECP's. There are five major separation
18 • Recombinant DNA Synthesis of Human Growth Hormone 253
999
Fold 99
Purification
Per 99.9
Step
99
2 3 4 5
Process Steps
Figure 14 Prototropin purification. The five major purification steps used in the Gl1015 are
analyzed to determine their efficiency in removing E. coli contaminants. The fold purifications
(bars) are calculated by taking the ECP values/mg of hGH before process steps and dividing by the
ECP values after those steps. The cumulative effect of the steps can be seen by plotting the overall
purity at the completion of the steps. The material after step 5 is purified bulk ready for formulation.
steps in the process, and each contributes significantly to the overall purification,
as can be seen in Fig. 14. The met-hGH produced by process GII015 has been
tested in GS-deficient children for 6 months and has shown itself to be as
efficacious as, but significantly less immunogenic than, material produced by
process code GII008 .
99.999
99.99
..
I
Ph ••• II
• •
Ph.s. II'
I
Percent
..
Purity
99.9
Figure 15 Purity of different hGH preparations. The approximate purity of various met-hGH
preparations used clinically is plotted versus time. The clinical trials that were run with each
preparation are indicated by the arrows. The purity of the GIl 002 was estimated by 35S-labeled
process validations and the Gll008 and Gll015 preparations using the immunoradiometric ECP
assay.
18 • Recombinant DNA Synthesis of Human Growth Hormone 255
2.2. Summary
Three different preparations of met-hGH have been tested by Genentech
in human clinical trials. Each new process was significantly purer than its
predecessor and required the development of new analytical tools before that
new level of purity could be achieved (Fig. 15). Our experience has led to the
conclusion that the different preparations of met-hGH behave very differently
in the clinic. Consequently, it is very important to differentiate between
manufacturers and between different processes from the same manufacturer
when examining clinical results. We must rely on rigid control of a specific
manufacturing process to guarantee the safety of the product.
ACKNOWLEDGMENTS. We thank Matt Field, Vince Anicetti, Kathy McKeown,
Greg Bennett, Baron Reed, and Michele Sanda for their excellent technical
assistance and Sara Wright for help in preparing this manuscript.
References
1. D. V. Goeddel, H. L. Heynecker, T. Hozumi, R. Arentzen, K. Itakura, D. G. Yansura, M. J.
Ross, G. Miozzari, R. Crea, and P. H. Seeburg, Direct expression in Escherichia coli of a DNA
sequence coding for human growth hormone, Nature 281:544, 1979.
2. P. H. Seeburg, I. Shine, J. A. Martial, J. D. Baxter, and H. M. Goodman, Nucleotide sequence
and amplification in bacteria of structural gene for rat growth hormone, Nature 270:486, 1977.
3. W. J. Kohr, R. Keck, and R. N. Harkins, Characterization of intact and trypsin-digested
biosynthetic human growth hormone by high-pressure liquid chromatography, Anal. Biochem.
122:348, 1982.
4. K. Olson, J. Fenno, N. Lin, R. N. Harkins, C. Snider, W. H. Kohr, M. J. Ross, D. Fodge, G.
Prender, and N. Stebbing, Purified human growth hormone from E. coli is biologically active,
Nature 293:408, 1981.
5. M. J. Ross, Production of medically important polypeptides using recombinant DNA technology,
in: Insulins, Growth Hormone, and Recombinant DNA Technology a. Gueriguian, ed.) Raven Press,
New York, 1981, pp. 33-48.
6. A. J. S. Jones and J. V. O'Connor, Chemical characterization of methionyl human growth
hormone, in: Hormone Drugs, Proceedings of the FDA-USP Workshop on Drug and Reference Standard
for Insulin, Somatotropins and Thyroid Axis Hormones, U.S. Pharmacopeial Convention, Inc., Rock-
ville, Maryland, 1981.
7. N. Stebbing, K. Olson, N. Lin, R. N. Harkins, C. Snider, M.J. Ross, F. Fields, L. May,J. Fenno,
D. Fodge, and G. Prendor, Biological comparison of natural and recombinant DNA-derived
polypeptides, in: Insulins, Growth Hormone, and Recombinant DNA Technology a. Gueriguian, ed.),
Raven Press, New York, 1981, pp. 117-132.
