Vascular Endothelial Growth Factors Biology Essay

4th Year Project Report. Submitted to the University of Manchester in part fulfilment of MPharm degree

Joseph Okeleke

Date submitted: 22/04/2013

School of pharmacy and pharmaceutical sciences

Professor Ian Stratford

Word Count: 20,894

Contents

Declaration

I understand the nature of plagiarism and that it is serious academic offence. I confirm that no material in the project has been plagiarised.

Joe Okeleke: 22/04/2013

Abstract

Acknowledgement

I would like to thank Professor Stratford for all the invaluable help and support he provided during this project. I can safely say that this project has sparked interest in oncology and I have future aspirations in that field.

I would also like thank my family and friends for all their prayers and support.

Abbreviations

Abbreviation

Name

VEGF

Vascular endothelial growth factor

HIF

Hypoxia-inducible factors

ECM

Extracellular matrix

muMAb

Murine monoclonal antibodies

MAb

Monoclonal antibodies

mCRC

Metastatic colorectal cancer

NSCLC

Non-small cell lung cell

mRCC

Metastatic renal cell carcinoma

GBM

Glioblastoma

FDA

Food and Drug Administration

Rhumab

Recombinant humanised monoclonal antibody

MTX

Methotrexate

WCB

Working cell banks

OS

Overall survival

PFS

Progression free survival

ORR

Objective Response Rate

OR

Objective Response rate

TTP

Time to progression response

ITT

Intention-to-treat analysis

CTC

Common toxicity criteria

Cmax

Maximum concentration

5-FU

Fluorouracil

LV

Leucovorin

IRF

Independent review facility

GI

Gastrointestinal

IFL

Irinotecan Flurouracil Leucovorin

Bv

Bevacizumab

RR

Reponse rate

DoR

Duration of response

Quality of Life

QoL

FOLFOX4

Oxaliplatin, Leucovorin, and Fluorouracil

MSKCC

Memorial Sloan-Kettering Cancer Center

6-PFS

Progression free survival at six months

CPT-11

Irinotecan

MRI

Magnetic resonance imaging

PET

Positron emission tomography

FOLFIRI

Infusional Flurouracil Leurovorin

ECOG

ECOG Performance Status

NICE

National Institute for Health and Care Excellence

INTRODUCTION

Cancer describes a group of diseases in which normal cells undergo numerous changes that cause deregulated cell proliferation and invasion from a local site of origin to other sites within the body1. Until recently, standard chemotherapy primarily focused on targeting cancer cells directly with the aim of disrupting cell cycle replication hence preventing the excessive proliferation, and ultimately causes cell death. However, the nature of cancer holds true in that they acquire numerous mutations as they grow rapidly, some of which provide these cells with the tools to evade or resist damage from chemotherapy as well as action from the immune system. Although over 200 different types of cancer exist, majority, if not all display six trademark features1. These include self-sufficiency in growth signals, insensitivity to anti-growth signals, evasion of apoptosis, indefinite replicative potential, sustained angiogenesis, and tissue invasion and metastasis2. These characteristics highlight the distinctive nature of cancer cells as compared to normal cells. The discovery of these characteristics have led to the development of new drugs which do not target cancer cells directly but instead disrupt the functioning of specific molecules and processes essential to the growth of these cancer cells. Avastin is one such drug, it targets the process of angiogenesis, which cancer cells utilise in order to maintain adequate supply of nutrients and oxygen in order to sustain rapid proliferation.

Figure This diagram displays the six essential characteristics noticeable in most tumours 3

Angiogenesis

Angiogenesis is the process by which new blood vessels develop from pre-existing ones4. It occurs naturally in embryo development, to help build the necessary blood circulation between mother and foetus5. Likewise angiogenesis restores blood flow to tissues after an injury hence aiding with wound recovery3,5. Besides this, angiogenesis is involved in repairing the uterus lining prior to ovulation during reproductive cycle.3 The angiogenic process is driven by an imbalance in expression of angiogenic factors which may stimulate or inhibit blood vessel growth. Over 200 angiogenic factors have already been identified and bind to their receptors located on surface of endothelial cells6. The role of angiogenesis in tumour growth is largely dependent on consistent imbalance of these angiogenic factors6. Tumour cells, in order to grow into a detectable mass, develop the capacity to induce and sustain angiogenic activity and hence promote constant blood vessel growth. In this manner its cells can gain access to a sustained supply of nutrients and oxygen. Tumour cells initiate a process known as the ‘angiogenic switch’ which alters the balance of pro-angiogenic and anti-angiogenic factors in favour of pro-angiognenic factors7. This is achieved by altering gene transcription of these proteins. Consequently, the expressions of pro-angiogenic proteins such as vascular endothelial growth factors (VEGF) and Fibroblast growth factors (FGF) are elevated while anti-angiogenic factors such as have reduced expression2. Following the ‘angiogenic switch’, tumour cells become angiogenic by phenotype meaning they are able to acquire their own constant blood supply. This coupled with cell proliferation results in rapid growth and possibly metastasis of the tumour.8,9 Angiogenesis performs an important role in the maturation of small, dormant and localised tumours into larger and expanding malignant tumours10.

Vascular endothelial growth factors

Vascular endothelial growth factor (VEGF) is recognised as one of the most potent inducers of angiogenesis and it is the target protein of Avastin. VEGF is a 45-kDa homodimeric glycoprotein with a variety of roles in both normal and abnormal processes of angiogenesis11,12. It is generally understood from in vitro and in vivo studies, that VEGF promotes the proliferation of vascular endothelial cells as well as promoting their survival. In vivo, VEGF acts a survival factor for endothelial cells by inhibiting apoptotic processes13. VEGF regulates vascular permeability by its ability to induce vascular leakage. Such activity explains its presence during inflammation as well as in other disease conditions13.

VEGF (also referred to as VEGF-A) is a member of the gene family containing other related ligands, namely VEGF-B, VEGF-C, VEGF-D, VEGF-E and placenta growth factor (PlGF).14 Each ligand binds to VEGF receptors and their corresponding co-receptors with different specificities, hence producing varying biological responses.4 VEGF-C and VEGF-D, for example bind to VEGFR-3 and induce proliferation, migration and survival of lymphatic endothelial cells (lymph-angiogenesis).12 PlGF induces recruitment of marrow derived cells.4

Alternative exon splicing of the VEGF-A gene results in several isoforms of the protein of which the most common variants are VEGF121, VEGF165, VEGF189, and VEGF20615. These variations are named according to the length of their amino acid sequence. VEGF121 the smallest isoform is freely diffusible, binding directly to VEGF receptors12. VEGF165 the predominant variant exists in both a diffusible and an ECM-bound form12. VEGF189 and VEGF206 the larger isoforms are not freely soluble, rather they are bound to heparin sulphate present in the extracellular matrix (ECM) and hence are poorly secreted16. However, these proteins can be cleaved by proteases into shorter diffusible fragments in order to be activated17. Alternatively ECM-bound isoforms can be released by interacting with matrix metalloproteinase16.

Expression of VEGF

Several in situ hybridization studies have revealed that VEGF mRNA is up regulated in many different tumours types including lung, breast, renal, kidney, bladder and ovarian cancer18. As tumour cells enlarge in mass, they become ischaemic; under these hypoxic conditions the transcription of VEGF mRNA is up-regulated8. The protein Hypoxia-inducible factor (HIF) is a basic hetero-dimeric protein that consists of two subunits, HIF- 1α and HIF 1β19. It is released by hypoxic cells and selectively binds to the promoter sequence of VEGF gene. This increases the rate of mRNA transcription hence more VEGF protein is produced15. Over expression of VEGF in tumours may also be a result of mutation or inactivation of a tumour suppressor gene known as von Hippel-Lindau (VHL). This protein is responsible for the degradation of VEGF as well as the inactivation of HIF-1 via oxygen dependent proteolysis20. Consequently, cells lacking the wild type VHL produce large amounts of VEGF leading to more sprouted capillaries15. VEGF expression is also up-regulated by the presence of other growth factors in tumours such as epidermal growth factor, platelet-derived growth factor, α-fibroblast growth factors and insulin growth factors15. Additionally, hormones such as oestrogen and inflammatory cytokines such as interleukin-1 and interleukin-6 have been shown to elevate VEGF mRNA expression8. It has also been shown that oncogenic forms of the tumour-suppressor genes up regulate the expression of VEGF8.

VEGF Receptors

VEGF binds to two tyrosine kinase receptors, VEGFR-1 (Flt-1) and VEGFR2 (Flk-1). Both of which are structurally similar and are located on the surface of vascular endothelial cells15. VEGFR3 is also a member if this family; however it does not interact with VEGF rather it binds to VEGFC and VEGFD resulting in proliferation, migration and angiogenesis of lymphatic endothelial cells (table 1)13. In addition to binding to receptor tyrosine kinases (RTKs), VEGFs also bind co-receptors such as neuropilins (NRP); these have a positive modulatory effect on VEGFRs21.

