Final Report Summary - CARESS (A new methodology based on an advanced molecular probe for early detection of DPD enzyme deficiency in oncological patients, also enabling a personalised and effective drug management)
Executive Summary:
In our age, cancer is the most common cause of death in Europe. Amongst all the chemotherapy agents, 5-FU (5-Fluorouracil) has been in use for 40 years, and is one of the most successful and widely employed. In general, 5-FU is relatively well tolerated at standard doses, but around 8% of patients manifest a genetic variation that leads to a deficiency of an enzyme called Dihydropyrimidine dehydrogenase (DPD) that is crucial for the metabolism and deactivation of 5-FU. This causes several toxic reactions. The CARESS Project is a joint effort of 3 research Institutions from Italy and France together with 4 SMEs (small and medium sized entities) from Austria, France and Italy, to develop an accurate, easy-to-use and cost efficient technology for determination of the DPD activity level.
The protocol has been setup by the CNR and the spectrophotometric evaluation of the DPD activity has involved four main activities and has given rise to many others side activities, crucial for the optimization of the developing system and for the solution of the encountered critical points. The main activities managed for the development of the protocol have been: definition of the experimental protocol for the reaction of the selected Molecular Probe (MP) with DPD; selection of a suitable MP able to release, after reaction with DPD, easily detectable ions; assessment of the protocol for total inhibition of Thymidine Phosphorylase (TP) reaction; definition of the protocol for the spectrophotometric determination of ions released by MP. ICO has validated the full CARESS protocol with HPLC method and then they have proceeded with the validation with UV measurement.
From the technological point of view, LABOR consolidated of the prototype concept, the mechanical and electronic design (carried on partially during the P1), the selection of all the components, the manufacturing of all the components and the preliminary tests on the single components before the integration of the whole prototype. The final machinery can be considered a real automation device having to perform rapid and precise operations and displacement of fluids.
Finally, the field tests were to present the protocol followed with the CARESS device and the controls test needed for the validation of the prototype. It aimed to carry out and to evaluate both CARESS protocol and prototype, mimicking the real operating conditions within a diagnostic laboratory. The protocol developed by CNR was modified and adapted (in terms of volumes and timing of dispensation) to the prototype. All the experimental tests made at ICO for the validation of the prototype have followed the protocol. In order to define the performance of the CARESS machinery in the detection of DPD activity, and its potential deficiency, results obtained both manually and by Caress device were compared.
Project Context and Objectives:
In our age, cancer is the most common cause of death in Europe, with around 3,2 million new cases and 2 million deaths each year (7,6 millions in the world), with a terrifying increasing trend which will lead to 15 new million cases per year within 2020, as reported by WHO.
It’s the therapeutic approach used to fight this plague that is casting a doubt on the “Primum non nocere” principle; everybody who had the misfortune to get to know the severe side effects of chemotherapy- based approaches can understand how the border line between care and harm is often, dramatically, crossed.
Amongst all the chemotherapy agents, 5-FU (5-Fluorouracil) has been in use for 40 years, and is one of the most successful and widely employed, used in the treatment of breast, colon, and skin cancer, which are three of the most frequently occurring malignancies.
In general, 5-FU is relatively well tolerated at standard doses, but there are some important issues that must be duly taken into consideration:
• Around 8% of patients manifest a genetic variation that leads to a deficiency of an enzyme called Dihydropyrimidine dehydrogenase (DPD) that is crucial for the metabolism and deactivation of 5-FU. This causes several toxic reactions like mucositis, diarrhea, neutropenia, cerebellar ataxia, cerebellar dysfunction, and can even be fatal at the very first dose of 5-FU, with a mortality rate of about 0,5%. This deficiency must be properly identified before the beginning of the therapy, since the patient who has this disorder will never be aware of it until they suffer from a cancer that is treated with 5-FU chemotherapy.
• About 30% of patients suffer from severe toxicity effects after being treated with 5-FU.
• Despite many approaches having been proposed for DPD deficiency screening, none of the current strategies is adequate to mandate routine DPD testing prior to starting a 5-FU based therapy.
• The current gold standard approach for calibrating 5-FU drug therapy, as also happens with many other chemotherapy agents, is based on body surface area estimation: the patients are administered a dose in milligrams based on the actual square meters of their body surface, and this has been established without a rigorous scientific support. Literature studies show in fact up to 100 fold variation in the levels of 5-FU in plasma between different subjects with the same body surface area , thus resulting in the ineffectiveness of the therapy and in harmful side effects for the patients. Besides, 5-FU has a narrow therapeutic window and to establish the right dose it is necessary to set a trade-off between toxicity and therapeutic effectiveness.
• Surprisingly enough, studies have found out that many patients are not being treated with the appropriate doses for achieving optimal 5-FU plasma concentration.
o Only 20 – 30% of patients are treated in the optimal dose range;
o Approximately 40-60% are underdosed;
o Approximately 10-20% are overdosed ;
• Over the last 20 years, the scientific community has agreed on the reduced toxicity and improved clinical outcomes of pharmacokinetic dose management. Pharmacokinetic in fact takes into account how the human body processes the drug, and by knowing the specific response of each subject to the administered dose it is possible to personalise the therapy around each patient’s specific profile, according to well-defined ranges .
All these facts boil down to a simple evidence:
If DPD deficiency is so dangerous, why isn’t there a mandatory approach to detect it? And, if a dynamic management of the therapy is more effective and less harmful, why is it so difficult to proceed this way?
Because currently there is not a standard, simple, fast, reliable, and cheap way to detect the DPD activity level, which also allows clinicians to calibrate the therapy following well-established clinical algorithms.
This revolutionary approach could be employed in every medical center dealing with cancer, and could greatly enhance the effectiveness of the therapy, with enormous savings for the Healthcare Systems and, most importantly, safeguarding the priceless value at stake: to improve the patient’s quality of life.
Despite the fact that many methods have been tried for DPD deficiency screening, none of the current strategies are adequate to mandate routine DPD testing prior to starting a 5-FU based therapy.
The CARESS Project is a joint effort of 3 research Institutions from Italy and France together with 4 SMEs (small and medium sized entities) from Austria, France and Italy, to develop an accurate, easy-to-use and cost efficient technology for determination of the DPD activity level.
The CARESS project proposes to implement the first, compact, cheap, accurate and standard system for measuring the DPD activity level, thus allowing for rapid screening for DPD deficiency and for real-time adjustment of the therapy administered to the patient through a pharmacokinetic approach, therefore maximizing the therapeutic efficacy, while at the same time reducing the costs associated with the management of toxicities by limiting the dangerous side effects for the patients. The system requires only a small blood sample and will rely on the chemical interaction between a 5-FU analogous chemical probe with the sample, which will release ions in the solution that can be easily detected and measured through optical analysis.
Project Results:
CARESS requirements and specification
The first step of the project was the identification of the stakeholder groups. When a new process is intended to partially or totally replace existing procedures and common practices, it is important to understand how the interested stakeholders are likely to react, and to define their needs and requirements, in order to put a strategy in place is the most likely to succeed. The stakeholder groups are:
1. Final beneficiaries of the system/method (indirect use)
2. Professionals
3. Public/Private Cancer care centers
4. Public/Private hospitals
5. Public/Private Labs
6. Private Clinics
7. Supporters/Institutions
Then, the analysis focused on the end-users of the CARESS system, their current unmet needs and their points of view. For each end-user identified by the analysis, a general description of the profile has been made, while their points of view are depicted taking into consideration the differences that could be present in the same profile working in a different reality. A further analysis of the CARESS end-users has brought the consortium to define three main categories of interest:
▪ the medical biologist or chemist,
▪ the laboratory technician,
▪ the oncologist (or physician with a clinical oncology practice).
The use cases, then, were really important in order to identify several possible interactions between the end-user and the CARESS system and served as a guide for leading to further development of functional requirements.
In conclusion, the requirements supplied by the beneficiaries (SME and project partners) have been collected and organized.
After the collection of user requirements, the system specifications have been defined. The main goal was the assessment of specifications able to address a defined set of user requirements.
In order to perform the best analysis of every technical and functional aspect of the system CARESS and to keep track of any changes, it was decided to organize the collection of the system requirements into sections:
▪ General system requirements (GR)
▪ Protocol system requirements (PR)
▪ Electronic system requirements (ER)
Finally, a preliminary configuration of the system layout and architecture was drafted, according to the requirements collected in the first period of the project, providing an overview of all the components chosen for the prototype to be developed in the project.
Protocol development
The setup of the protocol for the spectrophotometric evaluation of the DPD activity has involved four main activities and has given rise to many others side activities, crucial for the optimization of the developing system and for the solution of the encountered critical points.
The main activities managed for the development of the protocol are:
▪ Definition of the experimental protocol for the reaction of the selected Molecular Probe (MP) with DPD;
▪ Selection of a suitable MP able to release, after reaction with DPD, easily detectable ions;
▪ Assessment of the protocol for total inhibition of Thymidine Phosphorylase (TP) reaction;
▪ Definition of the protocol for the spectrophotometric determination of ions released by MP.
From the experimental point of view, the first three activities are strictly interconnected, and, in particular, both the selection of the MP and the assessment of the protocol for TP inhibition are subordinated to the definition of the protocol for the reaction of MP with DPD.
Definition of the experimental protocol for the reaction of the selected MP with DPD
According to the Dow, the DPD activity has to be evaluated on a blood sample, i.e. on the Peripheral Blood Mononuclear Cells (PBMCs) contained in the blood. The spectrophotometric determination of the CARESS system requires a transparency of the samples in the 400-800nm range of vis-spectrum. For this reason the first performed activity is the evaluation of the best protocol for PBMCs separation from the whole blood sample. For our purpose, the best protocol might produce a high yield of PBMCs and, as stated before, a transparency in the 400-800nm range of vis-spectrum. Three protocols were evaluated, two of which are very similar and isolate the PBMCs by a density gradient centrifugation called Ficoll gradient. The third method is the method currently used by one of the RTDPs (ICO) for the separation of small fraction of PBMCs from blood. This method is very simple and fast compared to the Ficoll gradient method and was evaluated with the aim of simplifying the CARESS methodology. We concluded that the Ficoll gradient obtains the highest yield of PBMCs (1.600.000/mL of whole blood) and transparency in the 400-880 spectral window.
CNR (IMC) exploited samples of blood from voluntary healthy blood donors, courtesy of the Unità Operativa Complessa di Immunoematologia e Medicina Trasfusionale – Azienda Policlinico Umberto I, in order to extract PBMCs and set up the experimental protocol. In order to simplify the processing of the samples and, most of all, the analysis of the obtained data, since the beginning of the second period, the blood samples allocated for each experiment were mixed together to obtain a pool. This pool can be reasonably considered representative of the mean population. Therefore, the degradation rates of the pool can be considered as the mean degradation rates values of the population.
Two different approaches for the reaction of MP with DPD in PBMCs were explored; the first exploits intact PBMCs, the second cytosolic extracts of PBMCs. In fact, we cannot be sure that the selected MP will enter PBMC and, if so, in a reasonable time consistent with the experiments. In both approaches the experiments were performed modulating the number of PBMCs, the time of the reaction and the concentration of the MP. In order to verify the effective degradation of MP (that is the reaction of DPD with MP) a dedicated HPLC method was developed.
The obtained results pointed out that one of the selected MP is able to produce ions upon the reaction with DPD, both in the presence of intact PBMCs and of cytosolic extracts, being the amount of ions produced higher with cytosolic extracts. Once clarified this crucial point, the consortium as a whole decided to carry on the experiments aimed at evaluating the degradation of MP only with intact PBMCs, because of the known advantages of this approach.
Selection of a suitable MP able to release, after reaction with DPD, easily detectable ions
The MP was selected in a pool of molecules (commercial or not) structurally correlated to the chemotherapeutic agent 5-FU and, therefore, capable of interacting with the enzyme DPD. The common feature of these potential MP molecules is their supposed ability to release, upon interaction with DPD, ions that can be easily detected by optical analysis. The starting selection of the MP was guided by its cost and by the chemical nature of the ions potentially released upon interaction; in fact the nature and the amount of the released ions have a strong influence on the spectrophotometric determination. During the project two different MP were tested in the experiments, one able to react with DPD but also with the other enzyme involved in pyrimidine metabolism, while the other capable of escaping the TP pathway. This second MP was evaluated in hope of making the overall CARESS protocol easier, but, unfortunately, this strategy was not successful.
Assessment of the protocol for total inhibition of TP reaction
Both the protocols for the total inhibition of the TP enzyme in cytosolic extracts and in intact PBMCs were set up, exploiting a molecule described in the literature as inhibitor of this enzyme in vitro. To set up the protocols, in the experiments the inhibitor concentration, the incubation time with the inhibitor before the addition of the MP and the number of PBMCs were varied. In order to verify the effective degradation of MP in the presence of the TP inhibitor (that is the reaction of DPD with MP) the previously developed HPLC method was exploited. A summary of the results is reported in Table 1.
Table 1. Results of the experiments with intact PBMCs in the presence of TP inhibitor at 1h and 2h. Δ[MP] is the variation of the molecular probe concentration.
The TP inhibitor at 0.75mM concentration, at both incubation times, and 0.5mM concentration, at 2h, appeared to alter the permeability of the cell membrane. In fact, in the presence of TP inhibitor 0.5 and 0.75mM at 2h and 0.75mM at 1h we often found a Δ[MP] higher than in the absence of the TP inhibitor. Therefore, the less time consuming inhibition of TP in intact PBMCs can be performed with the TP inhibitor 0.5mM with an incubation time of 1h. The difference in the degradation of MP in the absence and in the presence of the inhibitor is very small; for this reason, in order to obtain reliable data on real patients the measure on each blood sample has to be performed in triplicate.
Because the integration of a thermomixer in the CARESS automated system could be a problem both in terms of design and costs, experiments were performed in order to verify the influence of mixing on the reaction of both TP inhibitor and MP with intact PBMCs. To this end, we have evaluated, on samples from the same pool, the degradation rate of MP in a thermomixer at 37°C or in a thermostating apparatus at 37°C without mixing. The results are reported in Table 2.
Table 2. Results of the experiments with intact PBMCs in the absence and in the presence of thermomixing. Δ[MP] is the variation of the molecular probe concentration.
Surprisingly, in the absence of mixing the amount of MP degraded both in the absence and in the presence of TP inhibitor is higher, and the inhibition is more efficient. This evidence is of course advantageous for our purposes, both for the automation of the CARESS system and for the determination of ions produced by the degradation of MP.
Definition of the protocol for the spectrophotometric determination of ions released by MP
The selection of the spectrophotometric method for the determination of ions produced by the reaction between the molecular probe and the DPD enzyme started at the beginning of the project, together with the set-up of the experiments for the selection of the molecular probe. At the beginning the amount of ions to be detected was unknown, therefore we explored two different chromogenic reactions operating in two different dynamic ranges of the ion concentration, and therefore we set up the UV detection protocol according to method_1 and method_2. In the conditions set up for the reaction of MP with DPD in the presence of intact PBMCs, the amount of ions to be detected is of the order of 0.210 μg mL-1. This amount of ions can be detected by the chromogenic method_2. Therefore, the calibration curve for the ion of interest in the proper dynamic range was constructed and the influence of PBMCs matrix on the chromogenic reaction was evaluated.
