Single‑use (SU) systems have great potential for use in antibody‑drug conjugate (ADC) manufacturing. The use of organic solvents in the ADC process might, however, raise questions about potential leachables from the plastic and elastomeric materials of single‑use components. To address those concerns, we performed extractables studies on a disposable chromatography column housing, and two different disposable flow paths. The extractables studies were performed with two solvents commonly used in the ADC cytotoxin conjugation step: DMA and DMSO.

The studies were designed to ensure that conditions were exaggerated compared with existing ADC manufacturing processes. Extractable organic compounds and trace elements from the single‑use components were identified and semi‑quantitated with a complementary set of analytical techniques. The low levels of extractables found in this study support the use of ReadyToProcess™ columns, ÄKTA readyflux™ flow kits, and ÄKTA ready™ flow kits in ADC processes.

Introduction

ADCs are biotherapeutic molecules consisting of a cytotoxin coupled to a monoclonal antibody (mAb) by a linker. The target specificity of the mAb enables delivery of the toxic drug to cancer cells, while minimizing collateral damage to healthy cells. mAbs used in ADC production are typically manufactured according to traditional processes, including purification via protein A‑platform processes (1, 2). Before coupling the linker, the mAb needs to be transferred to a suitable solution. This solvent exchange is normally performed by an ultrafiltration/diafiltration (UF/DF) operation. After the linker coupling reaction, the next step is the conjugation reaction, which couples the cytotoxic drug. Figure 1 provides an example workflow for manufacturing an ADC from a bulk mAb product. Conjugation reactions are typically performed in a solvent containing either N,N‑dimethylacetamide (DMA) or dimethyl sulfoxide (DMSO) (3) at concentrations from 10% to 15%.

SU systems are well suited to ADC manufacturing for several reasons. Importantly, SU systems minimize operator exposure to toxins, while also protecting the product from the operator and environment. The high potency of ADCs means that relatively small amounts of products need to be produced per batch. The small batch sizes are well suited for incorporation of SU technology, which provides a cost‑efficient solution for multi‑product manufacturing. In addition to the lower upfront capital cost compared with reusable systems, SU technologies are quicker to start using because they are supplied ready to use. Cleaning and cleaning validation between manufacturing campaigns is unnecessary, and the risk of carryover of cytotoxin from one batch to another is minimized. Because cleaning is not performed, SU technologies minimize the volume of contaminated waste that must be handled and disposed of.

To support the adoption of SU technologies in ADC production, relevant extractables information is needed. Therefore, extractables studies were performed on three SU products: the ReadyToProcess™ 1 L column housing, disposable ÄKTA ready™ low flow kit for chromatographic systems, and ÄKTA readyflux™ flow kit for tangential flow filtration (Fig 2 and 3).

Fig 2. ÄKTA ready™ system with a 20 L ReadyToProcess™ column and ÄKTA ready™ flow kit. In this study, 1 L column housings were used.

Fig 3. ÄKTA readyflux™ tangential flow filtration system with flow kit.

Materials and methods

The goal of the extraction studies was to characterize extractables profiles with equipment and conditions relevant to current ADC manufacturing processes. The studies were designed and performed with advice from customers who use disposables in their ADC processes.

Extractables study design

The extractables studies were designed with consideration for test conditions representing a worst‑case scenario and for appropriate analytical techniques (3). Solvent concentrations used in typical cytotoxin conjugation reactions were exaggerated, as were surface area‑to‑volume ratios, temperatures, and contact times (Table 1). The experiments were set up to ensure contact with all wetted parts. Control samples of DMSO and DMA solution that had been stored at the same conditions but not in contact with the test article were included as blank references.

Table 1. Study design parameters compared with standard conditions

Extraction solutions

A 30% or 15% (v/v) solution of N, N‑dimethylacetamide (DMA) was prepared in ultrapure water at neutral pH. A 30% or 15% dimethyl sulfoxide (DMSO) solution was prepared the same way.

ÄKTA readyflux™ flow kit

Complete flow kits, manufactured with the standard method and gamma irradiated, were used. See Table 2 for a list of the materials in the wetted parts. Most wetted surface of the ÄKTA readyflux™ flow kit are made of TPE, silicone, and PP. Additionally, other materials are present in small components or in subassemblies. During extraction, additional silicone gaskets of the type recommended for this product were used to connect all open ends.