Table 21compares the biological function and receptor target of each VEGF isoform

Functions, Binding Properties, and Biological Implications of VEGFs

VEGF isoform

Receptor

Co-receptor

Biological function

VEGFA165

VEGFR1,VEGFR2

NRP1, NRP2

Angiogenesis (permeability, survival, migration of EC)

VEGFA121

VEGFR1, VEGFR2

NRP1

Angiogenic/antiangiogenic propertiesa

VEGFA145

VEGFR1,VEGFR2

NRP2

Angiogenesis

VEGFA189

VEGFR1, VEGFR2

NRP1

Angiogenesis

VEGFB

VEGFR1

NRP1

Fatty acid uptake in EC of the heart

VEGFC

VEGFR3 (VEGFR2)

NRP2

Lymph angiogenesis

VEGFD

VEGFR3 (VEGFR2)

NRP2

Lymph angiogenesis

PIGF

VEGFR1

NRP1, NRP2

Inflammatory cell recruitment

EC: Endothelial cells; a: a study by Nowak et al (2008) reported VEGFA121 as anti-angiogenic21.

VEGFR1

VEGFR1 is a single- trans membrane glycoprotein that is expressed in vascular endothelial cells 21. It is also expressed in a range of non- endothelial cells such as monocytes, macrophages, vascular smooth muscle cells and cancer cells21. However it is expressed at such low levels in the body that it has been difficult to elucidate enough amounts to investigate in detail, hence its signal transduction pathway is poorly understood (figure 2)21. What is known though is that it has a higher affinity for VEGFA than VEGFR2, though more poorly activated22. VEGFR1 is also able to bind to VEGFB and PlGF12. The function and signalling properties of VEGFR1 are dependent on the cell type and the ligand that binds to it21,22. The binding of VEGFB to the receptor results in increased fatty uptake by cells while interaction of PlGF with VEGFR1 regulates the recruitment of inflammatory cells23. It has been suggested that interaction of VEGF and VEGFR1 results in progression of tumours and increased tumour invasiveness.4 A study by Fan et al (2005) on the effect of VEGFR1 on colorectal cancer cell lines showed that the receptor increased migration, invasiveness and colony formation but not cell proliferation of cancer cell lines, thereby drawing a strong correlation between angiogenesis and metastasis24

VEGFR2

VEGFR2 a 200-300kda receptor has high affinity for VEGF but it may bind the processed forms of VEGF-C, VEGF-D and VEGF-E25. As such, it is expressed in both vascular and lymphatic endothelial cells25. It is also expressed in non-endothelial cells such as pancreatic duct cells, retinal progenitor cells, megakaryocytes and haematopoietic cells16. VEGFR2 is considered the main mediator of VEGF-A induced physiological and pathological responses25. VEGFR2 induces vascular permeability, EC proliferation, invasion, migration and overall survival of vascular endothelial cells25. The signalling pathways are better understood when compared to VEGFR1 and figure 3 illustrates the complexity of the processes involved. VEGFR2 plays an important role in cancer growth as evidence by Calvani et al (2006) showed that VEGFR-2 is responsible for mediating signalling in chronic hypoxic environments leading to increased tumour cell survival and tumour malignancy21,26.

VEGF Receptor Signal transduction

VEGF proteins bind to the N-terminus extracellular domain of VEGF receptors as free diffusible proteins or presented bound to a co-factor16. Such binding prompts dimerization of receptor subunits to form either homo- or hetero dimers. This process is necessary but not sufficient to incite activation of the kinase receptor. Parallel to this, VEGF also induces changes in intracellular domain conformation which stimulates ATP bind and auto or trans-phosphorylation of tyrosine residues16. This results in activation complex signal transduction pathways ultimately leading to specific biological responses.

Figure 2 and 3 display current knowledge on the signal transduction pathways for VEGFR1 and VEGFR2, both of which bind VEGFA.

Cover

Figure 2: schematic representation of the signal transduction pathway of VEGFR1 16

Cover

Figure 3: schematic representation of the signal transduction pathway of VEGFR2 16

Therapeutic significance

VEGF is expressed in numerous tumours, making it a good target for various cancers. VEGF circulates in the blood acting directly on endothelial cells; hence angiogenesis can be blocked without the need to directly targeting the tumour, as is the case with chemotherapeutic drugs. Additionally, VEGF stimulates mitosis in endothelial cells but has little effect on other cell type’s which means that targeting this protein it is unlikely disrupt normal physiological processes, manageable side effect profile and different range of side effects to that of chemotherapeutic drugs20.

Unlike in physiological angiogenesis, VEGF stimulates new bloods vessels to form in a disordered manner causing blood vessels to be vastly permeable and leaky20,27. This leads to poor tumour perfusion resulting in hypoxia which induces more VEGF production. Leaky blood vessels elevate the interstitial pressure around the tumour making drug delivery to tumour target sites difficult and reducing the efficacy of these drugs20. VEGF inhibitors reduces permeability of blood vessels hence interstitial pressure20. It therefore follows that combination therapy consisting of chemotherapy drug and a VEGF inhibitor can lead to additive or possibly synergistic anti-tumour activity.

Anti-angiogenic Drugs

Angiogenesis is a multi-faceted and complex process that involves growths factors such as VEGF and fibroblast growth factors, receptors such as receptor tyrosine kinases, transcription factors and many other signalling molecules28. Consequently, drug developers are investing substantial time and finance into developing drugs to that inhibit these signalling molecules. There are several types of anti-angiogenic drugs on the market or currently being developed21. Receptor tyrosine kinases and growth factors are desirable targets because they are located in the extracellular environments of cells hence developed drugs do not have to penetrate the cellular membrane, which is often a problem for chemotherapy drugs28. Some kinase inhibitors include Sunitinib which has been approved to treat gastro-intestinal stromal tumour and Sorafenib for the treatment of renal-cell cancer21. Table 2 details some anti-angiogenic drugs in clinical use and others still undergoing clinical development.

Table 21 Presents an array of anti-angiogenic drugs on the market or undergoing clinical testing

Therapeutic agent

Type

Target

Clinical development

Bevacizumab/Avastin

mAb

VEGFA

Approved in 2004 (CC),2006 (NSCLC),2008 (RCC),glioblastoma (2009)

Ramucirumab/IMC-1121B

mAb

VEGFR2

Phase II/III

VEGF-Trap/aflibercept

Fusion protein

VEGFA,PIGF

Phase II/III

VEGFAS/Veglin

Oligonucleotide

VEGFA,VEGFC, VEGFD

Phase I

SU11248/sunitinib (Sutent)

RTKI

VEGFR1-3, PDGFR, c-kit, Flt3

Approved in 2006 (GIST and RCC)

Sorafenib (Nexavar)

RTKI

VEGFR2-3, PDGFR,Raf-1, Flt-3, c-kit

Approved in 2005 (RCC), 2008 (HCC)

Pazopanib (Votrient)

RTKI

VEGFR1-3 PDGFR, Flt-3, c-kit

Approved in 2009 (RCC)

AG013736/axitinib

RTKI

VEGFR1-3, PDGFR, c-kit

Phase II/III

AZD6474/vandetanib (Zactima)

RTKI

VEGFR1-3, EGFR RET

Phase II/III

AZD2171/cediranib (Resentin)

RTKI

VEGFR1-3, c-kit

Phase II/III

Brivanib alanitate

RTKI

VEGFR2 FGFR1

Phase II/III

AV-951/tivozanib

RTKI

VEGFR1-3, PDGFR

Phase II/III

PTK787/vatalanib

RTKI

VEGFR1-3, PDGFR, c-kit

Phase II

Abbreviations: BC, breast cancer; CC, colorectal carcinoma; HCC, hepatocellular carcinoma; mAb, monoclonal antibody; NSCLC, non-small cell lung carcinoma; RCC, renal cell cancer; RTKI, receptor tyrosine kinase inhibitor21.

Avastin

Following the discovery of VEGF in 1989, several murine anti-VEGF monoclonal antibodies were produced in order to gain better understanding of the VEGF physiology. These were tested on a number of xenograft models and showed efficiency in a wide range of tumours8. The initial study carried out by Kim et al showed that a murine monoclonal antibody (muMAb VEGF A 4.6.1) unique to VEGF could suppress angiogenesis and growth in multiple tumour lines8. MuMAb VEGF had a high affinity for VEGF and could recognize all VEGF isoforms. Results also showed that this MAb inhibited the growth of a number of human tumour lines when transplanted into nude mice29. The main drawback of muMAb VEGF A.4.6.1 was that it caused significant immune responses when used in human therapy29.