The results of the UV determination of ions released after the reaction of MP with DPD by the method_2 are reported in Table 3. The results of the UV determination were compared with those obtained by HPLC analysis of residual MP on the same samples in order to validate the UV protocol.
Table 3.Results of UV determination of released ions by the chromogenic method_2 in the reaction of MP with PBMCs. CTRL1 and CTRL2 are samples quenched a t=0 for the evaluation of the initial concentration of MP; C1-C4 are the samples incubated with MP in the presence and in the absence of TP inhibitor.
Results reported in Table 3 indicate that the concentration of released ions evaluated by UV measurements is not in agreement with those by HPLC, i.e. the standard method used as a reference.
The discrepancy between HPLC and UV results could be ascribed to two different causes, the first one is that the selected MP is not an inert molecule with respect to the chromogenic reaction exploited in method_2, the other is that ions released by the reaction of DPD with MP escapes from PBMCs only partially. These hypotheses were verified and confirmed by a lot of ad hoc experiments. That being so, (partial escape of ions from lymphocytes and reactivity of MP in the selected chromogenic reaction) a solution was found for evaluating, still by UV measurements, the degradation of MP. Actually, instead of the ions released by the reaction of DPD with MP we decided to evaluate the residual amount of MP after the reaction. This evaluation can still be done (with some expedients) by the same chromogenic reaction and therefore, doesn’t involve a change in the UV protocol and in the detection system of the CARESS machinery.
Therefore, a new UV detection protocol (always based on the chromogenic reaction 2) was set up and the degradation of MP upon the reaction in PBMCs was evaluated. Results of degradation of MP evaluated by UV measurements and by HPLC are reported in Table 4. The results are the mean values of 4 experiments. In each experiments each sample (the control, the samples in the absence and in the presence of TP inhibitor) were performed in duplicate. The degradation of MP is reported as variation of the concentration of MP with respect to the controls.
Table 4. Degradation of MP evaluated by UV measurements and HPLC.
The degradation found by the UV protocol and the HPLC method used as a reference, are in agreement. We did not find significant differences in the UV/HPLC correlation between the samples in the presence and in the absence of TP inhibitor. In Figure 1 the correlation between the results of UV and HPLC analysis is reported. At a first glance the correlation between the two sets of data could seem not so good; actually, in the literature, the comparison between different analytical methods displays similar correlation coefficients.
Figure 1. Correlation between the variation of the concentration of MP with respect to the controls, evaluated by UV and HPLC analysis.
In order to confirm by another experimental technique the goodness of MP as molecular probe for the determination of DPD activity, we have performed experiments exploiting a method currently used for the evaluation of DPD deficiency in the laboratory of Advanced Molecular Diagnostics (DiMA) at “Azienda Ospedaliera S. Andrea - Facoltà di Medicina e Psicologia, Sapienza University”. This routine evaluation of DPD deficiency relies on the determination of 5-FU degradation in PBMCs, analyzed by a HPLC-Tandem Mass Spectrometry method. These experiments were performed thanks to the courtesy of Prof. Maurizio Simmaco and Dr. Luana Lionetto. We can confirm also by this technique that the selected MP is a suitable probe for the CARESS system because it is degraded by PBMCs and that the degradation is different in the presence and in the absence of the TP inhibitor.
Protocol evaluation and validation
The purpose of this activity was to validate the feasibility of the full CARESS protocol created by CNR with the optical system developed by LABOR and his possible use as a diagnostic routine method. ICO has validated the full CARESS protocol with HPLC method and then they have proceeded with the validation with UV measurement.
The protocol developed by CNR contemplates several steps. ICO followed the steps reported below:
A. PBMCs isolation from blood sample with a manual or integrated Ficoll method
We looked to optimise the full CARESS protocol to increase the feasibility of the protocol and to simplify it for the diagnostic routine. For this purpose, we also checked the possibility to isolate PBMCs into blood draw tubes with Ficoll gradient integrated and their conservation after isolation.
In the way to validate this method for the PBMCs isolation, we made the full CARESS protocol with PBMCs isolated manually or with the CPT Ficoll tubes®. The blood samples came from the same volunteer.
B. TP inhibition, molecular probe degradation by DPD
After separating the PBMCs, we proceeded with the preparation of 4 aliquots: two with the addition of the TP inhibitor and the other two without the addition of the TP inhibitor. For each pair of aliquots one of the two serves as a control (reaction stopped at time zero), while the other is the sample treated according to the protocol in order to evaluate the DPD activity. Each pair of sample and their controls has been evaluated with the HPLC and the UV measurement.
C. Chromogenic reaction for UV measurement allowing the indirect DPD activity evaluation
We performed an automatic UV measurement with a microplate reader. Moreover, we decided to perform 5 replicates for each sample, considering as final Absorbance value the mean of these 5 replicates. The HPLC and the UV results are quite in line each other, but they must be confirmed with a large series of samples. However, we can conclude that the results obtained at ICO and at CNR are comparable and the protocol is well defined.
D. Full CARESS manual protocol
After separating the PBMCs with the manual Ficoll method, we proceeded with the preparation of 4 aliquots: two with the addition of the TP inhibitor and the other two without the addition of the TP inhibitor. For each pair of aliquots one of the two serves as a control (reaction stopped at time zero), while the other is the sample treated according to the protocol in order to evaluate the DPD activity. To conclude, the full manual protocol developed by CNR was verified by ICO successfully.
E. CARESS protocol optimizations
In order to evaluate the conservation of the samples after dilution with sulfuric acid 0,2M obtained in the previous experiment, they were aliquoted and stored. In conclusion, we need to make directly the chromogenic reaction after the probe reaction of to store it not diluted at -20°C.
In order to evaluate the stability of PBMCs and the DPD activity after samples conservation, we explored different possibilities routinely used for maintaining PBMCs in cell cultures. PBMCs recovered from the same blood sample were divided in several aliquots to test the different conservation methods. One aliquot was treated following the CARESS protocol and was used as the “control” for the comparison with the conserved PBMCs. None of the methods tested thus far have achieved positive results. After 24h conservation we have observed a reduction of the number of living cells, compared to the number obtained for the control sample. Moreover, we observed a different DPD activity in the controls and in the conservation tests. Unfortunately, we have not be able to reduce or simplify the protocol in terms of PBMCs isolation or conservation.
Then, further activities were performed in order to evaluate the possible correlation between the CARESS evaluation of the DPD activity and:
▪ the DPD deficiency evaluates by a multiparametric analysis performed at the ICO before treatment with fluoropyrimidines (5-FUODPMTox TM);
▪ the 5-FU plasma concentration at the first cycle of 5-FU treatment (5-FUODPM Protocol TM).
The scope of those evaluations was to permit the development of new algorithms, by ODPM, for dosage adaptation of 5-FU throughout the entire treatment course. The frequencies of those mutations are very low: which is why validating a new protocol and comparing it with what is currently in use would require a very large prospective clinical study. At the ICO, the two most frequent variants are: IVS14+1 G>A and 2846A>T (D949V). Both of these have been founded in patient samples with an allelic frequency of 1-1,3% among more than 20,000 patients.
The frequency of the complete DPD deficiency is around 0,2%. In consequence, in order to reach a significant relevance of the results and to be able to evaluate the correlations defined previously, the study would need to include at least 3000 patients, each of whom would be evaluated using both methods for comparison.
For this explorative project, we have decided to use only the biological waste of the ICO patient samples, when there was enough blood left in the sample after routine testing had already been accomplished. Therefore, it was not possible to make a statistical correlation. A larger prospective clinical trial will be required to determine if the CARESS method is an effective measure of DPD activity (which could be submitted as a Horizon 2020 European project).
At this point, the CARESS method does not evaluate the plasmatic level of 5-FU, but a substitute of that molecule to determine the DPD enzyme activity. Thus, adjusting the CARESS protocol to quantify 5-FU during the treatment, and also use the CARESS prototype for this additional analysis is quite far beyond the scope of the Caress project. This aspect would need much further R&D and is not only outside the scope of this project, but is also outside of the scope of the follow-up clinical trial project for which the consortium will be seeking financing over the next few months. At this time, the best way to study the metabolism of 5-FU during treatment is to measure the 5-FU concentration in plasma by a chromatographic method, which is already in current practice at ICO.
In conclusion, the small sample size did not allow us to make desired correlations to clinically validate the protocol for determining the risk of toxicity to fluoropyrimidines before treatment. The results, however, have clearly suggested that the manual protocol is quite robust. When a prospective clinical study aimed at clinical validation of CARESS method is finalized, this manual protocol may be used without further modification.
Electronic development
During the P2, the project entered in the core of the activities with the consolidation of the prototype concept, the mechanical and electronic design (carried on partially during the P1), the selection of all the components, the manufacturing of all the components and the preliminary tests on the single components before the integration of the whole prototype.
These activities represented the critical path of the project, under the engineering of the prototype point of view, and a deep management of the timing has been fundamental for tracking the single steps and be sure that no major delays could affected the entire plan of activities.
After months of designing functional components (P1) and the integration of the device, the first hypothesis of the control system has become an expanded system. The final machinery can be considered a real automation device having to perform rapid and precise operations and displacement of fluids. The control unit manages a high number of “peripheral” units and heterogeneous movements and automations, i.e. micrometer movements of a three axes portal or the presence control of the involved consumables, passing from the monitoring of an incubation chamber.
The Caress instrument, from the mechanical design point of view, can be considered as a machine for dosing small amount of liquids (sample, reagents, cleaning fluids), in and out of defined volumes where other operation are taking place (centrifuge, mixing, optical measurement). To perform these tasks, an ad-hoc designed dispensing group is mounted on top of a Cartesian plotter, with X and Y-axis corresponding to the working area where liquids have to be transferred. A Z-axis movement is necessary to put the needle in contact with reagents and samples. The intersection of rows identifies the points of the working map.
The machinery consists of several sections controlled by a central control unit that it is connected to the front Panel-PC (the operator interface). The control unit has expansion boards that adapt the signals and are used as mechanical support where necessary.
Summing up, the machinery is composed by the following functional blocks which have been designed and manufactured independently before the integration into the prototype:
▪ Power Chain (from 220V to +48V, +24V, +12V and +5V): the distribution of the power is ensured by starting from the 220V power supplies (from a typical wall outlet) through stabilized power units which are connected directly to the related sections. The external components (like pumps, air blowers, etc.) are supplied directly from 12V entering inside the control board and are controlled by the relative relays. This section supplies all sensors, motors and the HMI section too. All selected power supply elements are integrated inside the machine. They have the appropriate certifications to be used in laboratory as prototype and they are all commercial components in compliance with all applicable regulations. Four power supply units provide: +48 V DC for centrifuge motor, +24 V DC for syringe, +12 V DC for three axes motor, for panel pc and external components, +5 V DC for uP control unit and extension boards.
▪ Control Unit (Main Unit and controllers): the Control Unit manages and controls the communication and the co-ordination between input/output and devices/sections. It reads and interprets instructions and determines the sequence for processing the data. It directs the operation of the other units by providing timing and control signals. It performs the tasks requested by Panel-PC. The control unit takes care of the handling and control of: gantry movements, reagent and sample dispenser, security, spin monitoring, thermostat monitoring, controls the filling level of tanks, cleaning monitoring system and HMI communication. To better adapt control to the machine, the central unit has been designed and built from scratch using a microprocessor custom architecture with a central board connected to other secondary boards.
The main board (also called Control Unit) is the core of the whole system. It is connected with all components. On the board there are two communication ports: one is a direct serial port used for status report of the machinery, the second one is composed by an isolator and a USB to serial UART interface to be connected directly to the panel pc. Two LEDs (D2 and D3) monitor the transmissions. This port is also used to reprogram and update the uP via HMI.
▪ HMI Section (with barcode scanner): the CARESS HMI is composed of a Panel PC with a resistive touchscreen (to be used with gloves). The construction has been chosen to be as robust as possible even in difficult environment. The display has a visible area of 10.1” (1024x600 resolution). The display is arranged to be connected to a barcode reader via USB connection. The display will communicate to the operator all the necessary information to prepare properly the test cycle. It will inform the operator of the correct procedures to check the consumables loading (deep-well plates, reagents and distilled water), and allows operating the software. The barcode scanner will read an identification code for both the samples and the reagents that are loaded into the instrument. After each scan, the software will indicate where to place the respective reagent or sample, and will verify its correct positioning using optical sensors and/or micro-switches.
▪ Centrifuge Section: the centrifuge can be considered as an independent section controlled in its functions by the central unit. A chemical centrifuge works on the principal of centrifugal force. The high-speed revolution of the basket separates the supernatant and the precipitate (and cells dishes). This section is thermostatically controlled to allow an incubation period without further moving the samples. This section is composed of several electronic and mechanical parts. The inner part of the centrifuge is designed to be warmed and heated to a controlled temperature of 37°C (as required in the measurement protocol). Under the exposed bucket is placed a box containing photo-detectors with a reduced beam. Six optical sensors are placed under the respective sites of the six polypropylene tubes to control the positioning.
▪ Gantry control and Syringe: the gantry lifts objects and can move horizontally on a rail fitted under a beam. It is driven by three autonomous stepper motors and has several safety limit switches for each axis. This section requires electronic controls to drive the movements. The Y- and Z-axes are controlled by two twins and autonomous controllers through two communication ports. The X-axis is composed of an alpha step closed loop stepper motor. The dispenser is composed of a syringe with a washable needle. It is moved by the gantry on the surface of the working area. This dispenser is able to pick up and release liquid from the reagents bottles into sample tubes and micro-plate. The syringe piston is actuated via a lead-screw with 30 mm of travel, driven by a stepper motor.
▪ Optical Unit: the scanner has eight emitters and eight detectors to speed up the readings. It reads a loaded 96-wells microplate (128 x 86 mm) prepared by machinery. This section is composed of three boards connected to the control unit. The Optical Unit mainly consists of light sources transmitting light through the samples and photodiodes detecting the absorbance of the samples. The unit is integrated by the software, which collects, records and analyses the generated signals. The set of two circuits measures the ratio of light not absorbed in the wells (and, then, in the reference ones) at the selected wavelengths (405 nm). The LEDs board is composed by eight UV lamps, which illuminate the microplate column where the scanner is positioned. In each branch, the flowing current is constant and stabilized by precision resistors and a reference voltage. The used LED has a very narrow bandwidth. The photodiodes on the opposite side of the light source generate a digital output proportional to the amount of light received (and therefore proportional to the absorbed light).
