Conference Report – Engineering Proteins: Interferon, Diphtheria Toxin, and Growth Hormone

Introduction

Very powerful technologies are available today for industrial production of biologic compounds and derivatives to be used for replacement therapy or other therapeutic intents. How are these proteins produced in laboratories in such large quantities, to a high degree of purity and activity?

At a session on protein engineering held at the Interphex Annual Conference of Pharmadiscovery, Dr. John Kopchick,[1] of Ohio University, Athens, Ohio, outlined the steps involved in the design and production of recombinant biologics, as well as the unexpected findings that may arise along the way. Three biologics, interferon beta (IFN), diphtheria toxin (DT), and growth hormone (GH), now approved for clinical use, have been discussed as examples of what it takes to bring a new biologic from the laboratory to clinical use.

Engineering Proteins

When engineering a novel protein for large-scale production in vitro, structure of the protein itself is of critical importance. Any significant change in protein structure may, in fact, yield a functionally inactive molecule, or a protein with significantly reduced biological activity.

The primary structure of a protein is defined by its amino acid sequence, the secondary structure by the presence of alpha helixes or beta sheets, the ternary structure by covalent bonds between adjacent protein stretches, and the quaternary structure by the association of multiple subunits in a large, multimodular protein complex (as in a pentameric immunoglobulin M).

A number of expression systems are used to express a recombinant protein (eg, bacteria, yeast, insect cells, mammalian cells), but not all yield proteins identical to their human counterpart. While most secondary and tertiary structures may be faithfully reproduced, quaternary structures and posttranslational modifications (eg, glycosylation, acetylation) are often differentially undertaken in different cell types from different species. If any of these properties is critical for the overall activity of the recombinant protein, such characteristics have to be faithfully reproduced in the engineered protein to obtain a powerfully active molecule.

In proteins rich in beta sheets, pairs of chains lying side by side are stabilized by hydrogen bonds. Formation of a chemical bond (disulfide bond) between 2 cysteine residues then contributes to the overall 3-dimensional (ternary) structure. Of note, while correctly encoded in most expression systems, disulfide bonds and the deriving ternary and quaternary structures may be irreversibly compromised by chemical processes associated with protein isolation and purification.

An additional factor that needs to be taken into consideration is the preferential codon utilization (the triplet of bases coding for a specific amino acid) that may vary from species to species, particularly from Escherichia coli to humans. If a faithful replica of a human protein is being sought, attention has to be paid to the DNA sequence being inserted in the expression system of choice.

Recombinant Interferon Beta

IFN is one of the cytokines (soluble mediators) produced in tiny amounts by white blood cells and endothelial cells in response to viral infection, with the ultimate effect of blocking viral replication in infected cells. IFN contains 3 cysteine residues that are critical for its structure and function, at residue 17, 31, and 141.[2] In natural IFN, a disulfide bond between Cys 31 and Cys 141 yields a correctly folded molecule with potent antiviral activity.

As recalled by Dr. Kopchick,[1] initial attempts at producing recombinant IFN in bacteria yielded a mixture of recombinant IFN forms, only a third of which were biologically active. Such heterogeneity stemmed from the occurrence of an incorrect disulphide bond between Cys 7 and Cys 141 that severely compromised the ternary structure, and hence the biological function of the recombinant molecules. A mix of active/inactive proteins cannot be employed in vivo because the inactive form, if still endowed with residual binding activity, may act as an antagonist or a partial agonist, thus altering the effector activity of the active protein. Alternatively, at best, patients would be injected with unnecessary “trash,” or they would be exposed to altered proteins with an increased risk of producing antibodies and ensuing immune complexes.

To bypass the incorrect Cys bonding seen with first-generation recombinant IFN, the easiest approach was to remove the unwanted Cys 7 by changing it to an amino acid residue that would not affect overall protein structure such as serine. Cysteine and serine have a similar structure, with the sulfur element being replaced by oxygen in serine. The change was achieved with an in vitro mutagenesis step using an ad hoc mutated oligonucleotide that inserted the serine-encoding sequence in the expression vector used for the production of recombinant IFN.[1]

The derived molecule, IFN beta (Ser 17), proved to be functional and, to the surprise of the investigators, even better than expected: its shelf half-life was significantly increased with the Cys17-Ser17 mutation. Activity was, in fact, still detectable 150 days following storage in deep freezers. Functional testing showed that its antiviral and antiproliferative activities were comparable with those of purified human IFN, and thus equally potent from a biologic standpoint.[1]

Such engineered molecule has later been approved by the US Food and Drug Administration (FDA) for treatment of patients with relapsing-remitting multiple sclerosis owing to its documented immunosuppressive activity. Phase 3 clinical trials are in progress to evaluate the long-term safety, tolerability, and efficacy in this group of patients. Risks, optimal dosing, treatment timing, and how to prevent induction of specific antibodies to the recombinant proteins are among the most pressing questions awaiting an answer.[3–6]