A recombinant humanised version (Avastin) of the murine antibody was later produced for clinical evaluation8,30. RhuMab VEGF was developed from the murine MAb by site directed mutagenesis of a human antibody framework. Six complementary determining regions were changed from human to murine MAb. Additionally, seven residues from the humanized variable domain, and one residue from the humanised variable light were replaced with murine residues. This was done in order to achieve binding equivalent to muMAb VEGF A.4.6.129. Consequently, Avastin binds with a similar affinity to murine MAb as well as neutralising all VEGF-A isoforms18. Crystal structure analysis of the Fab-ligand complex makes it possible to determine the epitope that is responsible for Avastin’s binding and specific activities, namely glycine 8818.

Figure 3- This diagram highlights Glycine 88 which is essential for binding as well as specificity31

Mechanism of action

Bevacizumab ( Avastin), a monoclonal antibody, neutralises the biological activity of VEGF by binding to it and creating a steric block between VEFR and its receptors VEGFR-1 and VEGFR-232. This prevents the signalling cascades that activate angiogenesis, leading to reduced tumour perfusion and vascular volume32. Furthermore the number of viable, circulating mature and progenitor endothelial cells in patients with carcinoma is lowered.32 . Neutralisation of VEGF by bevacizumab is relatively restricted to regions of tumour growth. In most normal tissues except renal glomeruli, the expression of VEGFR receptors is significantly low, in some cases undetectable. However, in many tumours such as colorectal cancer and glioblastomas, the expression of VEGFR is significantly up-regulated33. Hence it follows that tumour angiogenesis and thereby tumour growth and metastases is specifically inhibited by bevacizumab.

Approved Indications

Bevacizumab is approved for:

Metastatic colorectal cancer (mCRC) with IV 5-fluorouracil-based chemotherapy for first and second line treatment34.

Advanced non-squamous non–small cell lung cancer (NSCLC) in combination with carboplatin and paclitaxel as first line treatment34.

Metastatic kidney cancer (mRCC) when used with interferon alpha34.

Glioblastoma (GBM) as a single agent in adult patients with the progressive form of the disease following prior therapy34.

Drug development process of Avastin

Getting a drug to market is a long and complicated process that can take between 10-15 years to achieve. Additionally the cost of developing a drug is significantly expensive. On average a pharmaceutical company may accrue costs up to and beyond $1.3 billion, with a significant proportion of finance spent on the discovery process and Phase III clinical trials35. Although a complex process, developing a drug can be simplified into 4 steps (figure 4). The comprehensive diagram below (figure 4) attempts to chart the development of Avastin up until its first initial approval by the FDA.

Discovery

The process of drug discovery involves identifying the lead compounds that show activity against the target disease or condition. Subsequently, these compounds are evaluated for safety, in order to understand their disposition characteristic and to discover any potential adverse effects. Safety information is used to optimise the lead compound, which may result in analogous compounds being developed. Following this, the best optimised lead compounds move on to the preclinical testing.

Preclinical trials

Once a drug is developed, it undergoes rigorous laboratory testing (usually 3 years) to prove to the FDA that it is safe to be tested in humans36. The drug toxicology is tested in animals and living tissue in order to estimate the dosage and dosage regimen that is safe to administer to humans.

IND application

This is the first of two FDA reviews which assesses all the preclinical data concerning the drug. Submitted IND must be reviewed by the FDA within 30 days in order for pharmaceutical company to begin conducting trials in humans36.

Clinical Trials

Phase I

In this phase, safety of the drug is assessed in a small population of healthy volunteers (20-80), over a period of months36. The study is designed to determine the drugs absorption, metabolism and excretion profile. Phase I studies also investigated side effects that rise due to drug administration and as the dose is increased. Providing the drug is well tolerated and its side effects are acceptable, the drug can proceed to phase II testing.

Phase II

Phase II trials aim to determine the safety and efficacy of the drug in patients with specific diseases or condition. The trials involve a larger population group usually in the hundred and are often placebo-controlled or evaluated against the standard treatment for that specific disease or condition36.

Phase III

Phase III studies involve several hundreds to thousands of patients in order to further assess the safety and efficacy of the drug compared to the standard treatment.

NDA application

This is the second review conducted by the FDA. Information on all animal and human data and analyses of all this information, along with information on how the drug is manufactured is submitted as one whole document (usually 100,000 pages) to the FDA36. The FDA assigns a review team consisting doctors, chemists, statisticians, microbiologists, pharmacologists, and other experts to evaluate the application, which may take up to 2 years 36. The Drug is approved providing it is safe, effective for its purpose and its manufacturing process it is reproducible and meets the required standards .Once approved, the drug can be sold on the market.

http://www.alzdiscovery.org/wp-content/uploads/2009/04/pre-discovery.jpg

Figure 4: Chart representing the stages involved in drug development. Chart also describes the attrition rate of compounds as they move through the drug development process over time35.

Aims and Objectives

This project will analyse the relevant information concerning the design and development of Avastin. It will include both pre-clinical and clinical data, prior to its approval for therapeutic use. It will also explore the decision making processes that were involved at each stage of development.

The objectives are as follows:

To evaluate the current indications for Avastin.

To review clinical trial data for each disease where Avastin is approved

To review clinical trial data for disease where it wasn’t approved

To summarise the reasons why approval was granted.

To evaluate the decisions that were made to allow the progression of Avastin through developmental stages leading to its marketing authorization.

To explore the reasons why Avastin was only approved for certain tumour types and not others.

METHODOLOGY

Introduction

Chapter 2 describes in detail the approaches that were employed in order to collect relevant data needed to review the development of Avastin. Accordingly, it was then possible to piece together the information that accurately described the progression of Avastin from its initial design to its validation by the FDA for each of its indications. Consequently, evaluating the FDA’s decision to licence Avastin became conceivable. In order to lessen the burden associated with selecting and analysing a sizable amount of data, I utilised a systematic approach that aided in selecting the most relevant trials as well as minimising bias.

Data Collection

Targeted searching

After examining various databases, the decision was made to use PubMed as the main source of identifying trial publications ranging from 1992- 2010. Although bevacizumab was patented in 1997, preclinical trials were dated as far back as 1992. PubMed was considered the most ideal database because it was the more user-friendly compared to other databases such as Web of Science and Sciverse. Furthermore it has a large database with a wider range of accessible published studies including a number of relevant older publications. Moreover, PubMed, being a U.S managed database, increased the probability of retrieving relevant data regarding Avastin, which itself was developed by a US pharmaceutical company, Genentech. However, in order to avoid limiting the range of data that could be retrieved other minor databases such as Cochrane, web of knowledge and sciverse were also utilised to identify studies otherwise not found in PubMed. This ensures that the probability of introducing bias is minimised and the integrity of my analysis is maintained. Another method that aided in the compilation of trial studies was to review the reference list or citations of retrieved trial studies.

Search strategy

Search terms were limited to publication dates ranging from 1992 to 2010. The key search terms included: VEGF, bevacizumab, RhuMab VEGF, Phase I, Phase II, Phase II, recombinant humanized monoclonal antibody murine VEGF, muMAb, colorectal cancer, glioblastoma, renal cell carcinoma, NSCLC, breast cancer. Filters were employed to restrict the number of search results. It is important to note that cited references from relevant retrieved articles were also chosen based on the trial selection criteria.

Table 3 Provides an indication of how key words used employed during searching.

Database searched

Search Number

Search Terms including (AND/OR)

Results

Articles Retrieved

PubMed

1

((bevacizumab OR Avastin OR RhuMab vegf OR anti-vegf OR anti vascular endothelial growth factor)) AND FILTERS

467

30

2

(#1) AND Phase I

214

5

3

((bevacizumab OR avastin OR rhumab vegf OR anti-vegf OR anti vascular endothelial growth factor)) AND preclinical AND [Pub date]

341

12

4

(#1) AND Phase II

140

16

5

(#31) AND Phase III

62

20

Table 4 lists the filters utilised during data collection

Filters employed

Clinical Trials

Publication date: 1992-2010

Trial selection

All abstracts of retrieved articled were read and those that met the inclusion criteria were identified and were chosen to be analysed. The criteria differed depending on if the study was a preclinical or clinical study.

Preclinical inclusion criteria

Type of study should include in vitro, in vivo

The study must feature or mention humanised monoclonal antibody ( RhuMAb)

Number of animals and their species must be clearly stated

Routes of administration, duration of treatment and observation must be stated

Studies investigating any rhuMAb associated risks including:

Acute (single dose) toxicity

Repeat-dose toxicity

Reproductive toxicity

Genotoxicity

Tumorigenicity

Sensitisation/immunosuppression and stimulation

Local and other specific adverse event

Safety pharmacology

Clinical trial Inclusion Criteria

Study must be primarily about Bevacizumab

Study must be peer reviewed

If indication specified must be: CRC, glioblastoma, mBC, RCC, NSCLC and study should have been carried our prior to licencing for that particular indication.