▪ Auxiliary Section (reagents and tanks checks, and washing control): the machine is completed with further components and a control board for the reactants. This section is composed of three pumps for washing unit, an air blower for needle drying and three level sensors for the tanks.
According to the testing phase of the project, all components were tested before and after the integration of the component in the related section and then all functionality tests were performed for each sections. All tests ran smoothly with the exception for the functionality of the washing chamber that has been revised and reworked. After the modifications provided, this section completely responds to the required specifications.
System Integration and preliminary tests at lab scale
As reported in the previous section of this document, the CARESS prototype required a design phase and a subsequent assembly phase which took a lot of effort during the period from M7 until M19 approximately. The single components have been assembled, tested, and reviewed in all their parts in the period between M19 and M22.
The final specifications of the machine are:
• dimensions: 870 mm (width), 620 mm (depth) and 760 mm (height) without stands;
• weight: about 68 kg.
The working plane hosts the functional sections of the instrument. These are:
1. Centrifuge Section;
2. Optical Unit;
3. Dispenser Section;
4. Reagents Stand;
5. Washing Section.
Below, the list of the auxiliary section by which the machinery is composed:
1. Power Supply Section;
2. Control Unit;
3. HMI Section;
4. Three tanks (for the containment of the distilled water and the ethanol and for discharging the liquid);
5. Three Pumps (two to send the washing liquids and one to force liquid drain);
6. An airblower (to dry the needle).
Connections are placed on the side of the prototype: white reset button for only the Control Unit, the ethernet port, external USB port with cover and the power plug with main switch (red light).
The system integration of the prototype respected the timing as reported into the DoW. The activities started at the end of the year 2014 (November – December 2014, M19/M20) to be concluded at the end of January 2015 (M21).
The Consortium decided not to publish pictures of the final prototype integrated in order not to lose the innovation. In fact, the prototype requires additional work after the end of the project for reengineering and industrialization. The publication of pictures in this delicate phase could be risky for the innovation because the prototype is not protected by patent yet, and the SMEs could lose their foreground and the first mover advantage.
After an intensive integration of all the components into the prototype, the activities went on with the preliminary tests at laboratory scale.
The preliminary tests were carried out in the chemical laboratories of LABOR. The strategy followed for the initial tests is listed below:
• Tests for the control of the necessary movements of the prototype, in compliance with the protocol, without liquids and reagents;
• Tests for the control of the movements loading the machine with distilled water;
• Tests campaign for retracing each step of the protocol with the real reactants.
The test campaigns have been conducted from the end of January to the middle of March 2015 until the shipment of the prototype to ICO.
Together with the CNR, LABOR began the first tests session of the CARESS protocol at the end of January. Taking as reference the protocol, the control tests were performed by running an approach from the bottom upwards: firstly last operations have been performed (optical measurements), which actually represent the final step of the protocol before obtaining the data on the CVS file.
The different tests sessions have been gathered in order to follow the same engineering approach followed for the design and manufacturing the prototype. Therefore, in each tests session, a specific component and operation of the machine have been tested.
During the first days the machine has been loaded only with distilled water. Labor and the CNR reviewed the entire protocol to align the prototype to the manual procedure. To optimize the timing of the microplate preparation the dosage of H2SO4 in the microplate (column 7) has been moved. This action also allows a better cleaning of the needle in the reagent exchange.
During one of the firsts tests, the liquid did not drop completely, it was necessary to update the firmware of the control unit to insert a specific step of pre-loading air in each dosage and for each quantity. The firmware update solved the problem.
After these optimizations, a new test was performed by filling the microplate with a manual preparation. However a problem occurred due to a bug in the software of the prototype. The test would have closed properly, if a 15-minute break would have been respected. This pause was not respected and this permitted to identify a specific bug in the software, which was solved.
The air taken before each dosage was increased to 30 µL and the speed of the plunger of the syringe was reduced to facilitate the exit of the last drop during the release of liquids. To better appreciate the levels of light intensity detected by the scanner the reading precision of the photodiodes was set to three decimals.
During a test, it was put in evidence that the dosage of the Br-/BrO3- for 24 patients took more than the pause required. We decided to make an intermediate test to measure the timing of Br-/BrO3- dosage for 24 patients (filling from the 2nd to 6th column) including washing. The cycle took more time than the expected. From this, we decided to increase the waiting time. The CNR performed tests to ensure that the protocol did not have substantial changes.
Before to analyze the obtained data, we have characterized the scanner. We performed a test doing a scan of the entire microplate filled with water in the dark and with the presence of external light; then we scanned the entire microplate filled with Standard 1 and Standard 5 with light and dark. The data did not show significant differences between the tests carried out in the dark and with external light.
Each time that we performed a specific test, the machine was controlled. Several functionalities were tuned: timing of the waiting and the synchronization of the actions.
Over the following days we performed tests for the characterization of scanner tuning the lighting time and the timing after the microplate preparation. We noted that increasing the illumination time did not have significant improvements, but only an increase of the timing of the scanning step.
In the following days we completed all actions on the machinery. Unfortunately, the results of the calibration curves were non-regular. This was due to the precision of the assay of the syringe. This problem was not revealed during the tests with distilled water. Using real reagents and sample, the needle seems to get “dirty” and reduces the precision in low volumes (10µL) where an incomplete detachment of the last drop dosed has a greater influence. To improve the precision of dosage, small volumes were doubled entailing several changes in the preparation of microplate.
The whole firmware has been adapted to the new protocol. Through many internal tests and a session tests with the CNR, we have double-checked all steps in the CARESS cycle.
Using larger volumes we obtained a greater stability but we did not solve the problem: in some experiments, the problem of the drop was observed again, unexpectedly. One of the optimization and improvement of the prototype will be certainly focused on this aspect. This can be solved with a study for a suitable coating to perfectly control the fall of the liquid from the needle.
Internal tests had been crucial not only to improve the functional aspects of each steps, but also to highlight two mistakes in the design of the washing station. These were:
• the geometry of the washing chamber;
• the necessity of a suction pump of waste liquids.
These two points have been completely solved by reviewing the machine in its functional design.
The following corrective actions have been carried out:
• mechanics rework of the washing chamber to change the internal shape;
• addition of a drainage pump.
During the tests sessions, we noticed a shift in the measured light not due to repeatability. We noted a natural evaporation due to auto-heating of the machine (power supply, motor drivers, etc.). This weight loss can be greater after a cycle of thermostating the centrifuge. This aspect, in a future revision, should be concretely considered and compensated with the involved liquids (quantity, viscosity, timing of release, etc.).
The linearity of each channel was tested by filling the first six columns (those used for the measurement) with several standards solutions of chromogenic reagent at different concentrations. The seventh column was filled with H2SO4 to have a luminosity reference.
All the necessary changes to improve the protocol have been implemented. Thanks to the modularity of the machinery and to the numerous parameters of the firmware, the debugging time and the revision work have been greatly reduced. The preliminary tests for the electronics and the mechanics have been completely preformed, as well as the operative tests for the automation from the software and firmware point of view.
The test phase highlighted some points which, during the industrialization of the prototype, will have to be managed and improved. All of those have been analyzed and the relative solution has been identified.
In conclusion we can state the followings:
▪ The CARESS prototype is a complex machine which correctly automates the protocol setup by the CNR;
▪ A complete tests campaign has been held between January and March 2015 for the debugging of the machine.
▪ From the engineering point of view we identified some weaknesses. Those have been analyzed and a related solution has been found. These solutions can be easily implemented during the industrialization of the prototype.
▪ The design phase of the mechanics resulted to the successful. In such a complex machine, only two adjustments were needed, both solved during the tests phase. The efficient mechanical rework of the washing chamber and the installation of a draining pump permitted to complete the washing cycle as it was designed and required in the WP1.
▪ The software of the machine revealed some bugs during the tests campaign, as well as the firmware had to be managed in order to optimize the automation and the waiting times. All the bugs identified have been analyzed and solved.
▪ The prototype was sent to ICO ready to be installed. At dedicated session of a week (Feb 9th –Feb 13th) has been held between LABOR and ICO, in the labs in Rome, in order to train ICO technicians for the correct use of the machine during the field tests. Further, an installation manual has been prepared as user guide for the partners.
▪ The Prototype has been designed and manufactured to be completely controlled and analyzed remotely, including a camera, filming the working area. With these features, LABOR was able to support ICO during the whole duration of the field tests.
Pilot projects
The field tests were to present the protocol followed with the CARESS device and the controls test needed for the validation of the prototype. It aimed to carry out and to evaluate both CARESS protocol and prototype, mimicking the real operating conditions within a diagnostic laboratory. The protocol developed by CNR was modified and adapted (in terms of volumes and timing of dispensation) to the prototype. All the experimental tests made at ICO for the validation of the prototype have followed the amended protocol. In order to define the performance of the CARESS machinery in the detection of DPD activity, and its potential deficiency, results obtained both manually and by Caress device were compared.
We are aware that this will be an important, and industrially ground-breaking results, and that a complete validation of the Dose Management Approach would require time, resources and budget, which are beyond the scope of this project.
The machinery was delivered on March 23rd 2015 at ICO (Angers). The first tests were made to evaluate the basal functions with LABOR assistance.
For the validation of the CARESS device, we firstly used cultured cells. The cell lines used were human colorectal adenocarcinoma. They were cultured respecting the GLP (“Good Laboratory Practice”) procedures and the recommendation of the provider ATCC (cultured in RPMI – 10% SVF solution). The choice of these cell lines was made based on the enzymatic DPD characteristics reported in the literature. Moreover, we have verified the genetic profile of these cell lines (sequencing) in order to evaluate the presence of the main important DPYD polymorphisms generating DPD deficiency: no major polymorphisms, responsible for a reduction in DPD activity, were found. Using these cells could be useful to verify if the CARESS system and/or the CARESS protocol are able to discriminate samples with different DPD activity.
Tests have been performed in order to check:
▪ Internet or intranet connection
▪ Software
▪ Dispensing
▪ Centrifuge
▪ Reagents evaporation (microplate)
▪ Biological constrains
▪ Calibration curves
▪ Repeatability/ reproducibility
The machinery was evaluated in comparison with the same protocol manually made. It took into account the waiting times, volumes and modified solutions (compared to the CARESS manual protocol). It aims at carrying out and evaluating the CARESS solution in real operating conditions, measuring through comparative analysis the performance of the CARESS device in the detection of DPD activity, and potentially his deficiency. The manual and automated protocols were developed to be easy to use and correspond to the diagnostic lab constrains. In way to define and validate their use in lab, we made some evaluation at different steps and levels.
A. CARESS device evaluation
Firstly, we evaluated the CARESS prototype device regarding several parts of both the device itself and the adapted protocol.
During the first tests with the CARESS device, we made some observations about the machinery protocol. All tests were defined in agreement with Labor and CNR and supervised by LABOR remotely using the internet interface with the instrument.
The syringe showed a problem during dispensing: we observed a random and not reproducible drop remaining on the needle. This deviation from the distributed volumes induces errors that were verified by reproducibility tests and with measurement of the dispensed volume in each well of the plate. After discussion with Labor and CNR, some possibilities exist to avoid this problem. This issue would need to be rectified before further testing of the machine protocol can be done.
Some observations were made about the prototype components. All this observations aims to optimize the CARESS device.
The CARESS software developed by LABOR showed goods results, followed the CARESS device protocol defined by CNR and presented the results in a table form as comma-separated values (.csv) file convertible into an Excel document. A continue internet or Ethernet connection is needed to avoid the arrest of the protocol.
The CARESS device protocol was evaluated regarding calibrations curves, and with both cell lines and water for the reproducibility.
The results obtained for intra- and inter-assay for the cell lines tested (HCT116 and DLD1) presented a deviation standard of the molecular probe evaluated up to 10%. It is too high to validate the device protocol both manually made and with the prototype. The results with water looked better, but they had a high standard deviation too. For the following tests with the patient samples, we decided to use the manual CARESS protocol already validated.
B. Final CARESS protocol results
In a second part, the evaluation of DPD activity was made, starting from blood samples (volunteers) or biological wastes (patients). The CARESS manual protocol was used in way to be sure about the results obtained.
From March 2015 to June 2015, 49 samples were analyzed following the manual protocol. The 49 samples analyzed derived from biological waste from patients treated at ICO. The R2 obtained for the calibration curves were > 0,99. The results obtained shown that there was a degradation of the molecular probe in all samples and the degradation rate was variable among samples, as expected.
An inter-assay was made for two samples and the standard deviations obtained were close to 10%. Other tests must be done.
Intra-assays were made for 5 samples, with 3 to 5 replicates into the same UV measurement plate. These results are encouraging because the deviation standard obtained for all these tests was under 10%, even if a standard deviation of less than 5% would be preferred. These results, compared to those obtained by the CARESS device, indicate that the manual protocol is robust and repeatable.
All the calibrations curves had a R2 value equal or up to 0.99. These optimal coefficient values of calibrations curves allow us to calculate the molecular probe concentration in the unknown samples accurately and reliably.
The reported results are very encouraging as they demonstrate that even slight changes in DPD activity are carefully highlighted by the CARESS manual protocol. This protocol is therefore sensible in the stratification of enzyme activity.
In conclusion, the CARESS protocol is promising for the determination of the lymphocytic activity of DPD. It is possible that the CARESS protocol may also be able to detect high metabolizer, but that occurred in only one sample, so no conclusions can be drawn from this event.
To validate the use of the manual CARESS protocol to screening the DPD activity prior chemotherapy, it is necessary to establish a clinical validation trial, including the analysis of at least 3000 patients samples to make sure to include deficient samples in the study. This trial should be multicentric and multinational in order to have the number of samples required as quickly as possible.
The weakness of this protocol, at present, is related to the PBMCs isolation that requires some technical skills and equipment. Moreover, PBMCs samples are not stable and maintaining DPD activity for long time (more than a few hours) is not possible (further research into this field would be very useful for making this protocol a clinical reality): the cells have to be isolated within two hours of the blood sample collection and the experiment must be done immediately after the isolation of the PBMCs.
Potential Impact:
The analysis of the market (held during specific workshops with the SMEs) showed that there is an urgent need of the determination of DPD deficiency prior the chemotherapy treatments but, currently, the methods used to determine the dose of 5-FU to give to a patient prior to chemotherapy treatment:
▪ do not take into account deficits or over-expression DPD. The result is that patients could be under- or over-dosed;
▪ are not specific and, when the specificity is high enough, the technique is far from adequate to be a standard in clinical practice due to the high costs associated.
In order to start identifying the potential impact the exploitation scenario of the innovation, the first step has been the definition of the framework. In particular, the DPD testing market and the In Vitro Diagnostics (IVD) sector have been described in their trend taking in consideration the literature defining the state of art. As main conclusion, “factors like the increase in awareness and acceptance of personalized medicine, advancements in molecular techniques, and increasing investments in genomics & proteomics research will spur the molecular diagnostics market in the future. Unfortunately, reimbursement issues for non-coded tests and complex regulatory framework will restrict the growth of this market. Introducing a new successful product is a challenging and complex endeavour. It can be estimated that the CARESS innovation will be not ready for the market before 2018, due to the inherent engineering, industrialization and certification processes for the device.