Table 2. Materials in wetted parts of ÄKTA readyflux™ flow kit

ReadyToProcess™ 1 L column

Column housings, assembled at the manufacturing site, were used. See Table 3 for a list of the materials in the wetted parts. Additional tubing needed for the experimental setup with the column was polytetrafluoroethylene (PTFE).

Table 3. Materials in wetted parts of ReadyToProcess™ columns

ÄKTA ready™ low flow kit

Complete flow kits, manufactured with the standard method, were used. This method includes gamma irradiation of the parts except for the pump tubing, which is autoclaved. See Table 4 for a list of the materials in the wetted parts. Additional EPDM gaskets and TC clamps were used for the experimental setup with flow kits to connect the column tubing and the six inlets with the outlets.

Table 4. Materials in wetted parts of ÄKTA ready™ flow kit

Process setup for ÄKTA readyflux™ flow kit

Two ÄKTA readyflux™ flow kits were tested, using the smallest size of the flow kit (1/4 in. flow kit TC) to ensure the highest possible surface area‑to‑volume ratio, representing a worst case. The wetted materials of construction are listed in Table 2. The ÄKTA readyflux™ flow kits were filled with extraction solvent (30% DMSO and 30% DMA in water) placed on a shaker (50 rpm) and then incubated at 30°C for 24 h. The conditions chosen targeted the highest levels of solvent that might be use during ADC manufacturing. All open ends were closed by connecting to each other using TC clamps and silicone gaskets. The vent filters were excluded from the extraction study by placing a metal clamp on the tubing prior to the filters (Fig 4). Control samples were made and incubated simultaneously using similar preparation devices, but without contact with the flow kit.

Fig 4. ÄKTA readyflux™ flow kit with connections to obtain a closed loop. The red circles mark the clamps used to exclude the vent filters.

Process setup for column housings

The test articles were two 1 L ReadyToProcess™ column housings (Fig 5). Because of the wide variety of resins that could be used, this study was limited to the column hardware. The smallest size ReadyToProcess™ column was selected to ensure the highest possible surface area‑to‑volume ratio, representing a worst case.

The wetted materials of construction are listed in Table 3. Tubing was connected to the inlet of each column. The other end of each piece of tubing was placed into a volumetric cylinder containing 1 L of either 15% DMA or 15% DMSO. The extraction solutions were pumped into the columns using a peristaltic pump. After filling, the tubing was removed, and the inlet and outlet tubes of the columns were clamped.

Fig 5. ReadyToProcess™ 1 L column. Column housings without chromatography resin were used for this study.

The filled columns were placed upright on an orbital shaker (19 mm orbit diameter) and incubated at 100 revolutions per minute (rpm) for 24 h at 30°C. After incubation, the clamps from the inlet and outlet tubes were removed. The extracts were transferred to separate bottles by applying nitrogen pressure.

The control samples were prepared by pumping 1 L of each solution through two pieces of tubing from a volumetric cylinder into a glass bottle, using a peristaltic pump. The bottles were placed on an orbital shaker alongside the filled columns and agitated at 100 rpm for 24 h at 30°C.

After incubation, extraction solutions from the test articles and control samples were collected for analysis, divided into separate containers for the different analytical techniques, and stored at 5°C.

Process setup for ÄKTA ready™ low flow kit

The test articles were two ÄKTA ready™ low flow kits. The smallest size flow kit was selected to ensure the highest possible surface area‑to‑volume ratio, representing a worst case. The wetted materials of construction are listed in Table 4. The open ends of the tubing of a low flow kit were connected to each other using EPDM gaskets and TC 25 clamps. The pump tubing of each flow kit was connected to a peristaltic pump, and the pump was used to fill the flow kit with 700 mL of either 15% DMA or 15% DMSO. During filling, the air trap of the flow kit was mounted at the highest position to let the air escape from the flow kit and to make sure that all surfaces were wetted with extraction solvent. Subsequently, the solution was circulated through the flow kit for 24 h at 30°C in an incubator (Fig 6).