Adequately designed Phase II and III

Study purpose must be clearly expressed.

Adequate primary and secondary endpoints clearly expressed

RESULTS

Introduction

At total of 20 trials were selected to be included in this review. They consisted of 5 preclinical, 2 phase I, 9 phase II and 6 phase III trials. Due to time constraint, only the trials reviewed by the FDA were chosen. In general trials that need not progress past Phase II were rejected. Bevacizumab plus Gemicitabine in the treatment of pancreatic cancer was selected because it progressed into phase III testing however it was not reviewed by the FDA because its primary objectives were not met. The complete documentation of all the results can be found in the appendix section D and there the manufacturer sponsored trials are distinguished from independently sponsored trials.

Preclinical analysis

Chemical and Pharmaceutical analysis

Drug Substance

Nomenclature

Description

Manufacture

Specifications

Stability

INN Name: Bevacizumab

Brand name: Avastin

Chemical name: Recombinant humanised monoclonal antibody to VEGF

Other names: rhuMAb VEGF, anti-VEGF

Avastin: IgG1k type antibody with a total molecular weight of 149kDa

It is a humanised derivative of a murine monoclonal antibody. 93% of the overall protein sequence defined by the human frame work. Murine light and heavy chain CDR sequences constitute the other 7 %37.

Avastin derived from the humanisation of murine monoclonal antibody (muMAb A4.6.1).

Six A4.6.1 CDR sequences inserted into human Fab framework in place it’s the human CDR counter parts.

Plasmid encoding Avastin introduced into Chinese hamster ovary parental cells by process of lipofection. Isolated cells selected secretion of bevacizumab in presence of increasing concentrations of methotrexate(MTX)37.

Subsequently sub-clone 107N used in production of bevacizumab in phase I and II trials

This cell clone was cloned in presence of higher concentration of MTX to produce high yielding G7 cell line; G7 clone was then used in the production of Avastin in Phase III trials37.

G7 clone was expanded produce a two-tiered cell bank system of the master cell bank (MCB) and the working cell banks (WCB), which is prepared from the MCB

Cells from the WCB used in fermentation and purification processes of Avastin37.

Samples bevacizumab characterised and compared and a reference standard product was produced from both 107 cell line and G7 cell line. These standards used to assess the drug product used in phase I, II and II clinical trials37.

All batches containing Avastin subject to certain specification tests including: identity, osmolarity, pH and purity tests and potency assays among others38.

Storage time of bevacizumab proposed at 24 months at -20 ⁰C ± 5⁰C or

45 days if stored at 5⁰C ± 3⁰C allowing for 5 freeze/thaw cycles37.

Drug Product

Formulation

Stability

Drug developed as an intravenous solution each vial containing (100mg/4ml) or (400mg/16ml) of Avastin38.

Constitution during clinical trials

Phase I and early phase II: 10mg/kg bevacizumab, 10mM histidine, 100 mg/ml trehalose dehydrate and 0.02% polysorbate 20

Late phase II and Phase III: 25 mg/ml, 51mM sodium phosphate, 60 mg/ml trehalose dehydrate and 0.04% polysorbate 20 ( same formulation in marketed product)37.

Shelf life proposed to be 24 months when stored at 5⁰C ± 3⁰C37.

Pharmacology, Pharmacokinetics & Toxicolology

Preclinical results will be separated into three broad categories, studies involving pharmacodynamics, pharmacokinetics studies and toxicology studies. . The studies involving muMAb investigated primarily on its pharmacology, although there was a study that estimated the dose range of Avastin for use in clinical trials. Pharmacokinetics of bevacizumab was assessed in animals showing affinity for the monoclonal antibody such as cynomolgus monkeys and rabbits (although to a lesser degree). Studies were also carried out in mice and rats, although these had no affinity for Avastin. Toxicology studies were also investigated in cynomolgus monkeys.

Pharmacodynamics  

This study attempted to make clear the physiological role of VEGF and to determine if monoclonal antibodies would be capable of neutralising VEGFs biological activity. Several monoclonal antibodies including muMAb were assessed both in vitro and in vivo assays. All the monoclonal antibodies generated including muMAb (8 x10-10M) had a high affinity for VEGF, (dissociation constant ranged from 2.2 x10-9- 4 x 10-10)30. MuMAb was also reported as having high antigen specificity because of its low binding affinity for other growth factors such as PDGF or EGF was low. MuMAb was able to bind to all 3 types of VEGF isoforms ( VEGF 121 ,165 and 189) unlike the other isoforms, suggesting that the its binding site was located within the first 121 amino acids of VEGF protein30. The in vitro assay confirmed that VEGF stimulated growth of vascular endothelial cells in ACE cells. However muMAb A461 blocked this induced proliferation at much lower concentrations compared to the other monoclonal antibodies30. An in vivo assay demonstrated that muMAb A4.6.1 completely blocked VEGF induced vascular leakage at a 1: 10 molar ratio (VEGF to monoclonal antibody)30. No other MAb were this active. Another in vivo assay assessing the effect of muMAb A4 6.1 on angiogenesis, observed that a greater concentration was needed to block angiogenesis than was required to block vascular leakage. It was observed that the minimum concentration needed to for maximum inhibition of angiogenesis was a 1: 250 molar ratio of VEGF to muMAb A4.6.130.

A study was carried out to prove the role of VEGF in tumour angiogenesis in vivo using anti –VEGF monoclonal antibody. VEGF producing tumour cell lines, A673 rhabdomyosarcoma, G55 glioblastoma multiforme and SK-LMS-1 leiomyosarcoma, were injected into nude mice39. MuMAb was then administered as doses of 10ug, 50ug, 100ug, 200ug and 400ug twice weekly39. MuMAb inhibited the growth of the A673 and G55 cell producing tumours. Doses as little as 10ug led to significant inhibition and at doses between 50-100ug the maximum inhibitory effect was observed39. SK-LMS cells were inhibited by muMAb too, although it took while to be measured because of its slow proliferation rate. Besides this, growth inhibition was maintained throughout the study resulting in significant decrease in tumour mass in all three cell line types39. In situ hybridization of the muMAb treated tumours revealed that VEGF expression was its highest in tumour cells located nearby necrotic cells however vascular density was decreased. Microscopic examination of the tumours confirmed that muMAb treated tumours had a decreased density in vascular cells39. The most profound of this effect was seen in A673 cell line tumours.

Another study carried out both in vitro and vivo tests comparing the efficacy and potency of muMAb A4 6.1 and RhuMab VEGF (Avastin), a humanised version of the murine antibody. Results from in vitro tests showed that both monoclonal antibodies are equivalent both in potency and efficacy29. In both cases 90% inhibition of VEGF induced bovine capillary endothelial cell proliferation was achieved at 500ng/ml29. Similarly a separate in vivo test showed that both monoclonal antibodies have comparable inhibitory effects on tumour growth. Each monoclonal antibody was injected into nude mice containing A673 rhadomyosarcoma tumour cell lines at doses of 0.5 and 5 mg/kg. Compared to the control group muMAb significantly reduced tumour weight by 85% and 93% over a four week period29. Similarly for Avastin treated mice, tumour weight decreased by 90% and 95% over the same time period29.

Pharmacokinetics

A pharmacokinetic study was carried out in cynomolgus monkeys, rats and mice in which Avastin was administered by IV and subcutaneously (S.C.) and its disposition in humans was predicted using allometric scaling. Mice were administered a single Avastin dose of 9.3mg/kg, as either IV or s.c. injections and rats were injected a single IV bolus dose of either 0.66 or 10mg/kg Avastin40. Cynomolgus monkeys were randomised into 4 groups and given either IV injections of 2, 10, or 50mg/kg or 10 mg/kg S.C. injection40. For all animals, serum samples were harvested at regular intervals and assayed by an ELISA.

At an IV dose of 10mg/kg, (9.3 mg/kg in mice), clearance was 15.7ml/day/kg in mice, 5.59ml/day/kg in cynomolgus monkeys and 4.83ml/day/kg in rats40. For all specimens the terminal half-life ranged between 6 and 12 days at dose of 10mg/kg. The predicted serum clearance in humans was set at 12 days40.

Tissue distribution of IV bolus of Avastin in rabbits was measured 2 hours after administration using a gamma counter40. Radio-activation observed that Avastin was primarily located in the plasma with lower amounts in highly perfused tissues such as kidneys, heart and lungs40. After 48 hours radio activation was still highest in the plasma but now higher levels of radio activity were observed in the testes, kidneys, bladder, pancreas, lungs and heart41. This suggests that VEGF may be highly expressed in these organs.