In addition, the IVD market has also been studied on the Player level. Although the market for IVDs is fragmented with more than 500 companies, the study delves into the business dynamics of major players in the industry that together form more than 70% of the global IVD market.
Then, the analysis of the external environment resulted in the first release of the CARESS SWOT analysis comparing the internal framework and highlighting the main strengths and weaknesses of the innovation.
The exploitation plan started with the segmentation of the market in order to focus the attention on the stakeholders potentially interested to the innovation. In agreement with the application scenario, the two main segment of interest are:
1. Small structures, with no/little spending power, which cannot afford the purchase of the machine. In this scenario, the reagents kit can be sold with the support of the ODPM’s web calculator for the interpretation of results. This scenario is the most profitable for the ODPM’s outsourcing model of the CARESS testing service.
2. Large structures, with adequate/high spending power, which can afford the purchase of the automated device. This scenario is the most profitable for the direct sale model for all the SMEs.
In both “scenarios” the SMEs will be able to act as first mover in the market, being market leaders in their sectors, to be able to bring the innovation to identified targets quickly and efficiently. This represents one of the key advantages for the exploitation of the solution.
Focussing on the exploitation of the solution, the SMEs deepened their discussions regarding the priority of the project, given the two possible scenarios. All of them identified, as the top priority , the Large Structure scenario (scenario 2) in which to exploit the sale of the fully automated device. However, even given the relatively higher barriers to entry, the Small Structure scenario (scenario 1) will be also investigated with the purpose of entering the market with an alternative lower-cost solution and/or with an indirect sale model.
The SMEs also discussed the geographical segmentation of the market. It was decided to view the market in two geographical segments when talking about “SMALL STRUCTURES”:
▪ SEGMENT A - Highly developed markets – oncology Centres
▪ SEGMENT B - Lower developed markets (Central & Eastern Europe, Asia, Africa, South America)
At the end of CARESS post project phase, the SMEs have identified the following products/services:
▪ CARESS KIT
o KIT ALONE: reagent kit, ready to be validated in a comprehensive clinical trial.
o KIT + ODPM CALCULATOR: reagent kit, ready to be validated in a comprehensive clinical trial, in coordination with the ODPM calculator.
▪ CARESS DEVICE
o DEVICE + KIT: validated prototype to be industrialized; further steps needed at the end of the project; specific supply package for the kit;
o DEVICE + KIT + ODPM CALCULATOR: validated prototype to be industrialized; further steps needed at the end of the project; specific supply package for the kit.
For each business solution identified (reported above), a value chain has been defined. With the outsourcing business model, ODPM will be the entry point in the market with the DPD deficiency screening service. In the direct business model, AMEDA will promote and sell the CARESS kit for manual use with the ODPM calculator as an option worldwide except France. AMEDA will also distribute the automated device, the ODPM calculator as an option worldwide, except France, where ODPM has the exclusive rights for the both the CARESS kit and device. IRBM and OPTO are always present in the value chains as exclusive suppliers of the probe/reagents and the electronics, respectively.
The first step to make the SMEs independent has been the technology transfer from the RTDPs, in order to shift the knowledge:
▪ AMEDA, OPTO and ODPM: electronics (AMEDA + ODPM), optical sections (OPTO), software and hardware of the automated device (AMEDA+ ODPM);
▪ IRBM and AMEDA: molecular probe and protocol;
▪ ODPM: DPD deficiency screening.
Specific meetings have been organized for demonstration and training of the SMEs, as well as materials to be used as user guides.
The most important outcome of the exploitation strategy has been the agreement on the follow up of the project and the identification of the most suitable funding scheme. Further, since the application is foreseen before the end of the 2015, a concrete application schema has been shown and already identified. This is in a preliminary stage and the consolidation is needed by the end of sept 2015, when a dedicated meeting has been scheduled.
Here are the key points on which the SMEs have set up the exploitation plans, which summarize the state of art at July 2015:
1. The CARESS prototype has been tested at laboratory scale and in a relevant environment but it is still in an experimental phase. The prototype must be further optimized in order to be exploited on the wider market. This will have to be done not only from the mechanical point of view but also from the optical and electronic point of view. Some weaknesses, modifications and optimizations have already been identified by the RTDPs. The final reengineering of the prototype will need at least 24 months and will be covered in the post-project phase of the plan.
2. The protocol for the DPD activity screening has been created, analysed and optimized by the CNR and tested in their laboratories. The results are good. The protocol has been tested in parallel in a relevant environment (ICO) and it has been validated. However, in order to launch a new clinical practice in the market, a multi-center clinical trial is needed. This will be the second element of the post project phase.
3. Given the specificities of the healthcare sector as well as IVDs and pharmaceuticals market, the beneficiaries agree that it is not possible to go to the market with an innovation without any means of protection. Therefore, as soon as the next phase is well underway and the preliminary results obtained in CARESS are confirmed and improved, a patent application will be necessary. It is agreed to apply for a patent for both the device and the protocol together.
EXPLOITATION PLAN STRATEGIC ROADMAP
With the validation procedure scheduled within the project’s framework, at M26, the protocol is ready for the clinical validation, which is out of the scope of the project, but that could start right after September 2015.
ODPM underlined the importance to describe the context of this clinical validation trial. This clinical validation trial should be organized as a 3 years study with a 2-years inclusion period, a multi–country, multicenter with a minimum inclusion of 500 patients per country. In this case, the trial could cost between 3 and 6 million euro and investors will be necessary.
However, it is a common thought that the protocol needs to be clinically validated before the instrument prototype is reworked and put into production. At the beginning of the discussion, the SMEs agreed to work on the industrialization of the prototype instrument in parallel to the clinical validation of the protocol, with the purpose to speed up the process. However, later on in the discussions, this was seen as a mistake and a waste of money. In fact, if the protocol does not prove to be an efficient analysis of DPD activity in clinical practice, then more work needs to be done on the protocol, which would then likely engender more changes to the instrument.
It may make more sense to do a smaller clinical trial (one center study at ICO) for a period of time to be determined (9-12 months), so that there are preliminary clinical results before the next phase with the final changes to the prototype and industrialization phase begins.
The preliminary study could be a feasibility study done over about one year, which could then give solid results regarding the clinical validity of the protocol (not just that it works chemically speaking) and then the consortium could move forward with the H2020 proposal much more confidently. The picture below reports the organization of the phases.
As for the industrialization of the prototype, the possibility to launch the system in the market requires an extensive testing phase after the end of the project. This is needed both to determine the level of differentiation of the CARESS instrument in relation to the systems already available but also to fulfill the requirements of the IVDs directive.
The post-project plan for the device will be divided into two phases: the first phase will include the reengineering of the prototype with the goal of reaching full compliance with the EU IVDs directive. Given the current state of the prototype, these goals are not too ambitious and are not too difficult to achieve once the re-engineering phase and new clinical testing are completed. In the first part of this phase, a re-engineering process will be required to prepare the industrialized production of the CARESS device starting from the validated prototype, ready at the end of the 2016. The main scope of this process is to test the performance of the system and make it completely compliant with the Directive. In fact, while an acceptable level has been reached with the prototype, a deeper investigation and optimization will be required to pass from the prototype to a commercial product.
Then, a deep testing phase is needed both at laboratory scale and in the field with end users. The activities concerned the reengineering and the tests will be strictly connected.
Likely, not only the compliance with the directive will be necessary but also the adaptation to new customer requests. This brings to the second phase. During the field test the main activity will be the collection of feedback and new requirements from end users. The new contacts with the customers will be also exploited to set up the marketing plan. This phase will require further adaptation of the device. It will be crucial to follow these activities in parallel in order to decrease the time to market as much as possible.
Finally, the commercialization can start only after this period of design revision for market use has been completed. The final product, in fact, even if in line with the technical performance, will need a revision under the marketing point of view. The market requires specific conditions that should be taken into consideration before the official launch.
The final product will be ready for market in 30-36 months after the end of the CARESS project. This means that the product cannot be commercialized before the end of the 2018.
FINANCING THE EXPLOITATION PLAN
During the Review meeting, held in Brussels the 21st of March 2014, the REA’s Project Officer suggested to start thinking about the follow-up of the project and the possibility to find new funds in order to finance the next stages. The PO suggested the “Business Angels” or “Angel Investors” as a possible way to go. During the 1st Year meeting (June 2014) held in Paris, two other possibilities were explored: The Worldwide Innovation Challenge 2030 originating in France, and the SMEs instrument of the EU program Horizon 2020. At the end of the project, the consortium agreed on the following:
▪ Innovation 2030 is not the suitable funding scheme for the follow up of the CARESS project.
• The suitable Call for CARESS seems to be the Call of the biomarkers, where the proposal can be related to the clinical validation and In-vitro diagnostics validation.
• CARESS could be applied directly in a Phase 2 trial before the end of the Year 2015. However, a specific channel for this point needs to be activated right after the end of the project in order to understand if November 2015 is a realistic date for the completion of the proposal. We also must determine if phase 1 is a necessary step or if the project is ready to go directly into phase 2 of the program and how the pre-clinical trial can be managed. In this case, robust and realistic data will be necessary for a detailed business plan.
At the time of wiring, the SMEs have already a proposal for the proponent consortium on H2020.
INTELLECTUAL PROPERTY RIGHTS MANAGEMENT
The discussion among SMEs about the IPRs management and protection plan opened during the Exploitation/IPRs meeting (held the 22nd of January 2014). The results under discussion for a potential protection plan have been:
▪ The CARESS reagents kit;
▪ The CARESS molecular probe;
▪ The CARESS automated device;
▪ The CARESS DPD deficiency screening.
The CARESS reagent kit is definitely an innovation and, as such, needs to be protected by patent.
The molecular probe, the new one in trials at CNR, is not a new molecule but its use in the CARESS testing process is new and will be patented as soon as the results are satisfactory. IRBM suggested that there may be space for patent protection on the use of the molecular probe in the CARESS context. The molecule itself may be not patentable because it is consolidated, well-known in the scientific and medical community and not covered by patent.
The CARESS device, as an instrument for the automated processing of the entire analytical procedure, is an innovation as well, because at this time, there is not another automated microplate processor with integrated centrifuge on the market. However, an investigation into whether this project / product includes enough innovation for a patent must be conducted. At the end of the project, we can state that, even if the prototype includes enough innovation for the market, it is still in an early stage of maturity. The patent application will be done in parallel with the certification process of the re-engineered device.
At the end of the project (M26) and after discussion among the SME beneficiaries, an adjusted final allocation of the ownership has been agreed upon as follows:
▪ CARESS prototype software, firmware and HMI
(optical scanner not included): AMEDA 100%
▪ CARESS optical scanner: OPTO 100%
▪ CARESS Protocol process (excluding molecular probe) ODPM 50% - AMEDA 50%
▪ Molecular Probe: IRBM 100%
▪ DPD deficiency algorithm based on CARESS protocol: ODPM 100%
This new reallocation fits better with the real interests of the SMEs and permits them to maintain the ownership, and assume the costs associated with the patent applications and maintenance therefore, only on those results in which they have an interest for the future of the instrument in the market.
In fact, in the initial allocation and foreground identification (as reported into the DoW), it seemed that in listing the IP for only the molecular probe, there was a gap in the process. While it’s been decided that each reagent is not patentable, the PROCESS may be. It is agreed that ODPM and AMEDA own an equal share of the Protocol IP.
During the meeting held in Brussels in March 2014, the PO suggested to start thinking about the way the SMEs intend to protect their knowledge. IRBM suggested a possible way to proceed to the consortium.
Another point which emerged during the open discussion of the SMEs, was the interest of potential private investors in the technology. OPTO suggested considering that investors are interested in companies and not in the Consortium. The combination of the results can be patented. For instance, OPTO needs only to have the licence of use but they are not interested to be owner (or co-owner).
Another possibility is to create a new company, which would own the patents and manages the activities (production and commercialization). In this case, the new company has to be legally and financially structured. IRBM reminded that the SMEs will have to cover the costs of this now company. In any case, this scenario will only be realistic in 3 years, after the industrialization and clinical validation stage when the instrument is ready to put on the market but not at this juncture, when the results are in a preliminary stage (research in relevant environment).
At this time however, even if the project is in an early stage, it is important to understand how to approach investors, because the consortium is not an entity and each SME can only approach investors at this point with its own piece of knowledge, which would not be enough for an investor to decide to support the project. In any case, it is agreed that an approach to potential investors without a protection of the knowledge is a weak position.
Finally, it is agreed to try to patent the device and the process together. Due to the high costs, it is agreed to start with 1 country and then extend the patent within the appropriate timeframe.
CONCLUSION
The CARESS protocol went through several modifications during the project. In only 26 months the project passed from a conceptual model to a chemically validated protocol in relevant environment. It is likely that the protocol is not in its final version, so further modifications will be necessary but the pillars have been built. It is a strong result, considering how CARESS is a small project in terms of budget and time, but it is still not enough to make the result a profitable solution for the market at this time. For this reason the SMEs need to put a complete trial in place, but the preliminary results are promising.
All of the changes and adjustments made to the protocol during the project were quite successful, and the more recent ones were mostly to accommodate the instrument; mostly related to the fact that the parameters of the instrument were changed to include a single dispensing needle as opposed to 2 or more channels for the dispensing of the reagents.
In conclusion, the CARESS PROTOCOL is a solid chemically validated analysis which does provide the expected result.
As for the prototype, this technology is in a less mature state. However, it has to be considered that that is an alpha prototype, therefore, by definition, a unit not ready for the market, because it is imperfect and does not respond to all of specific requirements and regulations yet. In 26 months, starting from scratch, it is not possible to have more than an alpha prototype (the machine has been manufactured in less than 1 year), and this is what expected by the research project. Maybe better results could be obtained in the calibration curves, but all the problems encountered can be all managed during the “reengineering and optimization” phase after the initial project. These are activities which perfectly fit with the PHC-12 Call. There is nothing in the “technology” that cannot be modified to make the machine a viable final product.
It is imperative that the results of the project, both for the CARESS protocol and the CARESS Instrument remain confidential, in order to increase the probability of success once both the instrument and protocol are brought to market. It is important for all of the partners to be able to have the benefit of having the ‘first mover’ advantage when this innovative product is brought to market.
List of Websites:
The CARESS WEBSITE is one of the most important assets for the dissemination of the project. During the design and test phase of the website one of the goals has been the definition of the key message to be communicated through the site and the way to communicate it. Particular attention was paid to the “receivers” (the end users), what they need to know about the project and how the message should be communicated.
The idea behind the banner used is to transmit the message that the Caress technology provides hope to cancer patients. In this final release, the logo, the soft green palette and the sprouting plant evoke life and growth. The layout of the webpage has been studied with the purpose of creating a screenshot that uniquely identifies the project. The link of the website is: http://caress-project.eu.
