ARTICLE IN PRESS
Cytotherapy, 2016; ■■: ■■–■■
Defining quality attributes to enable measurement assurance for
cell therapy products
SHENG LIN-GIBSON1, SUMONA SARKAR1 & YUZURU ITO2
1
Biosystems and Biomaterials Division, Material Measurement Laboratory, National Institute of Standards and
Technology, Gaithersburg, MD, USA, and 2Stem Cell Engineering Research Group (SCERG), Biotechnology Research
Institute for Drug Discovery, AIST, Tsukuba, Ibaraki, Japan
Cell therapy products (CTPs) represent a broad class
of therapeutics poised to move toward commercial-
ization due to recent discoveries, technological advances
and increased investment. The industry is racing to
develop, manufacture and translate promising thera-
peutics for treating a range of diseases and injuries but
must address significant manufacturing and regula-
tory challenges, including the development of robust
analytical methods and assays. Measurements for pro-
ducing high-confidence data underpin the decision-
making process from research and development (R&D)
to regulatory submissions, including understanding
and establishing critical quality attributes (CQA) for
candidate CTPs. In the United States, the Food and
Drug Administration defines CQA as “a physical,
chemical, biological, or microbiological property or
characteristic that should be within an appropriate
limit, range, or distribution to ensure the desired
product quality” [1]. Each CTP will likely have its own
set of CQAs based on the products’ intended use and/
or mechanism of action (MOA). CQAs are generally
considered to be quality attributes tied to the intend-
ed clinical response or MOA. Here, we focus on the
process for defining quality attributes and designing
appropriate measurements for a given cell prepara-
tion (during manufacturing or as a product), regardless
of its association with clinical response or MOA.
It is well recognized that measurements for quality
attributes are influenced by variability in the analyt-
ical method and variability in the cell preparation.
CTPs are composed of dynamic living entities whose
properties can fluctuate with time. CTPs also contain
a complex mixture of biological and non-biological ma-
terials that support the living cells, and variability may
come from changes in these components. Changes in
a processing parameter, sample handling procedure
Correspondence: Sheng Lin-Gibson, PhD, Biosystems and Biomaterials Division, Material Measurement Laboratory, National Institute of Standards and
Technology, Gaithersburg, MD, USA. E-mail: slgibson@nist.gov
ISSN 1465-3249 Copyright Published by Elsevier Inc. on behalf of International Society for Cellular Therapy.
http://dx.doi.org/10.1016/j.jcyt.2016.07.002
and reagent can also affect the quality attributes of
the CTP. Inability to separate actual changes in the
sample from variability in the analytical method hinders
the subsequent decision making in R&D and trans-
lation. As such, many are eager to develop industry
standards to reduce variability in analytical methods
as well as in manufacturing processes. Some have called
for the development of standard methods or refer-
ence materials, including reference cells, to enable
comparability. Efforts are underway to address tech-
nology challenges associated with developing a stable
and homogenous (sample-to-sample consistency) ref-
erence cell that is fit-for-purpose. Discussions for
documentary standards have focused on those that
address common, industry-wide (horizontal) chal-
lenges for analytical methods and manufacturing and
have avoided highly restrictive standards that may un-
intentionally hamper the development of this rapidly
evolving industry.
We propose that a generalized framework for de-
signing and conducting cell measurements is needed
to improve the overall measurement confidence. Com-
ponents of this framework include (i) clearly defined
quality attributes, (ii) well-designed measurements that
are fit-for-purpose, (iii) robust measurement pro-
cesses with built-in measurement assurance, and (iv)
appropriate documentation, reporting and commu-
nication. This commentary focuses on the necessary
first two steps of this framework. The measurement
process may include sample collection, sample prep-
aration, sampling, measurement or data collection and
data analysis. Strategies for improving the measure-
ment assurance, such as through process controls,
reference materials and design of experiments, were
examined at a 2015 National Institute of Standards
and Technology Workshop [2,3]. Documentation,
 ARTICLE IN PRESS
2 S. Lin-Gibson et al.
reporting, and communication of measurement results
and key meta-data are critical for interoperability, but
are beyond the scope here.