Fig 6. Filling and circulation procedure of ÄKTA ready™ low flow kit in an incubator. The peristaltic pump is placed behind and connected to the pump tubing. The bottle with the control sample is seen to the right. Control samples were prepared by filling a glass bottle with 500 mL of either extraction solution and placed in the incubator at 30°C for 24 h. After incubation, extraction solutions from the test articles and control samples were collected for analysis, divided, and stored in the same manner as the solutions from the column housing and ÄKTA readyflux™ flow kit studies.

Analytical methods

Two classes of extractable compounds were analyzed: organic compounds and a spectrum of elements. Organic compounds were identified, and semi‑quantitative results obtained with liquid chromatography‑mass spectrometry (LC‑MS) and gas chromatography‑mass spectrometry (GC‑MS) methods. Test methods for volatile (VOC), semi‑volatile (SVOC), nonvolatile compounds (NVOC), and elements are listed in Table 5. Analyses of organic compounds were performed by Nelson Labs Europe in Leuven, Belgium. Elemental analysis was performed by ALS Scandinavia AB in Luleå, Sweden by inductively coupled plasma/sector field mass spectrometry (ICP‑SFMS).

Table 5. Overview of chemical analyses performed

* MS = mass spectrometry; GC = gas chromatography; HS = headspace; APCI = atmospheric pressure chemical ionization; ICP‑SFMS = inductively coupled plasma/sector field mass spectrometry; UPLC = ultra performance liquid chromatography

Sample preparation

Prior to GC‑MS and LC‑MS, liquid/liquid extraction was performed on samples of the test and control solutions to transfer organic compounds to a low boiling point organic solvent. Dichloromethane (DCM) was used as extraction solvent for the DMSO samples, while hexane was used as extraction solvent for the DMA samples because of the solubility of DMA in DCM. Liquid/liquid extractions were performed at three different pHs. The combined extracts of different pH were concentrated under nitrogen flow with a concentration factor of 10.

Differential peaks were determined and identified for VOC analysis. The concentration of detected SVOC was estimated for VOC, except that a different internal standard was used for calculations. The reporting limit was set at 50 µg/L.

NVOC analysis using LC-MS APCI, positive and negative modes

The liquid/liquid extraction sample prep described for GC‑MS was used also for LC‑MS. An internal standard (Tinuvin 327) was added to a sample of each test or control solution.

Separation was performed on a 3 × 100 mm 1.7 µm C18 column with a water:methanol gradient from 20% to 100% methanol. High resolution mass spectrometry (HRMS) detection was performed in alternating full scan polarity‑switching mode (positive and negative APCI, 100 to 1500 amu).

Differential analysis was performed with a software to find differences between the extract and control sample. For each differential peak, retention time, accurate mass of the molecular ion, and mass spectrum were matched against a database to allow identification. Identification level was assigned as: identified compound, most probable compound, tentatively identified compound, or unidentified.

The quantitation of a detected NVOC assigned as identified compound was performed with the compound‑specific relative response factor (RRF) available for identified compounds. For other compounds, the response was compared with the response of the internal standard. The reporting limit was set at 50 µg/L.

Elemental analysis

The analysis with ICP‑SFMS targeted 25 elements (aluminum, arsenic, barium, cadmium, calcium, chromium, cobalt, copper, iron, lead, lithium, magnesium, manganese, molybdenum, nickel, palladium, potassium, silicon, silver, strontium, sulfur, titanium, vanadium, zinc, and zirconium). Detection limit was in the range of 0.1 to 10 µg/L for all elements except for calcium (100 µg/L), iron (30 µg/L), magnesium (30 µg/L), potassium (100 µg/L), silicon (500 µg/L), sulfur (10 mg/L), and zinc (30 µg/L).

Results and discussion

ÄKTA readyflux™ flow kit

In the extract from the ÄKTA readyflux™ flow kit incubation, five organic compounds were found above reporting limit for each solvent (30% DMSO and 30% DMA). All compounds identities were assigned confirmed and the majority of them were related to antioxidants and silicone material. Some differences between the solvents could be seen, but the total amount of extractables found did not differ significantly. The compound with the highest abundance was a carboxylic acid, semi-quantified to 430 µg per ÄKTA readyflux™ flow kit.

For ICP-SFMS, the results were aligned between the solvents: four elements were above the reporting limit, the same elements in both solvents, with potassium having the highest abundance (3.8 mg/flow kit in 30% DMA).