In our age, cancer is the most common cause of death in Europe. Amongst all the chemotherapy agents, 5-FU (5-Fluorouracil) has been in use for 40 years, and is one of the most successful and widely employed. In general, 5-FU is relatively well tolerated at standard doses, but around 8% of patients manifest a genetic variation that leads to a deficiency of an enzyme called Dihydropyrimidine dehydrogenase (DPD) that is crucial for the metabolism and deactivation of 5-FU. This causes several toxic reactions. The CARESS Project is a joint effort of 3 research Institutions from Italy and France together with 4 SMEs (small and medium sized entities) from Austria, France and Italy, to develop an accurate, easy-to-use and cost efficient technology for determination of the DPD activity level.
The protocol has been setup by the CNR and the spectrophotometric evaluation of the DPD activity has involved four main activities and has given rise to many others side activities, crucial for the optimization of the developing system and for the solution of the encountered critical points. The main activities managed for the development of the protocol have been: definition of the experimental protocol for the reaction of the selected Molecular Probe (MP) with DPD; selection of a suitable MP able to release, after reaction with DPD, easily detectable ions; assessment of the protocol for total inhibition of Thymidine Phosphorylase (TP) reaction; definition of the protocol for the spectrophotometric determination of ions released by MP. ICO has validated the full CARESS protocol with HPLC method and then they have proceeded with the validation with UV measurement.
From the technological point of view, LABOR consolidated of the prototype concept, the mechanical and electronic design (carried on partially during the P1), the selection of all the components, the manufacturing of all the components and the preliminary tests on the single components before the integration of the whole prototype. The final machinery can be considered a real automation device having to perform rapid and precise operations and displacement of fluids.
Finally, the field tests were to present the protocol followed with the CARESS device and the controls test needed for the validation of the prototype. It aimed to carry out and to evaluate both CARESS protocol and prototype, mimicking the real operating conditions within a diagnostic laboratory. The protocol developed by CNR was modified and adapted (in terms of volumes and timing of dispensation) to the prototype. All the experimental tests made at ICO for the validation of the prototype have followed the protocol. In order to define the performance of the CARESS machinery in the detection of DPD activity, and its potential deficiency, results obtained both manually and by Caress device were compared.
Project Context and Objectives:
In our age, cancer is the most common cause of death in Europe, with around 3,2 million new cases and 2 million deaths each year (7,6 millions in the world), with a terrifying increasing trend which will lead to 15 new million cases per year within 2020, as reported by WHO.
It’s the therapeutic approach used to fight this plague that is casting a doubt on the “Primum non nocere” principle; everybody who had the misfortune to get to know the severe side effects of chemotherapy- based approaches can understand how the border line between care and harm is often, dramatically, crossed.
Amongst all the chemotherapy agents, 5-FU (5-Fluorouracil) has been in use for 40 years, and is one of the most successful and widely employed, used in the treatment of breast, colon, and skin cancer, which are three of the most frequently occurring malignancies.
In general, 5-FU is relatively well tolerated at standard doses, but there are some important issues that must be duly taken into consideration:
• Around 8% of patients manifest a genetic variation that leads to a deficiency of an enzyme called Dihydropyrimidine dehydrogenase (DPD) that is crucial for the metabolism and deactivation of 5-FU. This causes several toxic reactions like mucositis, diarrhea, neutropenia, cerebellar ataxia, cerebellar dysfunction, and can even be fatal at the very first dose of 5-FU, with a mortality rate of about 0,5%. This deficiency must be properly identified before the beginning of the therapy, since the patient who has this disorder will never be aware of it until they suffer from a cancer that is treated with 5-FU chemotherapy.
• About 30% of patients suffer from severe toxicity effects after being treated with 5-FU.
• Despite many approaches having been proposed for DPD deficiency screening, none of the current strategies is adequate to mandate routine DPD testing prior to starting a 5-FU based therapy.
• The current gold standard approach for calibrating 5-FU drug therapy, as also happens with many other chemotherapy agents, is based on body surface area estimation: the patients are administered a dose in milligrams based on the actual square meters of their body surface, and this has been established without a rigorous scientific support. Literature studies show in fact up to 100 fold variation in the levels of 5-FU in plasma between different subjects with the same body surface area , thus resulting in the ineffectiveness of the therapy and in harmful side effects for the patients. Besides, 5-FU has a narrow therapeutic window and to establish the right dose it is necessary to set a trade-off between toxicity and therapeutic effectiveness.
• Surprisingly enough, studies have found out that many patients are not being treated with the appropriate doses for achieving optimal 5-FU plasma concentration.
o Only 20 – 30% of patients are treated in the optimal dose range;
o Approximately 40-60% are underdosed;
o Approximately 10-20% are overdosed ;
• Over the last 20 years, the scientific community has agreed on the reduced toxicity and improved clinical outcomes of pharmacokinetic dose management. Pharmacokinetic in fact takes into account how the human body processes the drug, and by knowing the specific response of each subject to the administered dose it is possible to personalise the therapy around each patient’s specific profile, according to well-defined ranges .
All these facts boil down to a simple evidence:
If DPD deficiency is so dangerous, why isn’t there a mandatory approach to detect it? And, if a dynamic management of the therapy is more effective and less harmful, why is it so difficult to proceed this way?
Because currently there is not a standard, simple, fast, reliable, and cheap way to detect the DPD activity level, which also allows clinicians to calibrate the therapy following well-established clinical algorithms.
This revolutionary approach could be employed in every medical center dealing with cancer, and could greatly enhance the effectiveness of the therapy, with enormous savings for the Healthcare Systems and, most importantly, safeguarding the priceless value at stake: to improve the patient’s quality of life.
Despite the fact that many methods have been tried for DPD deficiency screening, none of the current strategies are adequate to mandate routine DPD testing prior to starting a 5-FU based therapy.
The CARESS Project is a joint effort of 3 research Institutions from Italy and France together with 4 SMEs (small and medium sized entities) from Austria, France and Italy, to develop an accurate, easy-to-use and cost efficient technology for determination of the DPD activity level.
The CARESS project proposes to implement the first, compact, cheap, accurate and standard system for measuring the DPD activity level, thus allowing for rapid screening for DPD deficiency and for real-time adjustment of the therapy administered to the patient through a pharmacokinetic approach, therefore maximizing the therapeutic efficacy, while at the same time reducing the costs associated with the management of toxicities by limiting the dangerous side effects for the patients. The system requires only a small blood sample and will rely on the chemical interaction between a 5-FU analogous chemical probe with the sample, which will release ions in the solution that can be easily detected and measured through optical analysis.
Project Results:
CARESS requirements and specification
The first step of the project was the identification of the stakeholder groups. When a new process is intended to partially or totally replace existing procedures and common practices, it is important to understand how the interested stakeholders are likely to react, and to define their needs and requirements, in order to put a strategy in place is the most likely to succeed. The stakeholder groups are:
1. Final beneficiaries of the system/method (indirect use)
2. Professionals
3. Public/Private Cancer care centers
4. Public/Private hospitals
5. Public/Private Labs
6. Private Clinics
7. Supporters/Institutions
Then, the analysis focused on the end-users of the CARESS system, their current unmet needs and their points of view. For each end-user identified by the analysis, a general description of the profile has been made, while their points of view are depicted taking into consideration the differences that could be present in the same profile working in a different reality. A further analysis of the CARESS end-users has brought the consortium to define three main categories of interest:
▪ the medical biologist or chemist,
▪ the laboratory technician,
▪ the oncologist (or physician with a clinical oncology practice).
The use cases, then, were really important in order to identify several possible interactions between the end-user and the CARESS system and served as a guide for leading to further development of functional requirements.
In conclusion, the requirements supplied by the beneficiaries (SME and project partners) have been collected and organized.
After the collection of user requirements, the system specifications have been defined. The main goal was the assessment of specifications able to address a defined set of user requirements.
In order to perform the best analysis of every technical and functional aspect of the system CARESS and to keep track of any changes, it was decided to organize the collection of the system requirements into sections:
▪ General system requirements (GR)
▪ Protocol system requirements (PR)
▪ Electronic system requirements (ER)
Finally, a preliminary configuration of the system layout and architecture was drafted, according to the requirements collected in the first period of the project, providing an overview of all the components chosen for the prototype to be developed in the project.
Protocol development
The setup of the protocol for the spectrophotometric evaluation of the DPD activity has involved four main activities and has given rise to many others side activities, crucial for the optimization of the developing system and for the solution of the encountered critical points.
The main activities managed for the development of the protocol are:
▪ Definition of the experimental protocol for the reaction of the selected Molecular Probe (MP) with DPD;
▪ Selection of a suitable MP able to release, after reaction with DPD, easily detectable ions;
▪ Assessment of the protocol for total inhibition of Thymidine Phosphorylase (TP) reaction;
▪ Definition of the protocol for the spectrophotometric determination of ions released by MP.
From the experimental point of view, the first three activities are strictly interconnected, and, in particular, both the selection of the MP and the assessment of the protocol for TP inhibition are subordinated to the definition of the protocol for the reaction of MP with DPD.
Definition of the experimental protocol for the reaction of the selected MP with DPD
According to the Dow, the DPD activity has to be evaluated on a blood sample, i.e. on the Peripheral Blood Mononuclear Cells (PBMCs) contained in the blood. The spectrophotometric determination of the CARESS system requires a transparency of the samples in the 400-800nm range of vis-spectrum. For this reason the first performed activity is the evaluation of the best protocol for PBMCs separation from the whole blood sample. For our purpose, the best protocol might produce a high yield of PBMCs and, as stated before, a transparency in the 400-800nm range of vis-spectrum. Three protocols were evaluated, two of which are very similar and isolate the PBMCs by a density gradient centrifugation called Ficoll gradient. The third method is the method currently used by one of the RTDPs (ICO) for the separation of small fraction of PBMCs from blood. This method is very simple and fast compared to the Ficoll gradient method and was evaluated with the aim of simplifying the CARESS methodology. We concluded that the Ficoll gradient obtains the highest yield of PBMCs (1.600.000/mL of whole blood) and transparency in the 400-880 spectral window.
CNR (IMC) exploited samples of blood from voluntary healthy blood donors, courtesy of the Unità Operativa Complessa di Immunoematologia e Medicina Trasfusionale – Azienda Policlinico Umberto I, in order to extract PBMCs and set up the experimental protocol. In order to simplify the processing of the samples and, most of all, the analysis of the obtained data, since the beginning of the second period, the blood samples allocated for each experiment were mixed together to obtain a pool. This pool can be reasonably considered representative of the mean population. Therefore, the degradation rates of the pool can be considered as the mean degradation rates values of the population.
Two different approaches for the reaction of MP with DPD in PBMCs were explored; the first exploits intact PBMCs, the second cytosolic extracts of PBMCs. In fact, we cannot be sure that the selected MP will enter PBMC and, if so, in a reasonable time consistent with the experiments. In both approaches the experiments were performed modulating the number of PBMCs, the time of the reaction and the concentration of the MP. In order to verify the effective degradation of MP (that is the reaction of DPD with MP) a dedicated HPLC method was developed.
The obtained results pointed out that one of the selected MP is able to produce ions upon the reaction with DPD, both in the presence of intact PBMCs and of cytosolic extracts, being the amount of ions produced higher with cytosolic extracts. Once clarified this crucial point, the consortium as a whole decided to carry on the experiments aimed at evaluating the degradation of MP only with intact PBMCs, because of the known advantages of this approach.
Selection of a suitable MP able to release, after reaction with DPD, easily detectable ions
The MP was selected in a pool of molecules (commercial or not) structurally correlated to the chemotherapeutic agent 5-FU and, therefore, capable of interacting with the enzyme DPD. The common feature of these potential MP molecules is their supposed ability to release, upon interaction with DPD, ions that can be easily detected by optical analysis. The starting selection of the MP was guided by its cost and by the chemical nature of the ions potentially released upon interaction; in fact the nature and the amount of the released ions have a strong influence on the spectrophotometric determination. During the project two different MP were tested in the experiments, one able to react with DPD but also with the other enzyme involved in pyrimidine metabolism, while the other capable of escaping the TP pathway. This second MP was evaluated in hope of making the overall CARESS protocol easier, but, unfortunately, this strategy was not successful.
Assessment of the protocol for total inhibition of TP reaction
Both the protocols for the total inhibition of the TP enzyme in cytosolic extracts and in intact PBMCs were set up, exploiting a molecule described in the literature as inhibitor of this enzyme in vitro. To set up the protocols, in the experiments the inhibitor concentration, the incubation time with the inhibitor before the addition of the MP and the number of PBMCs were varied. In order to verify the effective degradation of MP in the presence of the TP inhibitor (that is the reaction of DPD with MP) the previously developed HPLC method was exploited. A summary of the results is reported in Table 1.
Table 1. Results of the experiments with intact PBMCs in the presence of TP inhibitor at 1h and 2h. Δ[MP] is the variation of the molecular probe concentration.
The TP inhibitor at 0.75mM concentration, at both incubation times, and 0.5mM concentration, at 2h, appeared to alter the permeability of the cell membrane. In fact, in the presence of TP inhibitor 0.5 and 0.75mM at 2h and 0.75mM at 1h we often found a Δ[MP] higher than in the absence of the TP inhibitor. Therefore, the less time consuming inhibition of TP in intact PBMCs can be performed with the TP inhibitor 0.5mM with an incubation time of 1h. The difference in the degradation of MP in the absence and in the presence of the inhibitor is very small; for this reason, in order to obtain reliable data on real patients the measure on each blood sample has to be performed in triplicate.
Because the integration of a thermomixer in the CARESS automated system could be a problem both in terms of design and costs, experiments were performed in order to verify the influence of mixing on the reaction of both TP inhibitor and MP with intact PBMCs. To this end, we have evaluated, on samples from the same pool, the degradation rate of MP in a thermomixer at 37°C or in a thermostating apparatus at 37°C without mixing. The results are reported in Table 2.
Table 2. Results of the experiments with intact PBMCs in the absence and in the presence of thermomixing. Δ[MP] is the variation of the molecular probe concentration.
Surprisingly, in the absence of mixing the amount of MP degraded both in the absence and in the presence of TP inhibitor is higher, and the inhibition is more efficient. This evidence is of course advantageous for our purposes, both for the automation of the CARESS system and for the determination of ions produced by the degradation of MP.
Definition of the protocol for the spectrophotometric determination of ions released by MP
The selection of the spectrophotometric method for the determination of ions produced by the reaction between the molecular probe and the DPD enzyme started at the beginning of the project, together with the set-up of the experiments for the selection of the molecular probe. At the beginning the amount of ions to be detected was unknown, therefore we explored two different chromogenic reactions operating in two different dynamic ranges of the ion concentration, and therefore we set up the UV detection protocol according to method_1 and method_2. In the conditions set up for the reaction of MP with DPD in the presence of intact PBMCs, the amount of ions to be detected is of the order of 0.210 μg mL-1. This amount of ions can be detected by the chromogenic method_2. Therefore, the calibration curve for the ion of interest in the proper dynamic range was constructed and the influence of PBMCs matrix on the chromogenic reaction was evaluated.