Toxicology studies

The preclinical safety evaluation of bevacizumab included screening of human, cynomolgus monkey, and rabbit tissues for cross-reactivity study or non-target tissue binding at concentrations of 10ug and 400ug/ml42. No tissue cross-reactivity or non-target tissue binding was observed.

A multiple dose toxicity study was performed on healthy cynomolgus monkeys to determine the toxicity of bevacizumab following a twice weekly IV bolus dose. Cynomolgus monkeys were randomly assigned into 4 groups and given a dose of 2, 10 or 50mg/kg for a period of 4 or 13 weeks. At regular intervals, clinical measurements and observations and blood samples as well as other tests were performed. After treatment with Avastin, no signs of toxicity were observed. Furthermore there were no effects on body weight, blood pressure measurements, electrocardiograms, haematology or bone marrow parameters. In vitro test showed that Avastin is compatible with human serum and plasma as it did not induce haemolysis. The study demonstrated that an IV bolus twice weekly was well tolerated and no clinical signs of toxicity were observed at doses up to 50mg/Kg. However phsyeal dysplasia was observed doses between 10-50mg/kg in both male and female, although at lower levels, cynomolgus monkeys42.

Guidance on clinical trial end points

Endpoints

Overall Survival

Overall survival is as time dependent end endpoint that is defined as the time from randomisation till death and it is measured in the intent-to-treat population43. One of the advantages of this endpoint is the ease and precision of its measurement. The simplicity of measuring this endpoint, i.e. the documentation of time-of-death also eliminates bias usually associated with measuring endpoints hence blinding is not essential in studies that use this as its primary end point43. However, overall survival should be evaluated in randomisation studies as in this setting it is most reliable. A second advantage is that a study demonstrating statistically significant improvement in overall survival often correlates well with clinical significance and benefit, providing that its toxicity profile is acceptable44. So it follows that overall survival is universally accepted as a direct measure of benefit. The main disadvantages of this endpoint are that it requires large studies in order for its statistical analysis to be relevant. Additionally, its accuracy is also affected by cross over, sequential therapy and the documentation of non-cancer deaths.

Progression Free Survival

Progression free survival (PFS) is defined as the time from randomisation to objective tumour progression or death43. It is often used as a surrogate for accelerated approval or regular approval of cancer drugs. PFS is capable of measuring tumour growth and determining survival benefit of a medication45. Besides this, PFS also measures the stability of a tumour disease. The assessment of PFS is generally objective and quantitative, thus its accuracy is often validated. Unlike overall survival where large study sizes are needed, PFS can still be valid in smaller sample sizes and shorter follow-ups. However, it should be noted that inadequate follow-ups and consequently missing data could lead to bias in PFS measurements and possibly and overestimation of the PFS in that study43. Another limitation of PFS arises from the definition and subsequent variation in the definition of tumour progression. There are no standard regulatory criteria for defining progression hence some studies may apply a variety of different criteria such as RECIST43. The FDA recommends that trial protocols should communicate precisely the definition of tumour progression and methodology of preventing bias from arising43. One such bias is that of assessment bias which is particularly prevalent in open labelled studies. Consequently the blinding during studies is preferred or at least assessment should be subjected to independent investigators. Finally, PFS may not correlate entirely with overall survival in some malignant cases. This is usually due to the data collected being insufficient to adequately evaluate the correlation between survival benefits and PFS. Thus, it follows that PFS is not statistically validated as a surrogate for survival in all cancer studies.

Objective Response Rate

Objective response rate is defined as the proportion of patients who show tumour size reduction of a pre-specified amount and for a certain period of time43. Therefore ORR is a direct measure of a drug’s therapeutic effect, although not a comprehensive measure of its activity. Consequently it can be evaluated in singe arm studies as well as randomisation studies. A component of ORR is Response duration which is defined as the time of initial response (complete or partial) until recorded tumour progression46. It should be noted that stable disease is not a component of ORR as it can reflect both antitumor activity as well as natural history of disease. Rather stable disease can be better assessed by PFS. The FDA recommends that ORR be measured according to standardised response criteria e.g. RECIST criteria and subsequently such criteria should be stated in the protocol prior to the start of the study43.

Time to progression response

Time to progression response (TTP) similarly to disease free progression is defined as the time elapsed between treatment initiation and tumour progression43. However TTP does not take into consideration deaths that occur before tumour progression within its analysis, rather deaths are censored i.e. not recorded. It is often used as a surrogate marker for the approval of drugs but in some cases it may be serve as a primary end point for drug approval43.

How Endpoints are assessed

Intention-to-treat analysis

Intention to treat analysis (ITT) is a statistical analysis that includes in its population every subject that is randomised according to randomised treatment assignment whether or not the subject actually receives the treatment47. In a way it is reflective of clinical practice as issues such as non-compliance, withdrawal and protocol deviation also occur in clinical settings. ITT eliminates prognostic differences in RCTs that may be introduced by non-compliance and missing outcomes and thus within reason gives an unbiased estimate of treatment effect47. Furthermore, due to its all-inclusive nature, it preserves the sample size and thus also preserves the statistical power of the study47. Besides this, ITT prevents type I errors also known as false positives47. One of the limitations of ITT is that including non-compliance and drop outs may cause an underestimation of treatment effect as these issues may be independent of the treatment thus making ITT more susceptible to type II errors (false negatives)47. Despite this, the FDA recommends that clinical trials analyse both subjects that complete the study and the entire population as a whole; ITT achieves this objective43.

RECIST Criteria

RECIST criteria are a set of guidelines that provides a standard approach for clinicians to measure solid tumours and changes in those tumours during a clinical trial. It is therefore a useful tool in evaluating endpoints such as objective response, tumour progression or time to tumour progression since these clinical outcomes are based on the anatomical change in tumour size during a study46.

Kaplan- Meier survival analysis

Kaplan-Meier approach is a method that is used in time-to event models. It measures the distribution of time between two events, taking into account censored data48,49. Thus it is idea estimator of overall survival and progression free survival. It follows that with Kaplan-Meier, it is possible to plot a time-to endpoint curve providing precise documentation of the endpoint during the study. From this curve it is possible to estimate mean and median survival taking into account stand error and 95% confidence interval48. Furthermore time-to event curves can be analysed by Cox proportional hazards regression (Hazard ratio) producing a risk of complication comparison between two treatment groups being analysed. Thus in a clinical trial where the endpoint is survival, the hazard ratio estimates the likely of death in the treated group compared to the control subject at any given point in time.

Toxicity

Majority of drug trials use National Cancer institute common toxicity criteria (CTC) to assess the adverse effects of drugs administered during clinical trials50. CTC provides a standardised classification for assessing these adverse events51. It utilises a grading scale that ranks an adverse event based on its severity. The grades are as follows (table 5):

Table 5 describes the different grade levels of National cancer institute common toxicity criteria

Grade

Definition

0

No adverse event or within normal limits

1

Mild adverse event

2

Moderate adverse event

3

Severe and undesirable adverse event

4

Life-threatening or disabling adverse event

5

Death related to adverse event

Phase I clinical trials

The primary purpose of the following two dose escalation studies was to evaluate the pharmacokinetics of bevacizumab as well as its safety and tolerability. These phase I trials set the foundation for the dosage regimen administered in subsequent phase II and III trials.

Study 1

AVF0737g

In this phase I trial, bevacizumab was administered by IV infusion over a range of doses from 0.1 to 10 mg/kg in 25 patients over a trial period of 70 days52. Doses were given at days 0, 28, 35 and 42. Pharmacokinetic properties were sampled at regular intervals and toxicity was assessed in accordance with the NCI common toxicity criteria52.

Grade 3 or 4 adverse events (according to CTC scale) were reported in 4 patients including anaemia, dyspnoea and two episodes of bleeding52. These adverse events were classed as disease-related and hence not associated with bevacizumab treatment. Other Grade 1-2 adverse events were also reported, the most common of which included asthenia, headache and nausea. Other adverse events included mild rise in blood pressure and fever52. Since bevacizumab was well tolerated at doses ranging up to 10mg/kg for all patients, it was concluded that bevacizumab could be safely administered without dose limiting toxicity as those doses52.

In this study no patient experienced either an objective complete response or a partial response. However 2 patients showed minor regression of their metastases after completion of bevacizumab therapy suggesting potential anti-tumour activity52. Of the remaining patients, 12 patients experienced stable disease over the study period while the rest (11) demonstrated disease progression52.

Pharmacokinetics studies showed that the maximum concentration (Cmax) of bevacizumab increases in a dose related manner. The mean observed Cmax ranged from 2.8 ug/ml (0.1mg/kg) to 284 ug/ml (10mg/kg)52. However, there was not significant accumulation of bevacizumab in the body during multi-dosing at days 28, 35 and 4252. The mean clearance of bevacizumab was lower in the 10mg/kg group (2.8ml/kg/day) than in the 0.1 mg/kg (9.13ml/kg/day), additionally, at doses greater than 0.3mg/kg, the kinetics of bevacizumab was reported as linear with a half-life of approximately 21 days52. Thus, it was concluded that bevacizumab had low clearance and low volume of distribution properties.