ReadyToProcess™ column housings

No organic extractable compounds were found above the reporting limit with HS‑GC‑MS or GC‑MS in the extraction samples with 15% DMA or 15% DMSO. The results with LC‑MS showed one organic compound that was present in both 15% DMA and 15% DMSO. The extractable compound was assigned a confirmed identity level as an identified compound related to a curing agent used with elastomeric material. The concentration was estimated below 200 µg/L (ppb) in both extraction solvents (that is, less than 200 µg/ReadyToProcess™ column).

Analysis with ICP‑SFMS showed low levels of a few extractable elements. The most abundant were calcium (< 100 µg/L) and magnesium (< 25 µg/L), followed by zinc (< 10 µg/L) and barium (< 2 µg/L). The extractable elements were found at a similar level in both 15% DMA and 15% DMSO.

ÄKTA ready™ low flow kit

The results showed five organic extractable compounds above the reporting limit with HS‑GC‑MS, GC‑MS, and LC‑MS. Two of the compounds were found in 15% DMA, and three compounds were found in 15% DMSO. All five extractable compounds were assigned a confirmed identity level as identified compounds related to polyamide and silicone materials and one solvent. The highest abundant extractable compound was present at a concentration below 600 µg/L (that is, 410 µg/ ÄKTA ready™ low flow kit, because the extraction volume was 700 mL).

Analysis with ICP‑SFMS showed that the most abundant extractable element was silicon, which was present below 9 mg/L (ppm). Additionally, calcium, barium, zinc, and copper were found at lower levels. The extractable elements were found at a similar level in both 15% DMA and 15% DMSO.

Assessment of results

A general toxicity and safety evaluation of extractable compounds as a worst case was conducted. The evaluation can only be general, because the specific details regarding the route of administration, dosage level, or toxicity of the proposed drug compounds will differ between different ADCs.

Toxicological information and a derived risk index (RI) for eight out of the ten identified extractable compounds were listed in a reference containing compiled safety impact information (4). In that reference, risk indices were obtained by subjecting toxicological safety data such as no observed effect levels (NOELs), no observed adverse effect levels (NOAELs), lowest published toxic dose (TDLOs), and others to a systematic evaluation process using appropriate uncertainty factors. An RI value represents a daily intake value for life‑long intravenous administration. Two additional RI values were derived for the final identified extractable compounds from a reported NOAEL value and appropriate uncertainty factors.

Assumptions for the assessment were:

• Single‑use equipment included in this assessment comprise the column housing, the chromatography flow kit, and the tangential flow filtration flow kit. All extractables from these disposables end up as impurity in the ADC product.
• Batch size is 5 g, considered a small batch, representing a worst case.
• Dosage is 3.6 mg/kg given every three weeks (21 d cycle), that is, 3.6 mg/kg × 70 kg bodyweight (bw)/21 d = 12 mg/person/d.

Potential exposure to extractables was calculated from the highest result on extractables divided with the batch size of ADC multiplied with the daily dosage of ADC (Table 6).

Table 6. Calculations for the safety assessment of the eight extractable compounds (highest concentration chosen if reported more than once)

¹ÄKTA readyflux™ flow kit

²ReadyToProcess™ column

³ÄKTA ready™ flow kit

Assessment of the results according to the listed assumptions shows that the potential exposure to extractables is well below the RI for each extractable compound. Therefore, extractables from the ÄKTA readyflux™ flow kit, ReadyToProcess™ column housing, and ÄKTA ready™ flow kit should pose no safety concern for use in ADC manufacturing within the conditions of this study.

Conclusions

The low levels of extractables found in this study demonstrate the chemical compatibility of ÄKTA readyflux™ flow kits, ReadyToProcess™ columns, and ÄKTA ready™ flow kits with two organic solvents typically employed in ADC manufacturing processes: DMSO and DMA. Detailed results of these studies are available in the validation guides for these SU components to supplement existing data generated with aqueous solvents and ethanol. Along with other SU components and systems, the ÄKTA readyflux™ flow kits, ReadyToProcess™ columns, and ÄKTA ready™ flow kits offer a solution to some of the main challenges in ADC manufacturing.

Resources

Access validation guides via Cytiva Regulatory Support (registration required)

ÄKTA ready™ flow kits

ÄKTA readyflux™ flow kits

Learn more about ReadyToProcess™ columns