Establishing well-defined quality attributes
through clearly defined components
For a given cell preparation, one might ask, “What is
in the vial or container at this moment in time?” to
identify it or compare it with another cell prepara-
tion. Ideally, the full list of ingredients should be known
to properly answer this question. If the identity and
quantity of each ingredient is known and the func-
tion of each ingredient is well understood, it may be
possible to predict the performance of the product with
a high confidence. For a complex cell preparation, it
is generally not possible to list all ingredients. Instead,
CTPs are commonly described with a set of quality
attributes (e.g., identity, quantity, viability, purity, ste-
rility, stability, biological activity). Strict definitions for
quality attributes are impractical due to the unique-
ness of each CTP.
One strategy to define quality attributes for a
specific CTP is establishing clearly defined and
quantifiable components. In this process, we bin the
contents of a cell preparation into the following broad
categories: cells, potential impurities and suspension
medium (Figure 1), where each category can be further
Figure 1. Establishing clearly defined and quantifiable components and sub-components for a cell preparation at a given time.
divided into sub-categories. For example, potential
impurities can be divided into adventitious agents (e.g.,
bacterial and viral contaminants), biological impuri-
ties (e.g., cellular components), and non-biological
impurities (e.g., extractables, leachables and other
non-biological contaminants). Suspension medium is
a complex mixture that includes biologically and chem-
ically derived ancillary (raw) materials that could be
intentionally added to the preparation or present as
by-products from the manufacturing process. Simi-
larly, cells can be further divided according to
user-defined criteria. Each component can then be
examined individually and with respect to other com-
ponents, particularly when designing an appropriate
assay in which multiple components may contribute
to a measured response.
Ambiguities associated with applying criteria to a
heterogeneous cell population (illustrated as gradi-
ents in Figure 1) make the process for clearly defining
the cell components critical. A cell population can be
defined by its genomic profile, function, morpholo-
gy, viability and specific biomarker, for example. As
such, the quality attribute should be defined explic-
itly to reflect the specific cell population of interest,
particularly when a quality attribute has multiple
possible definitions. The user may have to establish
specifications for deciding whether a cell is or is not
a part of the defined population, particularly for
 ARTICLE IN PRESS
Defining quality attributes to enable measurement assurance for CTPs
population definitions that are expressed by a distri-
bution. It is important to note that while criteria for
defining cell populations are user defined, the ability
to distinguish these populations is dependent on the
quality of the measurement.
Design of measurements that are
fit-for-purpose
Well-defined quality attributes and a clear under-
standing of their intended uses guide the design of
measurements with sufficient selectivity, sensitivity and
resolution to enable subsequent decision making (fit-
for-purpose). When possible, a measurement output
(measurand) that directly assesses the quality attrib-
ute in question should be selected. It is important to
recognize that a measurand may in fact represent a
surrogate measure for assessing the quality attribute.
As with any good analytical method, the measure-
ment should have high selectivity for the measurand
without significant interference from other compo-
nents in the cell preparation.The measurement should
be sufficiently sensitive to detect the smallest rele-
vant quantity of the intended measurand and have
sufficient resolution to discern meaningful changes in
the quality attribute.
The intended use will guide the analytical require-
ments of the measurement. Applications such as routine
cell culture, in-process controls and product release
assays may have different analytical requirements with
regard to precision and accuracy, as well as different
practical considerations. A measurement with large
imprecision, for example, may not be able to resolve
quality attribute changes that are deemed significant
for a specific intended use. In this case, the measure-
ment method should be improved or a different
method selected to meet the needs of the applica-
tion. Measurement methods should also be robust such
that the results are not significantly affected by small
changes in the measurement process. Different ap-
plications may have different factors that will affect
robustness. Measurement assurance concepts may be
implemented to monitor measurements and mini-
mize sources of variability [2,3].
Viability as a clearly defined quality attribute
Here we consider an instance where cell viability is
identified as a quality attribute for a particular CTP.
Based on the framework presented here, we first define
viability as the ratio of user defined viable cell quan-
tity to total cell quantity, where both quantities must
be further defined and measured.
Total cell counting is conceptually straightforward.
