Viewpoint

pubs.acs.org/jmc

How Big Is Too Big for Cell Permeability?

Par̈ Matsson† and Jan Kihlberg*,‡
†
Department of Pharmacy, BMC, Uppsala University, Box 580, SE-751 23 Uppsala, Sweden
‡
Department of Chemistry, BMC, Uppsala University, Box 576, SE-751 23 Uppsala, Sweden

ABSTRACT: Understanding how to design cell permeable ligands for intracellular targets that have difficult binding sites, such
as protein−protein interactions, would open vast opportunities for drug discovery. Interestingly, libraries of cyclic peptides
displayed a steep drop-off in membrane permeability at molecular weights above 1000 Da and it appears likely that this cutoff
constitutes an upper size limit also for more druglike compounds. However, chemical space from 500 to 1000 Da remains
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

virtually unexplored and represents a vast opportunity for those prepared to venture into new territories of drug discovery.

T he human proteome is estimated to have 100 000 to (Figure 1a), i.e., at a MW of approximately 1000 Da. This
Downloaded via 119.160.102.45 on March 18, 2020 at 05:45:26 (UTC).

1 000 000 binary protein−protein interactions (PPIs),
which may constitute one of the most important sources of
novel targets for drug discovery. However, harnessing this
potential has been challenging, since PPIs and several other
nonclassical drug targets typically have large, featureless, and
sometimes flexible interfaces that are difficult to drug with small
molecules.1,2 In addition to the difficulties posed by the target
binding site, the majority of PPIs and nonclassical targets are
located inside cells, requiring drugs to cross cell membranes to
elicit their effect. Thus, even though biologics are often ideal for
difficult-to-drug targets, they are ruled out by their lack of cell
permeability, as well as by their low bioavailability on oral
administration. In order to unlock the potential of PPIs and
other intracellular targets for drug discovery, it is therefore of
utmost importance to understand how far chemical space that
contains cell permeable, druglike molecules extends. In this
issue of the Journal of Medicinal Chemistry, Cameron Pye et al.
report that carefully designed cyclic peptides display a sharp
reduction in passive cell permeability with increasing size.3
When put in the context of other recent studies, it appears
likely that this cutoff is applicable also to nonpeptidic and more
druglike compounds beyond traditional small molecule space.
Cameron Pye et al. evaluated the impact of molecular size
and lipophilicity on membrane permeability using libraries of
cyclic octa-, nona- and decapeptides in the 800−1200 Da size
range. Backbone amide bonds were completely N-methylated
to eliminate the possibility of intramolecular hydrogen bonding,
which could otherwise have influenced conformation prefer-
ences and thereby membrane permeability. As a consequence
of the synthetic procedure, all peptides contained one Tyr and
one Pro residue, while the remaining residues were restricted to
amino acids with natural or non-natural aliphatic side chains,
thereby reducing the influence of polar and charged groups.
The parallel artificial membrane permeability assay (PAMPA)
and a MDCK cell clone that expresses only low levels of P-
glycoprotein were used for determination of membrane
permeability to eliminate or reduce the influence of transporter
mediated efflux.
Interestingly, a sharp and unexpected reduction in passive
permeability was observed, beginning at molecular sizes just
over 800 Å3 and severely limiting permeability above 1000 Å3
permeability trend was also observed for a set of cyclic peptide
natural products that displayed conformational preferences but
had MWs in the same range as the investigated peptide
libraries. As pointed out by the authors, this size-dependent
permeability is not compatible with the traditional solubility−
diffusion based theories for cell permeability. Instead other
theories, for instance, approximating membrane diffusion with
diffusion through polymer networks, may be required to
provide mechanistic insight into the unexpected size depend-
ence. The study of Pye et al. also points to the importance of
keeping lipophilicity within a narrow window in order to
combine cell permeability and aqueous solubility at high MW.
This window is reduced as MW increases, putting an upper
limit on the size of cell-permeable compounds. A further
consequence of this narrow window is that exposed polar or
charged groups can be expected to be prohibitive for
permeability in this chemical space.
It is important to note that the sharp drop-off in permeability
(which becomes very pronounced at a MW > 1000 Da) agrees
very well with previous investigations that have highlighted that
almost no orally administered drugs and clinical candidates
exist at higher MWs than this (Figure 1b).4 It thus appears that
the results reported by Pye et al. are not limited to the
hydrophobic cyclic peptides studied by the authors but are
applicable across different classes of cell permeable and orally
administered drugs. As recently suggested,5,6 it also appears to
be essential for orally administered drugs with MW > 700 Da to
possess a certain flexibility in the molecular structure; this
allows them to adapt to their environment and thereby
combine aqueous solubility, cell permeability, and efficient
target binding.5,6 In the absence of such “chameleonic”
behavior, these important drug properties would be mutually
exclusive in this chemical space. A few orally available and cell-
permeable outliers do exist at even greater molecular sizes. Of
these, cyclosporin A has been studied most extensively and
appears as the master among molecular chameleons due to its
reversible formation of four intramolecular hydrogen bonds and
shielding of polarity by lipophilic side chains.5,6 In fact, as
Received: February 12, 2017
Published: February 24, 2017