Study 2

In a subsequent phase 1b trial, 12 patients were evaluated for tolerance and toxicities in combining bevacizumab 3mg/kg IV infusion with three other standard chemotherapy regimens used in lung, breast and GI cancers (table 5)53.

Table 6 describes the treatment regimen for study 2

Treatment regimen

Chemotherapy dose

Chemotherapy Schedule

Bevacizumab schedule

Doxorubicin 50mg/m2

Once every 4 weeks

3mg/kg weekly, weeks 1-8

Carboplatin AUC= 6T

Paclitaxel: 175 mg/m2

Once every 4 weeks

3mg/kg weekly, weeks 1-8

Fluorouracil (5-FU) 500mg/m2

Weeks 1-6

3mg/kg weekly, weeks 1-8

With Leucovorin 20 mg/m2

Every 8 weeks

Toxicity data showed that there were no significant toxicities or pharmacological interactions when bevacizumab was combined with other regimens. Grade 3 adverse events reported included leucopoenia, diarrhoea thrombocytopenia; however these toxicities at the time were attributed to the chemotherapy rather than bevacizumab53. In this study, no patient experienced any bleeding complications. Thus it was concluded that bevacizumab has a favourable tolerability when combined with other chemotherapy drugs.

The concentration time profile was consisted with the initial phase 1 study, when bevacizumab was given as a single agent. Furthermore the terminal half-life of bevacizumab was reported as 13 days53.

Due to the small sample size of patients, no conclusions could be drawn concerning the anti-tumour activity of bevacizumab53.

Rationale for chosen disease types

In a study exploring the correlation between p53, expression of VEGF and vessel count in colorectal cancers, established that VEGF is highly expressed in this cancer type. Furthermore and VEGF expression is at its highest in metastatic forms of colorectal cancer54. As wells colorectal cancer, VEGF is highly expressed in other cancer types. Several studies have shown that VEGF is highly expressed in a wide range of cancers including lung, breast, gastrointestinal tract, kidney, bladder, ovary, endometrium and several intracranial tumours including glioblastoma multiforme33. Thus it follows that Avastin be tested in the various cancers above to determine anti-tumour activity.

Metastatic colorectal cancer

Study 1 -Phase II

AVGF0780g

Study AVGF0780 was a randomised, open-label and multi-dose trial designed to evaluate the safety and efficacy of bevacizumab combined with fluorouracil (5-FU)/leucovorin (LV) in patients with previously untreated CRC. 104 eligible subjects were randomised to one of 3 arms:

Arm 1-: Control (5-FU/LV alone); 36 patients assigned

Arm 2-: 5-FU/LV + bevacizumab 5mg/kg administered once every 2 weeks; 35 subjects assigned

Arm 3-: 5-FU/LV + bevacizumab 10mg/kg administered once every 2 weeks; 33 subjects assigned

Patients taking the control medication could switch to treatment arm upon disease progression55.

The primary endpoints were progression free survival and objective rate response either complete or partial. Primary analyses of tumour status and disease progression was carried out independently by and independent review facility. Secondary endpoints included overall survival and duration of response55.

The efficacy results are displayed in table 6. Progression free survival was significantly longer in patients taking 5-FU/LV plus Avastin (5 mg/kg) compared to patients not receiving Avastin55. This meant that the resulting risk of disease progression was reduced by 54% in patients receiving Avastin 5mg/kg. However there was no significant difference in objective response rate and overall survival between the two patient groups55. There was also no significant difference between patients receiving Avastin (10mg/kg) plus 5-FU/LV and patients not receiving Avastin55.

55Table 7- Describes the efficacy results of Study AVF0780g

Endpoint

Control 5-FU/LV alone

N=36

Bevacizumab 5mg/kg + 5-FU/LV

N=35

Bevacizumab 10mg/kg + 5-FU/LV

N=33

Progression Free survival

Median PFS (months)

Hazard ratio

p-value

5.2

-

9.0

0.44

0.005

7.2

0.692

0.217

Objective response rate

Objective response

p-value

17%

-

40%

0.029

24%

0.434

Overall survival

Median (months)

Hazard ratio

P-value

13.6

-

-

17.7

0.521

0.073

15.2

1.009

0.978

Although bevacizumab was well tolerated, there was significantly more patients who experienced more adverse events while on bevacizumab (p=0.042). However this may be due to cross over of patients from the control arm to treatment arms leading to increased study duration of the treatment arms. Common adverse events observed included hypertension, thrombosis, fever, headaches rash and chills55. Some patients also experienced bleeding, the most common being epistaxis. Three patients on 10mg/kg experienced GI perforation of grade 3 -4 based on the common toxicity criteria55.

Study 2 -Phase II

AVF2192g

Study AVF2192 was a randomised, blinded placebo-controlled study designed to evaluate the efficacy of bevacizumab combined with bolus Fluorouracil/Leucovorin as a first-line treatment in CRC. 209 eligible subjects over the age of 65 were randomised into one of two arms:

Arm 1: 5-FU/LV given weekly for six weeks

Arm 2: -5-FU/LV + bevacizumab (5mg/kg) administered once every two weeks

Treatment cycle lasted 8 weeks for a total duration of 96 weeks.

Patients who showed either complete response or unacceptable toxicity could discontinue 5–FU/LV and just receive bevacizumab56.

The primary endpoints were overall survival estimated using Kaplan-Meier. Secondary endpoints included progression free survival (PFS), objective rate response (ORR), duration of response and Quality of life. PFS and ORR were assessed by both the investigator and independently by the independent radiology facility (IRF) using the RECIST criteria56. Safety and tolerability were also assessed by the NCI common toxicity criteria.

Efficacy results are expressed in table 8. The results demonstrated that bevacizumab (5mg/kg) + 5-FU/LV significantly prolonged disease progression by 3.7 months and reduced the risk of disease progression by 50% when compared to 5-FU/LV + placebo. Bevacizumab + 5-FU/LV also increased rate of response, duration of response and prolonged duration of survival, although these observations did not show statistical significance, hence a definite treatment-effect relationship was not established56. Bevacizumab had no negative impact on quality of life, which implies that it had some beneficial effect.

Table 8 – Describes the efficacy results of study AVF2192

Endpoint

Control 5-FU/LV +placebo

N=105

Bevacizumab (5mg/kg) + 5-FU/LV

N=104

Overall survival

Median (months)

Hazard ratio

P-value

12.9

-

-

16.6

0.79

0.160

Progression Free survival

Median PFS (months)

Hazard ratio

P-value

5.5

-

-

9.2

0.50

0.0002

Objective response rate

Objective response

Complete response

Partial response

P-value

15.2%

0

15.2

26.0%

0

26.0

0.055

Duration of response

Duration of response (months)

Hazard ratio

P-value

6.8

-

-

9.2

0.42

0.088

Compared to the placebo group, patient on bevacizumab showed a 16% increase in observed grade 3 or 4 adverse events (71% vs. 84%). However drug was well tolerated in the all patient demographics56. Bevacizumab associated-toxicities bleeding, thromboembolism and proteinuria were observed in the treatment group however there were no significant increases when compared to the placebo contro56l. Two patients in the treatment group experienced GI perforation of which one was fatal. Finally the most significant difference in toxicities between treatment and placebo group was seen in level of hypertensions. Grade 3 hypertension was seen in 16% patients on BV/5FU/LV compared to 3 % of placebo group (5FU/LV)56.

Study 3-Phase II

E2200

This was a non-randomised, open label phase II trial that was designed evaluate efficacy of combining high dose bevacizumab (10mg/kg) with irinotecan, 5-fluorouracil, leucovorin (IFL). In this single arm study 92 eligible subjects were recruited however only 81 were available for efficacy analysis and 87 used in toxicity analysis57. The overall response rate was 49.4%, of which 6.2% were complete responses57. The median overall was 26.3 months, 1-year survival 85% while the median progression free survival was 10.7 months57. Grade 3 or 4 bevacizumab associated toxicities included hypertension (2.3%), Thrombosis (11%), GI perforation (2.3%)57.

Study 4-Phase III

AVG2107g

This phase III trial was designed to compare the efficacy of bevacizumab (BV) combined with Irinotecan, fluorouracil and leucovorin (IFL) and IFL +placebo for previously untreated metastatic cancer. As such, the study was placebo-controlled and randomised. 813 patients where initially split into 3 arms:

Arm 1: IFL (once weekly for 4 weeks) +placebo

Arm 2: IFL + BV (5mg/kg every 2 weeks)

Arm 3: 5-FU/LV + bevacizumab (5mg/kg)

Interim analysis was performed to assess safety of Bv+IFL. The regimen was proved safe thus Bv+Fu/LV arm was discontinued and patients on that arm were randomly transferred to either of the other arms58.