The results of the UV determination of ions released after the reaction of MP with DPD by the method_2 are reported in Table 3. The results of the UV determination were compared with those obtained by HPLC analysis of residual MP on the same samples in order to validate the UV protocol.
Table 3.Results of UV determination of released ions by the chromogenic method_2 in the reaction of MP with PBMCs. CTRL1 and CTRL2 are samples quenched a t=0 for the evaluation of the initial concentration of MP; C1-C4 are the samples incubated with MP in the presence and in the absence of TP inhibitor.
Results reported in Table 3 indicate that the concentration of released ions evaluated by UV measurements is not in agreement with those by HPLC, i.e. the standard method used as a reference.
The discrepancy between HPLC and UV results could be ascribed to two different causes, the first one is that the selected MP is not an inert molecule with respect to the chromogenic reaction exploited in method_2, the other is that ions released by the reaction of DPD with MP escapes from PBMCs only partially. These hypotheses were verified and confirmed by a lot of ad hoc experiments. That being so, (partial escape of ions from lymphocytes and reactivity of MP in the selected chromogenic reaction) a solution was found for evaluating, still by UV measurements, the degradation of MP. Actually, instead of the ions released by the reaction of DPD with MP we decided to evaluate the residual amount of MP after the reaction. This evaluation can still be done (with some expedients) by the same chromogenic reaction and therefore, doesn’t involve a change in the UV protocol and in the detection system of the CARESS machinery.
Therefore, a new UV detection protocol (always based on the chromogenic reaction 2) was set up and the degradation of MP upon the reaction in PBMCs was evaluated. Results of degradation of MP evaluated by UV measurements and by HPLC are reported in Table 4. The results are the mean values of 4 experiments. In each experiments each sample (the control, the samples in the absence and in the presence of TP inhibitor) were performed in duplicate. The degradation of MP is reported as variation of the concentration of MP with respect to the controls.
Table 4. Degradation of MP evaluated by UV measurements and HPLC.
The degradation found by the UV protocol and the HPLC method used as a reference, are in agreement. We did not find significant differences in the UV/HPLC correlation between the samples in the presence and in the absence of TP inhibitor. In Figure 1 the correlation between the results of UV and HPLC analysis is reported. At a first glance the correlation between the two sets of data could seem not so good; actually, in the literature, the comparison between different analytical methods displays similar correlation coefficients.
Figure 1. Correlation between the variation of the concentration of MP with respect to the controls, evaluated by UV and HPLC analysis.
In order to confirm by another experimental technique the goodness of MP as molecular probe for the determination of DPD activity, we have performed experiments exploiting a method currently used for the evaluation of DPD deficiency in the laboratory of Advanced Molecular Diagnostics (DiMA) at “Azienda Ospedaliera S. Andrea - Facoltà di Medicina e Psicologia, Sapienza University”. This routine evaluation of DPD deficiency relies on the determination of 5-FU degradation in PBMCs, analyzed by a HPLC-Tandem Mass Spectrometry method. These experiments were performed thanks to the courtesy of Prof. Maurizio Simmaco and Dr. Luana Lionetto. We can confirm also by this technique that the selected MP is a suitable probe for the CARESS system because it is degraded by PBMCs and that the degradation is different in the presence and in the absence of the TP inhibitor.
Protocol evaluation and validation
The purpose of this activity was to validate the feasibility of the full CARESS protocol created by CNR with the optical system developed by LABOR and his possible use as a diagnostic routine method. ICO has validated the full CARESS protocol with HPLC method and then they have proceeded with the validation with UV measurement.
The protocol developed by CNR contemplates several steps. ICO followed the steps reported below:
A. PBMCs isolation from blood sample with a manual or integrated Ficoll method
We looked to optimise the full CARESS protocol to increase the feasibility of the protocol and to simplify it for the diagnostic routine. For this purpose, we also checked the possibility to isolate PBMCs into blood draw tubes with Ficoll gradient integrated and their conservation after isolation.
In the way to validate this method for the PBMCs isolation, we made the full CARESS protocol with PBMCs isolated manually or with the CPT Ficoll tubes®. The blood samples came from the same volunteer.
B. TP inhibition, molecular probe degradation by DPD
After separating the PBMCs, we proceeded with the preparation of 4 aliquots: two with the addition of the TP inhibitor and the other two without the addition of the TP inhibitor. For each pair of aliquots one of the two serves as a control (reaction stopped at time zero), while the other is the sample treated according to the protocol in order to evaluate the DPD activity. Each pair of sample and their controls has been evaluated with the HPLC and the UV measurement.
C. Chromogenic reaction for UV measurement allowing the indirect DPD activity evaluation
We performed an automatic UV measurement with a microplate reader. Moreover, we decided to perform 5 replicates for each sample, considering as final Absorbance value the mean of these 5 replicates. The HPLC and the UV results are quite in line each other, but they must be confirmed with a large series of samples. However, we can conclude that the results obtained at ICO and at CNR are comparable and the protocol is well defined.
D. Full CARESS manual protocol
After separating the PBMCs with the manual Ficoll method, we proceeded with the preparation of 4 aliquots: two with the addition of the TP inhibitor and the other two without the addition of the TP inhibitor. For each pair of aliquots one of the two serves as a control (reaction stopped at time zero), while the other is the sample treated according to the protocol in order to evaluate the DPD activity. To conclude, the full manual protocol developed by CNR was verified by ICO successfully.
E. CARESS protocol optimizations
In order to evaluate the conservation of the samples after dilution with sulfuric acid 0,2M obtained in the previous experiment, they were aliquoted and stored. In conclusion, we need to make directly the chromogenic reaction after the probe reaction of to store it not diluted at -20°C.
In order to evaluate the stability of PBMCs and the DPD activity after samples conservation, we explored different possibilities routinely used for maintaining PBMCs in cell cultures. PBMCs recovered from the same blood sample were divided in several aliquots to test the different conservation methods. One aliquot was treated following the CARESS protocol and was used as the “control” for the comparison with the conserved PBMCs. None of the methods tested thus far have achieved positive results. After 24h conservation we have observed a reduction of the number of living cells, compared to the number obtained for the control sample. Moreover, we observed a different DPD activity in the controls and in the conservation tests. Unfortunately, we have not be able to reduce or simplify the protocol in terms of PBMCs isolation or conservation.
Then, further activities were performed in order to evaluate the possible correlation between the CARESS evaluation of the DPD activity and:
▪ the DPD deficiency evaluates by a multiparametric analysis performed at the ICO before treatment with fluoropyrimidines (5-FUODPMTox TM);
▪ the 5-FU plasma concentration at the first cycle of 5-FU treatment (5-FUODPM Protocol TM).
The scope of those evaluations was to permit the development of new algorithms, by ODPM, for dosage adaptation of 5-FU throughout the entire treatment course. The frequencies of those mutations are very low: which is why validating a new protocol and comparing it with what is currently in use would require a very large prospective clinical study. At the ICO, the two most frequent variants are: IVS14+1 G>A and 2846A>T (D949V). Both of these have been founded in patient samples with an allelic frequency of 1-1,3% among more than 20,000 patients.
The frequency of the complete DPD deficiency is around 0,2%. In consequence, in order to reach a significant relevance of the results and to be able to evaluate the correlations defined previously, the study would need to include at least 3000 patients, each of whom would be evaluated using both methods for comparison.
For this explorative project, we have decided to use only the biological waste of the ICO patient samples, when there was enough blood left in the sample after routine testing had already been accomplished. Therefore, it was not possible to make a statistical correlation. A larger prospective clinical trial will be required to determine if the CARESS method is an effective measure of DPD activity (which could be submitted as a Horizon 2020 European project).
At this point, the CARESS method does not evaluate the plasmatic level of 5-FU, but a substitute of that molecule to determine the DPD enzyme activity. Thus, adjusting the CARESS protocol to quantify 5-FU during the treatment, and also use the CARESS prototype for this additional analysis is quite far beyond the scope of the Caress project. This aspect would need much further R&D and is not only outside the scope of this project, but is also outside of the scope of the follow-up clinical trial project for which the consortium will be seeking financing over the next few months. At this time, the best way to study the metabolism of 5-FU during treatment is to measure the 5-FU concentration in plasma by a chromatographic method, which is already in current practice at ICO.
In conclusion, the small sample size did not allow us to make desired correlations to clinically validate the protocol for determining the risk of toxicity to fluoropyrimidines before treatment. The results, however, have clearly suggested that the manual protocol is quite robust. When a prospective clinical study aimed at clinical validation of CARESS method is finalized, this manual protocol may be used without further modification.
Electronic development
During the P2, the project entered in the core of the activities with the consolidation of the prototype concept, the mechanical and electronic design (carried on partially during the P1), the selection of all the components, the manufacturing of all the components and the preliminary tests on the single components before the integration of the whole prototype.
These activities represented the critical path of the project, under the engineering of the prototype point of view, and a deep management of the timing has been fundamental for tracking the single steps and be sure that no major delays could affected the entire plan of activities.
After months of designing functional components (P1) and the integration of the device, the first hypothesis of the control system has become an expanded system. The final machinery can be considered a real automation device having to perform rapid and precise operations and displacement of fluids. The control unit manages a high number of “peripheral” units and heterogeneous movements and automations, i.e. micrometer movements of a three axes portal or the presence control of the involved consumables, passing from the monitoring of an incubation chamber.
The Caress instrument, from the mechanical design point of view, can be considered as a machine for dosing small amount of liquids (sample, reagents, cleaning fluids), in and out of defined volumes where other operation are taking place (centrifuge, mixing, optical measurement). To perform these tasks, an ad-hoc designed dispensing group is mounted on top of a Cartesian plotter, with X and Y-axis corresponding to the working area where liquids have to be transferred. A Z-axis movement is necessary to put the needle in contact with reagents and samples. The intersection of rows identifies the points of the working map.
The machinery consists of several sections controlled by a central control unit that it is connected to the front Panel-PC (the operator interface). The control unit has expansion boards that adapt the signals and are used as mechanical support where necessary.
Summing up, the machinery is composed by the following functional blocks which have been designed and manufactured independently before the integration into the prototype:
▪ Power Chain (from 220V to +48V, +24V, +12V and +5V): the distribution of the power is ensured by starting from the 220V power supplies (from a typical wall outlet) through stabilized power units which are connected directly to the related sections. The external components (like pumps, air blowers, etc.) are supplied directly from 12V entering inside the control board and are controlled by the relative relays. This section supplies all sensors, motors and the HMI section too. All selected power supply elements are integrated inside the machine. They have the appropriate certifications to be used in laboratory as prototype and they are all commercial components in compliance with all applicable regulations. Four power supply units provide: +48 V DC for centrifuge motor, +24 V DC for syringe, +12 V DC for three axes motor, for panel pc and external components, +5 V DC for uP control unit and extension boards.
▪ Control Unit (Main Unit and controllers): the Control Unit manages and controls the communication and the co-ordination between input/output and devices/sections. It reads and interprets instructions and determines the sequence for processing the data. It directs the operation of the other units by providing timing and control signals. It performs the tasks requested by Panel-PC. The control unit takes care of the handling and control of: gantry movements, reagent and sample dispenser, security, spin monitoring, thermostat monitoring, controls the filling level of tanks, cleaning monitoring system and HMI communication. To better adapt control to the machine, the central unit has been designed and built from scratch using a microprocessor custom architecture with a central board connected to other secondary boards.
The main board (also called Control Unit) is the core of the whole system. It is connected with all components. On the board there are two communication ports: one is a direct serial port used for status report of the machinery, the second one is composed by an isolator and a USB to serial UART interface to be connected directly to the panel pc. Two LEDs (D2 and D3) monitor the transmissions. This port is also used to reprogram and update the uP via HMI.
▪ HMI Section (with barcode scanner): the CARESS HMI is composed of a Panel PC with a resistive touchscreen (to be used with gloves). The construction has been chosen to be as robust as possible even in difficult environment. The display has a visible area of 10.1” (1024x600 resolution). The display is arranged to be connected to a barcode reader via USB connection. The display will communicate to the operator all the necessary information to prepare properly the test cycle. It will inform the operator of the correct procedures to check the consumables loading (deep-well plates, reagents and distilled water), and allows operating the software. The barcode scanner will read an identification code for both the samples and the reagents that are loaded into the instrument. After each scan, the software will indicate where to place the respective reagent or sample, and will verify its correct positioning using optical sensors and/or micro-switches.
▪ Centrifuge Section: the centrifuge can be considered as an independent section controlled in its functions by the central unit. A chemical centrifuge works on the principal of centrifugal force. The high-speed revolution of the basket separates the supernatant and the precipitate (and cells dishes). This section is thermostatically controlled to allow an incubation period without further moving the samples. This section is composed of several electronic and mechanical parts. The inner part of the centrifuge is designed to be warmed and heated to a controlled temperature of 37°C (as required in the measurement protocol). Under the exposed bucket is placed a box containing photo-detectors with a reduced beam. Six optical sensors are placed under the respective sites of the six polypropylene tubes to control the positioning.
▪ Gantry control and Syringe: the gantry lifts objects and can move horizontally on a rail fitted under a beam. It is driven by three autonomous stepper motors and has several safety limit switches for each axis. This section requires electronic controls to drive the movements. The Y- and Z-axes are controlled by two twins and autonomous controllers through two communication ports. The X-axis is composed of an alpha step closed loop stepper motor. The dispenser is composed of a syringe with a washable needle. It is moved by the gantry on the surface of the working area. This dispenser is able to pick up and release liquid from the reagents bottles into sample tubes and micro-plate. The syringe piston is actuated via a lead-screw with 30 mm of travel, driven by a stepper motor.
▪ Optical Unit: the scanner has eight emitters and eight detectors to speed up the readings. It reads a loaded 96-wells microplate (128 x 86 mm) prepared by machinery. This section is composed of three boards connected to the control unit. The Optical Unit mainly consists of light sources transmitting light through the samples and photodiodes detecting the absorbance of the samples. The unit is integrated by the software, which collects, records and analyses the generated signals. The set of two circuits measures the ratio of light not absorbed in the wells (and, then, in the reference ones) at the selected wavelengths (405 nm). The LEDs board is composed by eight UV lamps, which illuminate the microplate column where the scanner is positioned. In each branch, the flowing current is constant and stabilized by precision resistors and a reference voltage. The used LED has a very narrow bandwidth. The photodiodes on the opposite side of the light source generate a digital output proportional to the amount of light received (and therefore proportional to the absorbed light).
▪ Auxiliary Section (reagents and tanks checks, and washing control): the machine is completed with further components and a control board for the reactants. This section is composed of three pumps for washing unit, an air blower for needle drying and three level sensors for the tanks.