© 2017 American Chemical Society 1662 DOI: 10.1021/acs.jmedchem.7b00237

J. Med. Chem. 2017, 60, 1662−1664
Journal of Medicinal Chemistry
Figure 1. Membrane permeability of macrocyclic peptides, large drugs, and clinical candidates. (a) Using carefully designed libraries of cyclic
peptides, Pye et al. demonstrate a sharp decrease in membrane diffusion at high molecular volume that is not accounted for by the traditional
solubility-diffusion model.3 (b) A recent review of high-MW drugs and clinical candidates revealed that orally bioavailable compounds are extremely
rare at MW > 1000 Da (red and gray symbols represent oral and nonoral compounds, respectively; shaded surfaces indicate extended regions of
chemical space in which cell permeability is possible).4
demonstrated by Pye et al., cyclosporin A outperforms the
similarly sized model cyclic peptides when it comes to
permeability (Figure 1a). Incorporation of chameleonic
behavior in the design could thus be essential for discovery
of “oversized” drugs that extend cell-permeable drug space
beyond the current borders at approximately 1000 Å3 or 1000
Da (Figure 2).3,4 Even more importantly, it may also be an
Figure 2. Size boundaries for membrane-permeable molecules.
Membrane permeability is typically limited when polar surface area
(PSA) exceeds 140 Å2.13 Compounds that can alternately expose or
shield polar functionality depending on the environment (“chame-
leonic molecules”) represent an important opportunity to venture
beyond the borders of traditional drug space. The concept is
exemplified by cyclosporin A, which can adopt both membrane-
permeable (low PSA) and water-soluble (high PSA) conformations.
This extends the boundaries of cell-permeable chemical space and also
appears to be essential for orally administered drugs with MW > 700
Da. TPSA: topological polar surface area.
effective means of improving the pharmacokinetic properties of
compounds in the 700−1000 Da MW range, which is a
challenging region of chemical space but one that still presents
significant opportunities for discovery of orally administered
drugs.
The use of artificial membranes and efflux-reduced MDCK
cells allowed Pye et al. to study the size dependency of simple
transmembrane diffusion, in isolation from other complicating
transport mechanisms. This includes transporter-mediated
efflux and uptake and particle trafficking pathways like endo-
and transcytosis. The impact of such facilitated mechanisms in
the cellular transport of larger druglike molecules is not yet
1663
Viewpoint
clear. Analyses of cell permeability in traditional small-molecule
drug space indicate that transporter-mediated efflux increases
with greater molecular size.7 In line with this, cellular efflux that
increased with molecular size was observed for macrocyclic
compounds in the 400−800 Da size range.8 Thus, efflux
mediated by ATP-binding cassette (ABC) transporters such as
P-glycoprotein (MDR1/ABCB1), breast cancer resistance
protein (BCRP/ABCG2), and members of the multidrug
resistance-associated protein family (MRP/ABCC) can impose
additional limitations on the cell permeability of high-MW
molecules, below the drop-off described by Pye et al. However,
transporter-mediated efflux has not been systematically ex-
plored in the size ranges discussed by Pye et al., with some
notorious exceptions like cyclosporin A.
Correspondingly, uptake transporters in the solute carrier
(SLC) superfamily can have pronounced impact on cellular
drug permeability, but this has not been systematically studied
for larger molecules. Available literature suggests that only
some of the known drug-transporting SLCs may accept
substrates near the size ranges discussed by Pye et al.5
Examples include the transport of larger HCV NS3/4A
protease inhibitors by members of the organic anion trans-
porting polypeptide (OATP/SLCO) family.9 However, the
SLC family is notably understudied,10 with the majority of
human SLCs referred to in less than 15 reports in the scientific
literature. Clearly, much work is needed in defining the
substrate specificities of these transporters and their potential
impact on drug disposition.
Endocytosis and transcytosis mechanisms have mostly been
studied for nanoparticulate drug delivery systems and biologics
but could also play a role for intermediate-size drug molecules.
For instance, a family of cell-penetrating peptides that for the
first time provide for highly efficient delivery of drugs to the
cytosol was recently described.11 Again, fundamental research is
needed, but given that most transcellular permeability
coefficients reported for nanodelivery systems12 are lower
than those observed by Pye et al. at MW of ∼1000 Da, it
appears that simple transmembrane diffusion may still be the
more effective pathway for compounds in the size-range studied
by Pye, et al, which are still, by comparison, rather small.
In conclusion, intracellular targets such as protein−protein
interactions constitute valuable new opportunities for drug
DOI: 10.1021/acs.jmedchem.7b00237
J. Med. Chem. 2017, 60, 1662−1664
Journal of Medicinal Chemistry Viewpoint