The primary end point was overall survival and the secondary endpoint were progression-free survival (PFS), response rate (RR), duration of response (DoR), safety and Quality of Life (QoL)58.

Patients taking bevacizumab combined with IFL had a significant longer duration of survival (20.3 months) than patients taking IFL +placebo (15.6)58. This meant that patients on the treatment arm had a 34 % lower risk of dying compared to their counterparts on the control arm (Figure 5)58. It was also the case that PFS, duration of response and response rate was significantly greater in patients on IFL + bevacizumab compared to patients taking IFL + placebo (table 9)58.

The adverse events experienced by patients were consistent with information derived from previous trials. Overall 10% more patients taking IFL+ bevacizumab experienced Grade 3 /4 AE (84.9%) than patients taking IFL + Placebo (74%)58. Table 10 compares common adverse events experienced by both treatment groups. It should be noted that grade 3 hypertension was significantly more likely to occur in patients taking IFL + bevacizumab than the control58. GI perforation (1.5%) was only observed in arm 2 (IFL + bevacizumab) resulting in one patient dying as a result58.

Figure 5: Comparison of OS in patients taking IFL + placebo or IFL + bevacizumab

Table 9 compares the efficacy of IFL + placebo and IFL +bevacizumab

Analysis of Efficacy

Endpoint

IFL plus Placebo

(N= 411)

IFL plus bevacizumab

(N=402)

P value

Median survival (months)

15.6

20.6

<0.001

One year survival rate (%)

63.4

74.3

<0.001

progression free survival (months)

6.2

10.6

<0.001

Overall response rate (%)

34.8

44.8

0.004

Median duration of response (months)

7.1

10.4

0.001

Significant adverse effects

Adverse event

IFL plus Placebo (%) (N= 397)

IFL plus bevacizumab (%) (N= 393)

Any grade 3 or 4 adverse event

74

84.9

Grade 3 Hypertension

2.3

11

Grade 3 or 4 Leukopenia

31.1

37

Grade 3 or 4 diarrhoea

24.7

32.4

Table 11 comparison of selected adverse events in patients taking IFL+ placebo or IFL+ bevacizumab

Study 4-Phase III

E3200

This study is a randomised and open labelled study designed to compare the efficacy of bevacizumab combined with FOLFOX4 and FOLFOX4 alone and Bevacizumab for previously treated metastatic colorectal cancer59. 829 patients were split into 3 arms:

Arm 1: FOLFOX4 (Control)

Arm 2: Bevacizumab (10mg/kg) + FOLFOX4 every 2 weeks

Arm 3: Bevacizumab (10mg/kg) alone

The primary end point was overall survival while PFS, overall response and toxicity were secondary endpoints59.

BV + FOLFOX4 improved overall survival by 2.1 months (figure 6) when compared to the control: FOLFOX4. Although modest, this improvement is significant and represents 25% reduction in risk of death if taking BV +FOLFOX4 than FOLFOX4 alone. There were no differences in OS between taking BV alone or FOLFOX4 alone. The PFS and overall response rate were also significantly greater in BV +FOLFOX4 group than the control group (figure 7, 8) hence providing support for the notion that BV+FOLFOX4 may be more effective as 2nd line therapy for mCRC than the standard treatment of FOLFOX4. Bevacizumab as a monotherapy did not improve PFS; its overall response rate was not reported in this study and subsequently it was discontinued prematurely59.

Overall patients on Bevacizumab +FOLFOX4 experienced significantly more grade 3 / 4 adverse events (49.5%) than patients taking FOLFOX4 alone (36.1%)59. The overall incidence in patients taking bevacizumab alone was 36%. Figure 9 displays the most common and significant grade 3 /4 adverse events observed in patients. Bevacizumab associated toxicities observed includes hypertension and bleeding. There were 6 reports of GI perforations, three in each of the bevacizumab containing arms and this resulted in two deaths59.

Figure 6: Comparison of Overall survival between FOLFOX4+BV FOLFOX4 and BV alone

Figure 7: comparisons of progression free survival between FOLFOX4 +BV, FOLFOX and BV alone

Figure 8: comparison of response rate between FOLFOX4 +bevacizumab and FOLFOX alone

Significant Adverse events observed in patients taking FOLFOX + BV, FOLFOX and bevacizumab for mCRC

Figure 9 comparisons of significant adverse events between treatment groups in study E3200

Non-small cell lung cancer (NSCLC)

Study 1 -Phase II

Table 9- describes efficacy results of study AVF0757g

Endpoint

Control Carboplatin/

Paclitaxel

N=32

Bevacizumab 7.5mg/kg + C/P

N=33

Bevacizumab 15mg/kg + C/P

N=34

Time to disease progression

Median TTP (months)

p-value

4.2

-

4.3

-

7.4

0.023

Objective response rate

Objective response

18.8%

28%

31.5%

Overall survival

Median (months)

P-value

14.9

11.6

0.84

17.7

0.63

Adverse events were categorised using NCI common toxicity criteria. The most serious adverse events (grade 3 /4) in patients taking bevacizumab included leukopenia, thrombosis, Haemoptysis, bleeding60. 6 patients experienced life threatening bleeding of which 4 were fatal. 4 out of the 6 bleeding events were in patients with squamous carcinomas. An exploratory analysis revealed that the risk of serious bleeding was reduced in non-squamous histology patients60. Common bevacizumab associated events were also observed including hypertension, epistaxis, and rashes.

Study 2-Phase III

E4599

In this phase III trial the efficacy of bevacizumab combined with Paclitaxel-Carboplatin was compared to Paclitaxel-Carboplatin alone for the treatment of NSCLC. 878 patients were randomly separated into one of two arms. Bevacizumab dose was 15mg/kg and was given once every 3 weeks61. Patients with squamous cell carcinoma histology were excluded because study AVF0757g showed that patients in this category had an increased risk of pulmonary haemorrhaging when bevacizumab was administered61.

Table 12 description of marketed formulations of Avastin

Dosage Form and strength

100mg/4ml, single use vial

400 mg/16ml, single use vial

Table 11 summary of serious and common adverse effects of Avastin

Selective list of Adverse Effects

Serious

Arterial thromboembolism

Uncontrolled hypertension

Hemorrhage

Gastrointestinal perforations

Common

Epistaxis; headache; hypertension; rhinitis; proteinuria; taste alteration, dry skin, back pain,

The approval of Avastin for the treatment of metastatic breast cancer was withdrawn by the FDA in 20011, 3 years after initially granting its approval. The FDA careful review of subsequent studies, that Avastin neither shows safety or efficacy for use in metastatic breast cancer treatment68. Initially approval of Avastin for metastatic breast cancer was granted under the accelerated approval program providing that subsequent phase IV studies showed clinical benefit. However, this was not the case as the two phase IV studies carried out by Genentech proved Avastin offers little improvement in clinical outcomes compared to the standard treatment.

DISCUSSION

The methodology used in this project relied heavily on published peer reviewed literature but it was sufficient to accumulate the relevant data needed to understand the rationale behind the development and subsequent approval of Avastin for each of its different indications. The accrual of relevant data made it possible to construct an overall picture of the progression of Avastin from its design to its introduction onto the market. However, the organisation of the data could be improved by pooling and analysing the pooled data rather than individual trials hence allowing for a better evaluation of the trials. Furthermore it highly probable that some literature was not collected using my search strategy hence possibly affecting my analysis. It is important to note that the majority of trials reviewed were directly related to Avastin, it is possible that certain literatures, indirectly citing Avastin, could have been relevant to this study. Nonetheless the applied methodology was sufficient to achieve the aims and objectives of this project.

Pre-clinical analysis

Pre-clinical studies demonstrated that Avastin and its murine parental homolog A4 .6.1 have high affinity for human VEGF isoforms and inhibit the biological activity of VEGF resulting in decreased angiogenesis in tumours. Majority of pharmacological studies were performed using A4.6.1 however its equivalence with Avastin in both potency and efficacy was adequately documented37. The findings from the studies reviewed show that Avastin inhibits primary growth of tumours as well as metastases. This may be explained by the fact that Avastin inhibits VEGF induced vascular permeability and VEGF induced angiogenesis37.