8. R. G. Rosenfeld, B. B. Aggarwal, R. L. Hintz, and L. A. Dollar, Recombinant DNA derived
methionyl human growth hormone is similar in membrane binding properties to human
pituitary growth, Biochem. Res. Commun. 106:202, 1982.
9. J. L. Anderson, H. W. Marshall, B. E. Bray, J. R. Lutz, P. R. Frederick, F. G. Yanowitz, F. L.
Datz, S. C. Klausner, and A. D. Hagan, A randomized trial of intracoronary streptokinase in
the treatment of acute myocardial infarction, N. Engl. J. Med. 308: 1312, 1983.
10. E. L. Grinnan, L-Asparaginase: a case study of an E. coli fermentation project, in: Insulins,
Growth Hormone, and Recombinant DNA Technology, a.Gueriguian, ed.), Raven Press, New York,
1981, pp. 99-108.
11. H. DeBoer, H. Heyneker, L. Comstock, A. Wieland, M. Vasser, and T. Horn, Construction of
three hybrid promoters and their properties in Escherichia Coli, in: From Gene to Protein: Translation
256 111 • Biosynthesis of Human Growth Hormone
into Biotechnology (F. Ahmad, J. Schultz, E. E. Smith, and W. J. Whelan, eds.), Academic Press,
New York, 1982, pp. 309-327.
12. A. J. S. jones and J. V. O'Connor, Control of recombinant DNA produced pharmaceuticals by
a combination of process validation and final specifications, in: Developments in Biological
Standardization: Symposium on Standardization and Control of Biologicals Produced by Recombinant
DNA Technology, Vol. S9, (S. Kerger, ed.), Internation Association of Biological Standardization,
Basel, Switzerland, 1985, pp. 175-180.
13. C. A. Dinarello, J. V. O'Connor, G. LoPreste, and R. L. Swift, Human leukocytic pyrogen test
for detection of pyrogenic material in growth hormone produced by recombinant Escherichia
coli, J. Clin. Micro. 20:323-329, 1984.
14. J. W. Ross, R. S. Baker, C. S. Hooker, I. S. johnson, J. R. Scmidtke, and W. C. Smith, Procedure
for detection of potential E. coli peptides (ECP's) in biosynthetic human insulin (BHI), antibodies
to ECP's in patients treated with BHI, and measurement of bacterial endotoxins in BHI, in:
Hormone Drugs, United States Pharmocopeial Convention, Inc., Rockville, Maryland, 1982,.pp.
127-138.
15. R. S. Baker, J. R. Schmidtke, J. W. Ross, and W. C. Smith, Preliminary studies on the
immunogenicity and amount of Escherichia Coli polypeptides in biosynthetic human insulin
produced by recombinant DNA technology, Lancet 2:1139, 1981.
16. R. L. Hintz, D. M. Wilson, J. Fenno, R. G. Rosenfeld, A. Bennett, B. McClellan, and R. Swift,
Biosynthetic methionyl-human growth hormone is biologically active in adult men, Lancet
1:1276, 1982.
17. j. H. Morrissey, Silver stained for proteins in polyacrylamide gels: A modified procedure with
enhanced uniform sensitivity, Anal. Biochem. 117:307, 1981.
18. B. R. Oakley, D. R. Kirsch, and N. R. Morris, A simplified ultrasensitive silver stain for detecting
proteins in polyacrylamide gels, Anal. Biochem. 105:361, 1980.
19. F. C. Kafatos, C. W. jones, and A. Efstratiadis, Determination of nucleic acid sequence
homologies and relative concentration by a dot hybridization procedure, Nucelic Acid Res.
7:1541,1979.
20. W. N. Burnette, Western blotting electrophoretic transfer of protein from sodium dodecyl
sulfate polyacrylamide gels to unmodified nitrocellulose and radiographic detection of antibody
and radioiodinated protein, Anal. Biochem. 112:195, 1981.