According to the testing phase of the project, all components were tested before and after the integration of the component in the related section and then all functionality tests were performed for each sections. All tests ran smoothly with the exception for the functionality of the washing chamber that has been revised and reworked. After the modifications provided, this section completely responds to the required specifications.
System Integration and preliminary tests at lab scale
As reported in the previous section of this document, the CARESS prototype required a design phase and a subsequent assembly phase which took a lot of effort during the period from M7 until M19 approximately. The single components have been assembled, tested, and reviewed in all their parts in the period between M19 and M22.
The final specifications of the machine are:
• dimensions: 870 mm (width), 620 mm (depth) and 760 mm (height) without stands;
• weight: about 68 kg.
The working plane hosts the functional sections of the instrument. These are:
1. Centrifuge Section;
2. Optical Unit;
3. Dispenser Section;
4. Reagents Stand;
5. Washing Section.
Below, the list of the auxiliary section by which the machinery is composed:
1. Power Supply Section;
2. Control Unit;
3. HMI Section;
4. Three tanks (for the containment of the distilled water and the ethanol and for discharging the liquid);
5. Three Pumps (two to send the washing liquids and one to force liquid drain);
6. An airblower (to dry the needle).
Connections are placed on the side of the prototype: white reset button for only the Control Unit, the ethernet port, external USB port with cover and the power plug with main switch (red light).
The system integration of the prototype respected the timing as reported into the DoW. The activities started at the end of the year 2014 (November – December 2014, M19/M20) to be concluded at the end of January 2015 (M21).
The Consortium decided not to publish pictures of the final prototype integrated in order not to lose the innovation. In fact, the prototype requires additional work after the end of the project for reengineering and industrialization. The publication of pictures in this delicate phase could be risky for the innovation because the prototype is not protected by patent yet, and the SMEs could lose their foreground and the first mover advantage.
After an intensive integration of all the components into the prototype, the activities went on with the preliminary tests at laboratory scale.
The preliminary tests were carried out in the chemical laboratories of LABOR. The strategy followed for the initial tests is listed below:
• Tests for the control of the necessary movements of the prototype, in compliance with the protocol, without liquids and reagents;
• Tests for the control of the movements loading the machine with distilled water;
• Tests campaign for retracing each step of the protocol with the real reactants.
The test campaigns have been conducted from the end of January to the middle of March 2015 until the shipment of the prototype to ICO.
Together with the CNR, LABOR began the first tests session of the CARESS protocol at the end of January. Taking as reference the protocol, the control tests were performed by running an approach from the bottom upwards: firstly last operations have been performed (optical measurements), which actually represent the final step of the protocol before obtaining the data on the CVS file.
The different tests sessions have been gathered in order to follow the same engineering approach followed for the design and manufacturing the prototype. Therefore, in each tests session, a specific component and operation of the machine have been tested.
During the first days the machine has been loaded only with distilled water. Labor and the CNR reviewed the entire protocol to align the prototype to the manual procedure. To optimize the timing of the microplate preparation the dosage of H2SO4 in the microplate (column 7) has been moved. This action also allows a better cleaning of the needle in the reagent exchange.
During one of the firsts tests, the liquid did not drop completely, it was necessary to update the firmware of the control unit to insert a specific step of pre-loading air in each dosage and for each quantity. The firmware update solved the problem.
After these optimizations, a new test was performed by filling the microplate with a manual preparation. However a problem occurred due to a bug in the software of the prototype. The test would have closed properly, if a 15-minute break would have been respected. This pause was not respected and this permitted to identify a specific bug in the software, which was solved.
The air taken before each dosage was increased to 30 µL and the speed of the plunger of the syringe was reduced to facilitate the exit of the last drop during the release of liquids. To better appreciate the levels of light intensity detected by the scanner the reading precision of the photodiodes was set to three decimals.
During a test, it was put in evidence that the dosage of the Br-/BrO3- for 24 patients took more than the pause required. We decided to make an intermediate test to measure the timing of Br-/BrO3- dosage for 24 patients (filling from the 2nd to 6th column) including washing. The cycle took more time than the expected. From this, we decided to increase the waiting time. The CNR performed tests to ensure that the protocol did not have substantial changes.
Before to analyze the obtained data, we have characterized the scanner. We performed a test doing a scan of the entire microplate filled with water in the dark and with the presence of external light; then we scanned the entire microplate filled with Standard 1 and Standard 5 with light and dark. The data did not show significant differences between the tests carried out in the dark and with external light.
Each time that we performed a specific test, the machine was controlled. Several functionalities were tuned: timing of the waiting and the synchronization of the actions.
Over the following days we performed tests for the characterization of scanner tuning the lighting time and the timing after the microplate preparation. We noted that increasing the illumination time did not have significant improvements, but only an increase of the timing of the scanning step.
In the following days we completed all actions on the machinery. Unfortunately, the results of the calibration curves were non-regular. This was due to the precision of the assay of the syringe. This problem was not revealed during the tests with distilled water. Using real reagents and sample, the needle seems to get “dirty” and reduces the precision in low volumes (10µL) where an incomplete detachment of the last drop dosed has a greater influence. To improve the precision of dosage, small volumes were doubled entailing several changes in the preparation of microplate.
The whole firmware has been adapted to the new protocol. Through many internal tests and a session tests with the CNR, we have double-checked all steps in the CARESS cycle.
Using larger volumes we obtained a greater stability but we did not solve the problem: in some experiments, the problem of the drop was observed again, unexpectedly. One of the optimization and improvement of the prototype will be certainly focused on this aspect. This can be solved with a study for a suitable coating to perfectly control the fall of the liquid from the needle.
Internal tests had been crucial not only to improve the functional aspects of each steps, but also to highlight two mistakes in the design of the washing station. These were:
• the geometry of the washing chamber;
• the necessity of a suction pump of waste liquids.
These two points have been completely solved by reviewing the machine in its functional design.
The following corrective actions have been carried out:
• mechanics rework of the washing chamber to change the internal shape;
• addition of a drainage pump.
During the tests sessions, we noticed a shift in the measured light not due to repeatability. We noted a natural evaporation due to auto-heating of the machine (power supply, motor drivers, etc.). This weight loss can be greater after a cycle of thermostating the centrifuge. This aspect, in a future revision, should be concretely considered and compensated with the involved liquids (quantity, viscosity, timing of release, etc.).
The linearity of each channel was tested by filling the first six columns (those used for the measurement) with several standards solutions of chromogenic reagent at different concentrations. The seventh column was filled with H2SO4 to have a luminosity reference.
All the necessary changes to improve the protocol have been implemented. Thanks to the modularity of the machinery and to the numerous parameters of the firmware, the debugging time and the revision work have been greatly reduced. The preliminary tests for the electronics and the mechanics have been completely preformed, as well as the operative tests for the automation from the software and firmware point of view.
The test phase highlighted some points which, during the industrialization of the prototype, will have to be managed and improved. All of those have been analyzed and the relative solution has been identified.
In conclusion we can state the followings:
▪ The CARESS prototype is a complex machine which correctly automates the protocol setup by the CNR;
▪ A complete tests campaign has been held between January and March 2015 for the debugging of the machine.
▪ From the engineering point of view we identified some weaknesses. Those have been analyzed and a related solution has been found. These solutions can be easily implemented during the industrialization of the prototype.
▪ The design phase of the mechanics resulted to the successful. In such a complex machine, only two adjustments were needed, both solved during the tests phase. The efficient mechanical rework of the washing chamber and the installation of a draining pump permitted to complete the washing cycle as it was designed and required in the WP1.
▪ The software of the machine revealed some bugs during the tests campaign, as well as the firmware had to be managed in order to optimize the automation and the waiting times. All the bugs identified have been analyzed and solved.
▪ The prototype was sent to ICO ready to be installed. At dedicated session of a week (Feb 9th –Feb 13th) has been held between LABOR and ICO, in the labs in Rome, in order to train ICO technicians for the correct use of the machine during the field tests. Further, an installation manual has been prepared as user guide for the partners.
▪ The Prototype has been designed and manufactured to be completely controlled and analyzed remotely, including a camera, filming the working area. With these features, LABOR was able to support ICO during the whole duration of the field tests.
Pilot projects
The field tests were to present the protocol followed with the CARESS device and the controls test needed for the validation of the prototype. It aimed to carry out and to evaluate both CARESS protocol and prototype, mimicking the real operating conditions within a diagnostic laboratory. The protocol developed by CNR was modified and adapted (in terms of volumes and timing of dispensation) to the prototype. All the experimental tests made at ICO for the validation of the prototype have followed the amended protocol. In order to define the performance of the CARESS machinery in the detection of DPD activity, and its potential deficiency, results obtained both manually and by Caress device were compared.
We are aware that this will be an important, and industrially ground-breaking results, and that a complete validation of the Dose Management Approach would require time, resources and budget, which are beyond the scope of this project.
The machinery was delivered on March 23rd 2015 at ICO (Angers). The first tests were made to evaluate the basal functions with LABOR assistance.
For the validation of the CARESS device, we firstly used cultured cells. The cell lines used were human colorectal adenocarcinoma. They were cultured respecting the GLP (“Good Laboratory Practice”) procedures and the recommendation of the provider ATCC (cultured in RPMI – 10% SVF solution). The choice of these cell lines was made based on the enzymatic DPD characteristics reported in the literature. Moreover, we have verified the genetic profile of these cell lines (sequencing) in order to evaluate the presence of the main important DPYD polymorphisms generating DPD deficiency: no major polymorphisms, responsible for a reduction in DPD activity, were found. Using these cells could be useful to verify if the CARESS system and/or the CARESS protocol are able to discriminate samples with different DPD activity.
Tests have been performed in order to check:
▪ Internet or intranet connection
▪ Software
▪ Dispensing
▪ Centrifuge
▪ Reagents evaporation (microplate)
▪ Biological constrains
▪ Calibration curves
▪ Repeatability/ reproducibility
The machinery was evaluated in comparison with the same protocol manually made. It took into account the waiting times, volumes and modified solutions (compared to the CARESS manual protocol). It aims at carrying out and evaluating the CARESS solution in real operating conditions, measuring through comparative analysis the performance of the CARESS device in the detection of DPD activity, and potentially his deficiency. The manual and automated protocols were developed to be easy to use and correspond to the diagnostic lab constrains. In way to define and validate their use in lab, we made some evaluation at different steps and levels.
A. CARESS device evaluation
Firstly, we evaluated the CARESS prototype device regarding several parts of both the device itself and the adapted protocol.
During the first tests with the CARESS device, we made some observations about the machinery protocol. All tests were defined in agreement with Labor and CNR and supervised by LABOR remotely using the internet interface with the instrument.
The syringe showed a problem during dispensing: we observed a random and not reproducible drop remaining on the needle. This deviation from the distributed volumes induces errors that were verified by reproducibility tests and with measurement of the dispensed volume in each well of the plate. After discussion with Labor and CNR, some possibilities exist to avoid this problem. This issue would need to be rectified before further testing of the machine protocol can be done.
Some observations were made about the prototype components. All this observations aims to optimize the CARESS device.
The CARESS software developed by LABOR showed goods results, followed the CARESS device protocol defined by CNR and presented the results in a table form as comma-separated values (.csv) file convertible into an Excel document. A continue internet or Ethernet connection is needed to avoid the arrest of the protocol.
The CARESS device protocol was evaluated regarding calibrations curves, and with both cell lines and water for the reproducibility.
The results obtained for intra- and inter-assay for the cell lines tested (HCT116 and DLD1) presented a deviation standard of the molecular probe evaluated up to 10%. It is too high to validate the device protocol both manually made and with the prototype. The results with water looked better, but they had a high standard deviation too. For the following tests with the patient samples, we decided to use the manual CARESS protocol already validated.
B. Final CARESS protocol results
In a second part, the evaluation of DPD activity was made, starting from blood samples (volunteers) or biological wastes (patients). The CARESS manual protocol was used in way to be sure about the results obtained.
From March 2015 to June 2015, 49 samples were analyzed following the manual protocol. The 49 samples analyzed derived from biological waste from patients treated at ICO. The R2 obtained for the calibration curves were > 0,99. The results obtained shown that there was a degradation of the molecular probe in all samples and the degradation rate was variable among samples, as expected.
An inter-assay was made for two samples and the standard deviations obtained were close to 10%. Other tests must be done.
Intra-assays were made for 5 samples, with 3 to 5 replicates into the same UV measurement plate. These results are encouraging because the deviation standard obtained for all these tests was under 10%, even if a standard deviation of less than 5% would be preferred. These results, compared to those obtained by the CARESS device, indicate that the manual protocol is robust and repeatable.
All the calibrations curves had a R2 value equal or up to 0.99. These optimal coefficient values of calibrations curves allow us to calculate the molecular probe concentration in the unknown samples accurately and reliably.
The reported results are very encouraging as they demonstrate that even slight changes in DPD activity are carefully highlighted by the CARESS manual protocol. This protocol is therefore sensible in the stratification of enzyme activity.
In conclusion, the CARESS protocol is promising for the determination of the lymphocytic activity of DPD. It is possible that the CARESS protocol may also be able to detect high metabolizer, but that occurred in only one sample, so no conclusions can be drawn from this event.
To validate the use of the manual CARESS protocol to screening the DPD activity prior chemotherapy, it is necessary to establish a clinical validation trial, including the analysis of at least 3000 patients samples to make sure to include deficient samples in the study. This trial should be multicentric and multinational in order to have the number of samples required as quickly as possible.
The weakness of this protocol, at present, is related to the PBMCs isolation that requires some technical skills and equipment. Moreover, PBMCs samples are not stable and maintaining DPD activity for long time (more than a few hours) is not possible (further research into this field would be very useful for making this protocol a clinical reality): the cells have to be isolated within two hours of the blood sample collection and the experiment must be done immediately after the isolation of the PBMCs.
Potential Impact:
The analysis of the market (held during specific workshops with the SMEs) showed that there is an urgent need of the determination of DPD deficiency prior the chemotherapy treatments but, currently, the methods used to determine the dose of 5-FU to give to a patient prior to chemotherapy treatment:
▪ do not take into account deficits or over-expression DPD. The result is that patients could be under- or over-dosed;
▪ are not specific and, when the specificity is high enough, the technique is far from adequate to be a standard in clinical practice due to the high costs associated.
In order to start identifying the potential impact the exploitation scenario of the innovation, the first step has been the definition of the framework. In particular, the DPD testing market and the In Vitro Diagnostics (IVD) sector have been described in their trend taking in consideration the literature defining the state of art. As main conclusion, “factors like the increase in awareness and acceptance of personalized medicine, advancements in molecular techniques, and increasing investments in genomics & proteomics research will spur the molecular diagnostics market in the future. Unfortunately, reimbursement issues for non-coded tests and complex regulatory framework will restrict the growth of this market. Introducing a new successful product is a challenging and complex endeavour. It can be estimated that the CARESS innovation will be not ready for the market before 2018, due to the inherent engineering, industrialization and certification processes for the device.