Many methods can be used for total cell counting,
some of which count cells directly, whereby a count
3
is registered for each cell; others count cells indi-
rectly, where an observation is made for the entire cell
population and then extrapolated using a calibration
curve to derive cell quantity. The user must decide if
each cell should be counted, as in the case of direct
counting, and if so, whether cells should be defined
by the presence of each nucleus or as cellular objects.
The user may also decide that a single population-
based measure, such as the packed cell height, may
be more relevant for the intended use to represent
the total cell quantity. The selection of an appropri-
ate measurement method should be based on the
fit-for-purpose criteria, while considering potential con-
tributions from other components. For example,
potential biological impurities, such as micelles, lipo-
somes or beads may be miscounted as a cell depending
on the method selected.
User-defined viable cells (e.g., defined by ability
to replicate, metabolic activity, membrane integrity,
the expression of apoptotic markers, nuclear changes,
morphological changes) must be clarified to avoid am-
biguity. Each definition is based on a different biological
phenomenon and therefore cannot reasonably be ex-
pected to be quantitatively or qualitatively comparable.
Even among seemingly similar viability criteria, via-
bility results may not be comparable. For example,
membrane integrity based on trypan blue dye exclu-
sion may result in a different viability results than
membrane integrity based on fluorescence-based dye
exclusion because of factors such as the size of the dye
molecule, the mechanism of staining and morpho-
logical changes induced by the stains.
The selection of a cell counting method may be
further dictated by the composition of the cell prep-
aration and intended use. Direct cell counting methods
are generally more selective for specific cell popula-
tions than indirect methods. On the other hand, direct
cell counting methods may require cells to be well dis-
persed without large debris for optimal performance.
Indirect methods may be more useful for online moni-
toring or for cell aggregates.
Moving forward
The approach presented here facilities the process for
defining the quality attributes for CTPs by breaking
down a complex cell preparation into components
with clearly defined properties that can be confi-
dently measured. Such approaches are expected to
facilitate further discussions on the design of appro-
priate measurements. Critical measurements may then
be identified and further evaluated for their fitness-
for-purpose. Additional steps are needed to develop
robust protocols for measurement validation and
reporting/communication. Measurement assurance
tools, including appropriate statistical methods, should
 ARTICLE IN PRESS
4 S. Lin-Gibson et al.
be applied to CTP quality attributes to establish robust
characterization strategies and facilitate product com-
parability and translation.
The generalized framework for designing and con-
ducting cell measurements presented here may have
broader applications. The research community has an
urgent need for improved reproducibility. This larger
discussion has many commonalities with concepts de-
scribed here: the need for clearly defined attributes,
well-designed methods, a robust and validated mea-
surement process with controlled starting materials and
process controls and data interoperability through ap-
propriate and organized documentation of data and
metadata.
The cell therapy industry has repeatedly called for
the development of standards for measuring quality
attributes. Over the past 2 years, a group of interna-
tional experts has been working within ISO/TC 276:
Biotechnology to develop standards for cell counting,
an important component of cell measurement. A gap
analysis showed that a common understanding of
quality attributes and recommendations for design-
ing cell measurements are necessary to improve the
quality of cell measurements. Efforts are underway to
develop standards that address these gaps and needs.
We welcome comments and input from the broader
research community.
Acknowledgments
We thank members of ISO/TC 276: Biotechnology,
WG3: Analytical Methods, particularly members of
the Japan Mirror Committee for ISO/TC 276, for
valuable discussions. We also thank members of the
Office of Cellular,Tissue and Gene Therapies, CBER/
FDA, particularly Judith A. Arcidiacono, for their
insightful comments and input.
References
[1] U.S. Food and Drug Administration. Guidance for Industry;
Q8(R2) Pharmaceutical Development. <http://www.fda.gov/
downloads/Drugs/. . ./Guidances/ucm073507.pdf>; 2009.
[2] National Institute of Standards and Technology. NIST
workshop: strategies to achieve measurement assurance for cell
therapy products, <http://www.nist.gov/mml/bbd/biomaterials/
measurement-assurance-for-cell-therapy-products.cfm>; 2015
[updated July 8, 2015; accessed 12.01.15].
[3] Simon CG Jr, Lin-Gibson S, Elliott JT, Sarkar S, Plant AL.
Strategies for achieving measurement assurance for cell therapy
products. Stem Cells Transl Med 2016;5:705–8.