Diphtheria Toxin and Cancer Treatment

The dimeric DT is what makes some Corynebacteria so dangerous following infection of the upper respiratory tract in humans. Release of DT is a potentially life-threatening event associated with accumulation of pseudomembranous material and edema. The intrinsic potency of DT, as often happens, has however attracted the attention of other researchers who were looking for a powerful way of killing unwanted cells, particularly in the context of malignant tumor growth.[7]

Dissection of the cytotoxic activity of DT revealed that following endocytosis and an acid-triggered conformational change, DT is inserted in the cell membrane through the action of the beta chain, with release of the alpha chain into the intracellular space. Once inside the cell, the DT alpha chain leads to potent inhibition of protein translation by binding to the elongation factor 2, and cell death.

Hence came the thought to use the DT alpha chain as a killing effector, after coupling to another delivery system to specifically target malignant cancer cells such as leukemia or lymphoma cells.[6] The interleukin (IL)-2 receptor, highly upregulated on these cancer cells, was chosen as the surface target for intracellular delivery of the truncated, hybrid DT. The recombinant DAB389-IL-2 toxin was indeed effective in blocking protein synthesis. While 50% cell killing was observed at 10E-7 M with DT, 50% cell killing was seen with the DAB389-IL-2 toxin at 10E-9 M.

Later approved for use in patients with cutaneous T-cell lymphoma in advanced stage,[8] it is being further tested for efficacy and toxicity in other cancers. An approximately 20% response rate has been reported so far. The hybrid nature of the recombinant molecule has, however, conferred immunogenicity: repeated treatments induce production of specific antibodies that may lead to loss of in vivo activity of the DAB389-IL-2 toxin. Further studies are in progress to find ways to reduce immunogenicity and construct new hybrids with different targeting activity (eg, the epidermal growth factor receptors overexpressed by subsets of solid tumors).[1]

Growth Hormone

Although estrogens and testosterone still hold a role of prominence among hormones in terms of citations in professional and lay publications, mentions of GH seem to be ubiquitously present in the cyberspace. Hardly a day goes by that its mention doesn't appear in our email inboxes. More clarity on what it is, what it really does and does not do, as well as the risks associated with its intake is probably needed, if not mandatory.

As mentioned by Dr. Kopchick,[1] shall we say right away what GH does not do?

• GH does not increase life span.

• GH does not increase memory retention.

• GH does not increase sexual performance.

• GH does not induce hair growth.

• GH does not reduce wrinkles.

Involved in metabolism, growth, and lactation, GH can have both anabolic and catabolic effects: while it increases lean mass by increasing muscle and bones, it reduces fat. Some effects are mediated directly, others are mediated indirectly through the action of insulin growth factor 1 (IGF-1).[9]

A somehow shocking report was released in the press a while ago — up to 50% of the athletes participating to the Sydney Olympics have been suspected of taking recombinant GH. Such estimate was reached counting the number of syringes and needles being discarded by the athletes. GH, in fact, is not active orally, but needs to be injected. (Yes, GH pills being touted as a miracle remedy for many ailments are at best totally ineffective or not GH at all). Why would some of our finest athletes resort to GH? What will be the pay back at the end of the road?

Current abuse of GH stems from its selective anabolic properties. More muscle power is perceived as an advantage in many disciplines. Monitoring is not very effective as the recombinant GH protein has a quite short half-life in vivo (about 20–30 minutes) and thus it is hard to detect at random checks. The pay back is, however, too high to see it continuing.

Inappropriate treatment with GH (for those who already have endogenous production) leads to severe long-term toxicity, particularly in the kidneys, with development of glomerulosclerosis (tissue scarring), kidney insufficiency, and need for dialysis treatment. As few as 4 months may be enough to see these effects in some subjects.[1] Acute toxic effects seen also in subjects with low levels of endogenous GH include edema, arthralgia, decreased sensitivity to insulin, and increased production of IGF-1.

Injectable, recombinant GH has been approved by the FDA for a limited number of conditions including children with growth failure who don't produce enough endogenous GH, subjects with Turner's syndrome and short stature, AIDS patients with wasting disease, and adults with a GH deficit (eg, secondary to trauma-related deficit of GH-releasing factor). GH should not be used in patients with cancer or for growth promotion in children with a GH deficit who have reached bone maturity (closed epiphyses).