In addition, the IVD market has also been studied on the Player level. Although the market for IVDs is fragmented with more than 500 companies, the study delves into the business dynamics of major players in the industry that together form more than 70% of the global IVD market.
Then, the analysis of the external environment resulted in the first release of the CARESS SWOT analysis comparing the internal framework and highlighting the main strengths and weaknesses of the innovation.
The exploitation plan started with the segmentation of the market in order to focus the attention on the stakeholders potentially interested to the innovation. In agreement with the application scenario, the two main segment of interest are:
1. Small structures, with no/little spending power, which cannot afford the purchase of the machine. In this scenario, the reagents kit can be sold with the support of the ODPM’s web calculator for the interpretation of results. This scenario is the most profitable for the ODPM’s outsourcing model of the CARESS testing service.
2. Large structures, with adequate/high spending power, which can afford the purchase of the automated device. This scenario is the most profitable for the direct sale model for all the SMEs.
In both “scenarios” the SMEs will be able to act as first mover in the market, being market leaders in their sectors, to be able to bring the innovation to identified targets quickly and efficiently. This represents one of the key advantages for the exploitation of the solution.
Focussing on the exploitation of the solution, the SMEs deepened their discussions regarding the priority of the project, given the two possible scenarios. All of them identified, as the top priority , the Large Structure scenario (scenario 2) in which to exploit the sale of the fully automated device. However, even given the relatively higher barriers to entry, the Small Structure scenario (scenario 1) will be also investigated with the purpose of entering the market with an alternative lower-cost solution and/or with an indirect sale model.
The SMEs also discussed the geographical segmentation of the market. It was decided to view the market in two geographical segments when talking about “SMALL STRUCTURES”:
▪ SEGMENT A - Highly developed markets – oncology Centres
▪ SEGMENT B - Lower developed markets (Central & Eastern Europe, Asia, Africa, South America)
At the end of CARESS post project phase, the SMEs have identified the following products/services:
▪ CARESS KIT
o KIT ALONE: reagent kit, ready to be validated in a comprehensive clinical trial.
o KIT + ODPM CALCULATOR: reagent kit, ready to be validated in a comprehensive clinical trial, in coordination with the ODPM calculator.
▪ CARESS DEVICE
o DEVICE + KIT: validated prototype to be industrialized; further steps needed at the end of the project; specific supply package for the kit;
o DEVICE + KIT + ODPM CALCULATOR: validated prototype to be industrialized; further steps needed at the end of the project; specific supply package for the kit.
For each business solution identified (reported above), a value chain has been defined. With the outsourcing business model, ODPM will be the entry point in the market with the DPD deficiency screening service. In the direct business model, AMEDA will promote and sell the CARESS kit for manual use with the ODPM calculator as an option worldwide except France. AMEDA will also distribute the automated device, the ODPM calculator as an option worldwide, except France, where ODPM has the exclusive rights for the both the CARESS kit and device. IRBM and OPTO are always present in the value chains as exclusive suppliers of the probe/reagents and the electronics, respectively.
The first step to make the SMEs independent has been the technology transfer from the RTDPs, in order to shift the knowledge:
▪ AMEDA, OPTO and ODPM: electronics (AMEDA + ODPM), optical sections (OPTO), software and hardware of the automated device (AMEDA+ ODPM);
▪ IRBM and AMEDA: molecular probe and protocol;
▪ ODPM: DPD deficiency screening.
Specific meetings have been organized for demonstration and training of the SMEs, as well as materials to be used as user guides.
The most important outcome of the exploitation strategy has been the agreement on the follow up of the project and the identification of the most suitable funding scheme. Further, since the application is foreseen before the end of the 2015, a concrete application schema has been shown and already identified. This is in a preliminary stage and the consolidation is needed by the end of sept 2015, when a dedicated meeting has been scheduled.
Here are the key points on which the SMEs have set up the exploitation plans, which summarize the state of art at July 2015:
1. The CARESS prototype has been tested at laboratory scale and in a relevant environment but it is still in an experimental phase. The prototype must be further optimized in order to be exploited on the wider market. This will have to be done not only from the mechanical point of view but also from the optical and electronic point of view. Some weaknesses, modifications and optimizations have already been identified by the RTDPs. The final reengineering of the prototype will need at least 24 months and will be covered in the post-project phase of the plan.
2. The protocol for the DPD activity screening has been created, analysed and optimized by the CNR and tested in their laboratories. The results are good. The protocol has been tested in parallel in a relevant environment (ICO) and it has been validated. However, in order to launch a new clinical practice in the market, a multi-center clinical trial is needed. This will be the second element of the post project phase.
3. Given the specificities of the healthcare sector as well as IVDs and pharmaceuticals market, the beneficiaries agree that it is not possible to go to the market with an innovation without any means of protection. Therefore, as soon as the next phase is well underway and the preliminary results obtained in CARESS are confirmed and improved, a patent application will be necessary. It is agreed to apply for a patent for both the device and the protocol together.
EXPLOITATION PLAN STRATEGIC ROADMAP
With the validation procedure scheduled within the project’s framework, at M26, the protocol is ready for the clinical validation, which is out of the scope of the project, but that could start right after September 2015.
ODPM underlined the importance to describe the context of this clinical validation trial. This clinical validation trial should be organized as a 3 years study with a 2-years inclusion period, a multi–country, multicenter with a minimum inclusion of 500 patients per country. In this case, the trial could cost between 3 and 6 million euro and investors will be necessary.
However, it is a common thought that the protocol needs to be clinically validated before the instrument prototype is reworked and put into production. At the beginning of the discussion, the SMEs agreed to work on the industrialization of the prototype instrument in parallel to the clinical validation of the protocol, with the purpose to speed up the process. However, later on in the discussions, this was seen as a mistake and a waste of money. In fact, if the protocol does not prove to be an efficient analysis of DPD activity in clinical practice, then more work needs to be done on the protocol, which would then likely engender more changes to the instrument.
It may make more sense to do a smaller clinical trial (one center study at ICO) for a period of time to be determined (9-12 months), so that there are preliminary clinical results before the next phase with the final changes to the prototype and industrialization phase begins.
The preliminary study could be a feasibility study done over about one year, which could then give solid results regarding the clinical validity of the protocol (not just that it works chemically speaking) and then the consortium could move forward with the H2020 proposal much more confidently. The picture below reports the organization of the phases.
As for the industrialization of the prototype, the possibility to launch the system in the market requires an extensive testing phase after the end of the project. This is needed both to determine the level of differentiation of the CARESS instrument in relation to the systems already available but also to fulfill the requirements of the IVDs directive.
The post-project plan for the device will be divided into two phases: the first phase will include the reengineering of the prototype with the goal of reaching full compliance with the EU IVDs directive. Given the current state of the prototype, these goals are not too ambitious and are not too difficult to achieve once the re-engineering phase and new clinical testing are completed. In the first part of this phase, a re-engineering process will be required to prepare the industrialized production of the CARESS device starting from the validated prototype, ready at the end of the 2016. The main scope of this process is to test the performance of the system and make it completely compliant with the Directive. In fact, while an acceptable level has been reached with the prototype, a deeper investigation and optimization will be required to pass from the prototype to a commercial product.
Then, a deep testing phase is needed both at laboratory scale and in the field with end users. The activities concerned the reengineering and the tests will be strictly connected.
Likely, not only the compliance with the directive will be necessary but also the adaptation to new customer requests. This brings to the second phase. During the field test the main activity will be the collection of feedback and new requirements from end users. The new contacts with the customers will be also exploited to set up the marketing plan. This phase will require further adaptation of the device. It will be crucial to follow these activities in parallel in order to decrease the time to market as much as possible.
Finally, the commercialization can start only after this period of design revision for market use has been completed. The final product, in fact, even if in line with the technical performance, will need a revision under the marketing point of view. The market requires specific conditions that should be taken into consideration before the official launch.
The final product will be ready for market in 30-36 months after the end of the CARESS project. This means that the product cannot be commercialized before the end of the 2018.
FINANCING THE EXPLOITATION PLAN
During the Review meeting, held in Brussels the 21st of March 2014, the REA’s Project Officer suggested to start thinking about the follow-up of the project and the possibility to find new funds in order to finance the next stages. The PO suggested the “Business Angels” or “Angel Investors” as a possible way to go. During the 1st Year meeting (June 2014) held in Paris, two other possibilities were explored: The Worldwide Innovation Challenge 2030 originating in France, and the SMEs instrument of the EU program Horizon 2020. At the end of the project, the consortium agreed on the following:
▪ Innovation 2030 is not the suitable funding scheme for the follow up of the CARESS project.
• The suitable Call for CARESS seems to be the Call of the biomarkers, where the proposal can be related to the clinical validation and In-vitro diagnostics validation.
• CARESS could be applied directly in a Phase 2 trial before the end of the Year 2015. However, a specific channel for this point needs to be activated right after the end of the project in order to understand if November 2015 is a realistic date for the completion of the proposal. We also must determine if phase 1 is a necessary step or if the project is ready to go directly into phase 2 of the program and how the pre-clinical trial can be managed. In this case, robust and realistic data will be necessary for a detailed business plan.
At the time of wiring, the SMEs have already a proposal for the proponent consortium on H2020.
INTELLECTUAL PROPERTY RIGHTS MANAGEMENT
The discussion among SMEs about the IPRs management and protection plan opened during the Exploitation/IPRs meeting (held the 22nd of January 2014). The results under discussion for a potential protection plan have been:
▪ The CARESS reagents kit;
▪ The CARESS molecular probe;
▪ The CARESS automated device;
▪ The CARESS DPD deficiency screening.
The CARESS reagent kit is definitely an innovation and, as such, needs to be protected by patent.
The molecular probe, the new one in trials at CNR, is not a new molecule but its use in the CARESS testing process is new and will be patented as soon as the results are satisfactory. IRBM suggested that there may be space for patent protection on the use of the molecular probe in the CARESS context. The molecule itself may be not patentable because it is consolidated, well-known in the scientific and medical community and not covered by patent.
The CARESS device, as an instrument for the automated processing of the entire analytical procedure, is an innovation as well, because at this time, there is not another automated microplate processor with integrated centrifuge on the market. However, an investigation into whether this project / product includes enough innovation for a patent must be conducted. At the end of the project, we can state that, even if the prototype includes enough innovation for the market, it is still in an early stage of maturity. The patent application will be done in parallel with the certification process of the re-engineered device.
At the end of the project (M26) and after discussion among the SME beneficiaries, an adjusted final allocation of the ownership has been agreed upon as follows:
▪ CARESS prototype software, firmware and HMI
(optical scanner not included): AMEDA 100%
▪ CARESS optical scanner: OPTO 100%
▪ CARESS Protocol process (excluding molecular probe) ODPM 50% - AMEDA 50%
▪ Molecular Probe: IRBM 100%
▪ DPD deficiency algorithm based on CARESS protocol: ODPM 100%
This new reallocation fits better with the real interests of the SMEs and permits them to maintain the ownership, and assume the costs associated with the patent applications and maintenance therefore, only on those results in which they have an interest for the future of the instrument in the market.
In fact, in the initial allocation and foreground identification (as reported into the DoW), it seemed that in listing the IP for only the molecular probe, there was a gap in the process. While it’s been decided that each reagent is not patentable, the PROCESS may be. It is agreed that ODPM and AMEDA own an equal share of the Protocol IP.
During the meeting held in Brussels in March 2014, the PO suggested to start thinking about the way the SMEs intend to protect their knowledge. IRBM suggested a possible way to proceed to the consortium.
Another point which emerged during the open discussion of the SMEs, was the interest of potential private investors in the technology. OPTO suggested considering that investors are interested in companies and not in the Consortium. The combination of the results can be patented. For instance, OPTO needs only to have the licence of use but they are not interested to be owner (or co-owner).
Another possibility is to create a new company, which would own the patents and manages the activities (production and commercialization). In this case, the new company has to be legally and financially structured. IRBM reminded that the SMEs will have to cover the costs of this now company. In any case, this scenario will only be realistic in 3 years, after the industrialization and clinical validation stage when the instrument is ready to put on the market but not at this juncture, when the results are in a preliminary stage (research in relevant environment).
At this time however, even if the project is in an early stage, it is important to understand how to approach investors, because the consortium is not an entity and each SME can only approach investors at this point with its own piece of knowledge, which would not be enough for an investor to decide to support the project. In any case, it is agreed that an approach to potential investors without a protection of the knowledge is a weak position.
Finally, it is agreed to try to patent the device and the process together. Due to the high costs, it is agreed to start with 1 country and then extend the patent within the appropriate timeframe.
CONCLUSION
The CARESS protocol went through several modifications during the project. In only 26 months the project passed from a conceptual model to a chemically validated protocol in relevant environment. It is likely that the protocol is not in its final version, so further modifications will be necessary but the pillars have been built. It is a strong result, considering how CARESS is a small project in terms of budget and time, but it is still not enough to make the result a profitable solution for the market at this time. For this reason the SMEs need to put a complete trial in place, but the preliminary results are promising.
All of the changes and adjustments made to the protocol during the project were quite successful, and the more recent ones were mostly to accommodate the instrument; mostly related to the fact that the parameters of the instrument were changed to include a single dispensing needle as opposed to 2 or more channels for the dispensing of the reagents.
In conclusion, the CARESS PROTOCOL is a solid chemically validated analysis which does provide the expected result.
As for the prototype, this technology is in a less mature state. However, it has to be considered that that is an alpha prototype, therefore, by definition, a unit not ready for the market, because it is imperfect and does not respond to all of specific requirements and regulations yet. In 26 months, starting from scratch, it is not possible to have more than an alpha prototype (the machine has been manufactured in less than 1 year), and this is what expected by the research project. Maybe better results could be obtained in the calibration curves, but all the problems encountered can be all managed during the “reengineering and optimization” phase after the initial project. These are activities which perfectly fit with the PHC-12 Call. There is nothing in the “technology” that cannot be modified to make the machine a viable final product.
It is imperative that the results of the project, both for the CARESS protocol and the CARESS Instrument remain confidential, in order to increase the probability of success once both the instrument and protocol are brought to market. It is important for all of the partners to be able to have the benefit of having the ‘first mover’ advantage when this innovative product is brought to market.
List of Websites:
The CARESS WEBSITE is one of the most important assets for the dissemination of the project. During the design and test phase of the website one of the goals has been the definition of the key message to be communicated through the site and the way to communicate it. Particular attention was paid to the “receivers” (the end users), what they need to know about the project and how the message should be communicated.
The idea behind the banner used is to transmit the message that the Caress technology provides hope to cancer patients. In this final release, the logo, the soft green palette and the sprouting plant evoke life and growth. The layout of the webpage has been studied with the purpose of creating a screenshot that uniquely identifies the project. The link of the website is: http://caress-project.eu.