Avastin’s pharmacokinetics was initially established in rats, mice, rabbits and cynomolgus monkeys by means of single dosing study. A multi-dosing toxicology study on cynomolgus monkeys provided pharmacokinetic results that supported the single dosing study. The aim of these studies was to predict human exposure to Avastin and to determine the initial dose or doses to be used in clinical trials. The studies showed that Avastin has a low clearance value and low volume of distribution as its distribution was limited to the vasculature37,40. Additionally Avastin showed a consistent disposition between the doses of 2mg/kg and 50mg/kg in the multi-dose study37. The predictive human clearance value was comparable to observed human clearance values suggesting that clearance mechanism between species is similar40. Avastin’s metabolic profile proved to be similar to other antibodies in its family class of IgG 37.

In the toxicology study involving cynomolgus monkeys, Avastin was well tolerated leading to few adverse events. However, physeal dysplasia was experienced by both male and female monkeys in a dose related manner. This dose dependent toxicity followed blood vessel development in bone growing plates, suggesting that Avastin should not be recommend to patients with bone disease so as not to worsen the disease42. Avastin decreased the ovarian and uterus weights of female monkeys inhibit the maturation of ovarian follicles and copora lutea. Although this toxcicity was dose dependent, the findings suggest that Avastin should not be administered during pregnancy as it may cause damage to the mother and the possibly foetus42. A separate study observed that wound healing was delayed in rabbits at doses of 0.5mg/kg (data not shown)37. This finding strongly suggests that future exclusion criteria should include patients with major wounds or ulcers. Nonetheless, the well tolerated toxicity profile of Avastin was a key factor in its progression to clinical trials.

As far as can be determined the manufacturing process of Avastin for clinical trial testing was validated and the quality of production was controlled by adequate test methods and specifications37.

Phase I clinical trials

Only two phase I trials including study AVF0737g were reviewed because they were the most cited references in majority of the phase II trials reviewed in this project. Furthermore the searching strategy employed did not produce any other relevant phase I study. According to FDA reports, the pharmacokinetics of Avastin was well characterised based on 4 dose escalating studies, AVFO737g, AVF0776g, AVF0757g, AVF0780g, which establish the administration dose to be between 3mg/kg and 20mg/kg every 2 or 3 weeks interval37. This project did not have the capacity to evaluate this claim or analyse the pharmacokinetic data in detail. Study AVF0737g in which drugs doses were administered between 0.1 and 10mg/kg, the maximum tolerated dose was not established52. However based on this study the doses 5mg/kg and 10mg/kg were selected for phase II trial involving colorectal. Additionally this study estimates the half-life of Avastin to be approximately 20 days52.

Metastatic colorectal cancer

Colorectal cancer (CRC) is the third most common cancer worldwide and in 2008 it accounted for 10%of all cancers globally69. Besides this colorectal cancer is more prevalent in developed countries- almost 60% of all cases occur in developed Countries in contrast to the much lower incidence rates in developing countries. Incidence of colorectal cancer is significantly higher in men than in women, a 10 fold difference exists between the incidence in men and that in women69. Current trends show that the incidence of colorectal cancer is on the rise both in the developed world and similarly in developing countries69. These trends highlight the growing importance and dependence on chemotherapy drugs to combat this disease. Before the introduction of Avastin on to the market, the standard treatments of CRC were as follows.

Irinotecan, in combination with either infusional 5-FU/FA (FOLFIRI) or bolus 5-FU/FA (IFL) superseded 5-FU/FA alone as the first line treatment for chemotherapy-naive patients with mCRC37.

Oxaliplatin added to infusional 5-FU/FA (FOLFOX4) as a safe first line therapy for mCRC37.

To conclude, prior to the development and approval of AVASTIN, FOlFOX4 and were considered the first line treatment for mCRC. They both had comparable efficacy but alternate safety profile.

Five studies were reviewed, three of those studies were sponsored by Genentech and two were linked to NCI/ECOG collaboration. Study AVGF0780g and AVG2107g formed the basis for market authorisation of Avastin for use in first line treatment of mCRC. Study E2200 formed the basis for approval of Avastin as a second line treatment for mCRC. The published trials included in this review with the exception of study AVGF780g, were well designed including balanced population groups between treatment groups. Although the population group in phase III study AVG2107g was well balanced between the control and the treatment arm, in general patients treated with Avastin were relatively young (<59 years of age) with favourable ECOG status, usually 0. It is possible that this difference in demographic may have impacted on the prognosis of mCRC in a favourable direction during the study.

Phase III study AVG2107g showed that Avastin administered with bolus IFL prolonged survival of patients by 4.7 months compared to giving bolus IFL alone (20.3 vs 15.6 months)58. Study E2200 showed that the addition of Avastin to FOLFOX4 significantly improved the survival of patient who had progressed from 1st line therapy treatment by 2.1 months57. Although study met its primary objective of longer survival duration, its improvement was modest compared to the stand treatment. However, significant improvements in both PFS and ORR may have been a deciding factor in FDA approving Avastin for this indication.

The two phase II studies AVF0780g and AVF2192g provide good evidence that Avastin when combined with 5-FU based chemotherapy regimens improves on the clinical outcomes compared to giving 5-FU based chemotherapy alone37.

Non-small cell lung cancer

Phase III study E4599 assessing the safety and efficacy of Bevacizumab combined with paclitaxel and Carboplatin offered a significant improvement in OS, PFS and response rate compared to the standard treatment61. This demonstration of clinical benefit prompted the FDA to approve the treatment regimen70. However, this treatment regimen also increases the risk of drug toxicity related deaths which needs to be taken into consideration when offering treatment choice to patients61. Phase II trial AVF0757g contributed to the exclusion criteria documented in study E4599 as it discovered that bevacizumab treatment in patients with NSCLC, especially those with squamous cell cancer histology, increased their risk of fatal pulmonary haemorrhage60.

There are conflicting views as to whether Paclitaxel-carboplatin or paclitaxel- cisplatin, another standard cytotoxic regiment for NSCLC, should have been combined with bevacizumab. A study by rossell et al showed that paclitaxel-cisplatin combination increased duration of survival compared to paclitaxel-carboplatin71. However another study comparing 4 standard chemotherapy agents including cisplatin and carboplatin based regimens demonstrated that no significant advantages existed between the regimens72. In the light of these investigations, the trial investigators were justified in choosing to study the combination of bevacizumab and paclitaxel-carboplatin for NSCLC.

Recurrent Glioblastoma Multiforme

The FDA approved bevacizumab as a monotherapy for the treatment of recurrent glioblastoma multiforme through accelerated approval regulations. Approval was based on two phase II trials of which the pivotal trial was study AVF3708g. In the pivotal trial the efficacy end points were 6-months PFS and objective response. 6-month progression free survival (6-PFS) is regarded as a clinically relevant endpoint when evaluating treatments in recurrent glioblastomas as it is strongly correlated with overall survival45,73. To reduce the risk of bias associated with interpreting PFS values, the FDA strongly recommends that trials be randomised, controlled and blinded43,74. Study AV3708g was designed to be open labelled and non-comparative, thus rendering 6-PFS inadequate as a clinical endpoint. The FDA, therefore, based its accelerated approval on the objective response which it regards as an adequate surrogate endpoint provided that the objective response during the trial is greater than 30%45. Study AVF3708 met the criteria as its objective response was 28.2%; 97.5% CI, 18.5% to 40.3%66,74.

Issues related to development of Avastin

Several issues presented itself while assessing the clinical trials pertaining to Avastin. One of which was sponsorship bias. The vast majority of trials evaluated were sponsored by the pharmaceutical company Genentech. It is now well documented that industry funded trials are more likely to report positive outcomes than trials that are funded independently41. A systematic review looking into pharmaceutical industry funding and clinical trials results reported strong association between industry support and results that favour the industries’ interest. For example 17 studies reviewing pharmaceutical industry bias published since 2003 showed this positive correlation while 2 studies dispelled the notion75. Several primary studies, systemic reviews and meta analyses also confirm this clear association76. Sponsorship bias just like any other type of bias potentially skews the result outcome leading to a disparity between the trial out come and actual clinical presentation.

Another that issue is the expensive pricing associated with Avastin. As well as being new, Avastin is considerably expensive. According to a report, treating patients in the US for colorectal cancer may cost between $42, 800 and $55,000 a year77. This is clearly an expensive drug that is available only to those that can afford it. Due to its high cost and the fact that it only prolongs life rather than cure cancer, questions have been raised as to whether the drug provides good value for money and if it is actually beneficial to patients. This criticism was further compounded when the FDA revoked the use of Avastin in the treatment of metastatic breast cancer. It is worth noting that NICE does not recommend the use of bevacizumab due to it extravagant cost. NICE maintains the position that bevacizumab should not be funded by the NHS because it does not believe that the evidence supporting the using of Avastin shows adequate benefits for the patient to justify the extravagant costs78. Also the Scottish consortium does not recommend bevacizumab for use within the NHS for the same reasons stated above. This highlights the issue that the clinical benefits of bevacizumab although significantly better than the standard regimens may not be worth the extra price.