GH as a classic endocrine hormone is distributed by the general circulation and acts by binding to the target tissues. Engagement of the dimeric GH receptor leads to activation of the JAK2 kinase, GH receptor phosphorylation, recruitment of STAT5, entry of STAT5 in the nucleus, and modulation of gene expression.[9] More genes are turned off by GH than turned on.[1]

Its molecular structure resolved in 1987 revealed 191 amino acids with a 26 amino-acid secretory sequence that is clipped by the pituitary cells before release of the mature form of GH in the circulation. Five exons code for the full-length GH sequence. Recombinant GH has now been produced by 4 companies in bacteria, but it can be produced also in mammalian cells (L cells). When produced in bacteria, the start-site methionine has to be clipped off to avoid conferring high immunogenicity to the recombinant GH molecules.

Of note, transgenic mice overexpressing GH showed a significant reduction in life span (from an average of 12 to 9 months), increased size, development of abnormalities, and organ insufficiencies in the liver, kidney, and heart.[10] To avoid toxicity in humans who need GH replacement therapy, children with defects in GH and short stature should receive only enough GH to reach a normal range of concentration, with interruption of the GH replacement therapy as soon as bones reach maturity.

Past attempts at producing recombinant GH with increased potency did not yield a molecule significantly superior to natural GH. In vitro mutagenesis of amino acid residues in the third alpha helix and expression in mammalian cells yielded a functional molecule with binding parameters very similar to those of wild-type human GH.[1]

Development of a GH Antagonist

Engineering of GH by introduction of only one amino acid change (glycine to alanine), on the other hand, yielded, to the researchers' surprise, a mutant recombinant GH molecule with antagonistic activity, able to prevent normal growth in experimental animals.[10,11] In light of these results, researchers at the National Institute of Health are studying children with growth defects to define the molecular basis of their GH deficit.

From a structural point of view, the Gly-to-Ala mutation substantially reduces the “cleft” present on the dimeric GH receptor. Of note, dimerization precedes GH hormone binding. GH-signals are transmitted to the intracellular milieu by changes occurring upon ligand binding in the conformation of the receptor's dimer. The antagonistic activity of the mutated GH molecule developed by Dr. Kopchick and colleagues is due to its inability to induce functional dimerization of the GH receptors and signal transduction.[1]

The main indication for treatment with a GH antagonist is acromegaly, the “disease of the giants.” Acromegaly, due to an increased GH production by pituitary tumors, leads to high levels of circulating IGF-1. If left untreated, acromegalic individuals have been reported to live only rarely past their 50th year, and generally 10 years within diagnosis. Patients may come to observation because of compression symptoms due to the pituitary adenoma, or other symptoms and signs related to the abnormal production of GH. Surgery is the therapeutic option of choice, whenever possible.[12]

An alternative strategy to antagonizing GH in peripheral tissues is to block or reduce production of the GH-releasing factor produced by the hypothalamus. A 8mer peptide, somatostatin, has been developed to this end and is effective in about half of patients with acromegaly. It is, however, still unclear why it does not affect GH release in about 40% to 50% of patients with GH overproduction.[1] Dopamine agonists that can be easily administered orally are also effective in about 35% to 40% of patients in normalizing GH and IGF-1 levels. Treatment with a GH antagonist would then represent a useful alternative for those patients in whom surgery is not a viable option, and those who have failed surgery, radiotherapy, or treatment with somatostatin analogs or dopamine agonists.[1,12]

The first GH antagonist developed, identical to natural GH with the exception of 1 amino acid difference, however, showed a very short half-life (about 20 minutes) when injected in vivo, as well as immunogenicity with induction of an antibody response. Further in vitro modification of the molecule (PEGylation) has significantly improved half-life that now reaches approximately 72 hours following injection.[13] Indexes of liver function should be monitored during treatment with pegvisomant.

Endogenous levels of IGF-1 have been selected as clinical end points in recent clinical trials. Treatment was found to be effective in normalizing symptoms and IGF-1 levels in the vast majority (93%) of the more than 200 patients treated. Of note, however, 2 patients in the 40- to 54-year-old age group had severe acromegaly but normal levels of IGF-1.[1] In further studies, pegvisomant was also shown to significantly reduce ring size (P < .005), another clinical measure of the severity of the disease in acromegalic patients.[1] Approval by the FDA for the treatment of acromegaly followed in 2003.

The potential use of a GH antagonist for the treatment of cancer (breast, colon, and prostate cancer) or diabetes is being evaluated in various preclinical and experimental studies currently in progress. Some breast cancer cells, for instance, have been found to overexpress the GH receptor. More information on its potential efficacy and toxicity in these groups of patients is expected to come in the next few years.[1]

The mechanism of action of a GH antagonist, however, is different from that of somatostatin and its analogs. Blocking the peripheral GH receptor, in fact, leads to downregulation of IGF-1, but also to an increase in circulating GH. How and where to manipulate the effects of these hormones, and whether any combination of these antagonists may indeed be effective for cancer treatment, is still an unfolding story.