discovery but are often difficult to modulate using small
molecules and inaccessible for biologics that lack cell
permeability. It is therefore important to understand to what
extent such “difficult” intracellular targets can be modulated by
ligands outside traditional small molecule drug space and how
such ligands should be designed to successfully reach their
targets. A few recent studies have begun to provide a first level
of understanding of how such nonconventional drugs interact
with their targets,1,2 what the outer limits of cell permeability
are, and what structural features are important for their
design.4−6 Pye et al. have now taken another important step to
clarify the scope and limitations of drug discovery beyond the
rule of 5, but we still lack even a basic understanding of how to
design drugs in this space. Thus, we can only second the
conclusion of Pye et al. that mining of this chemical space for
therapeutically valuable compounds will be a challenge for the
coming decades.
Edwards, A. M.; Superti-Furga, G. A call for systematic research on
solute carriers. Cell 2015, 162, 478−487.
(11) Qian, Z.; Martyna, A.; Hard, R. L.; Wang, J.; Appiah-Kubi, G.;
Coss, C.; Phelps, M. A.; Rossman, J. S.; Pei, D. Discovery and
mechanism of highly efficient cyclic cell-penetrating peptides.
Biochemistry 2016, 55, 2601−2612.
(12) Lundquist, P.; Artursson, P. Oral absorption of peptides and
nanoparticles across the human intestine: Opportunities, limitations
and studies in human tissues. Adv. Drug Delivery Rev. 2016, 106, 256−
276.
(13) Palm, K.; Stenberg, P.; Luthman, K.; Artursson, P. Polar
molecular surface properties predict the intestinal absorption of drugs
in humans. Pharm. Res. 1997, 14, 568−571.

■ AUTHOR INFORMATION
Corresponding Author
*E-mail: jan.kihlberg@kemi.uu.se. Phone: +46 (0)18 4713801.
ORCID
Pär Matsson: 0000-0002-9094-2581
Jan Kihlberg: 0000-0002-4205-6040

■ REFERENCES
(1) Scott, D. E.; Bayly, A. R.; Abell, C.; Skidmore, J. Small molecules,
big targets: drug discovery faces the protein−protein interaction
challenge. Nat. Rev. Drug Discovery 2016, 15, 533−550.
(2) Doak, B. C.; Zheng, J.; Dobritzsch, D.; Kihlberg, J. How beyond
rule of 5 drugs and clinical candidates bind to their targets. J. Med.
Chem. 2016, 59, 2312−2327.
(3) Pye, C. R.; Hewitt, W. M.; Schwochert, J.; Haddad, T. D.;
Townsend, C. E.; Etienne, L.; Lao, Y.; Limberakis, C.; Furukawa, A.;
Mathiowetz, A. M.; Price, D. A.; Liras, S.; Lokey, R. S. Nonclassical size
dependence of permeation defines bounds for passive absorption of
large drug molecules. J. Med. Chem. 2017, DOI: 10.1021/acs.jmed-
chem.6b01483.
(4) Doak, B. C.; Over, B.; Giordanetto, F.; Kihlberg, J. Oral druggable
space beyond the rule of 5: Insights from drugs and clinical candidates.
Chem. Biol. 2014, 21, 1115−1142.
(5) Matsson, P.; Doak, B. C.; Over, B.; Kihlberg, J. Cell permeability
beyond the rule of 5. Adv. Drug Delivery Rev. 2016, 101, 42−61.
(6) Whitty, A.; Zhong, M.; Viarengo, L.; Beglov, D.; Hall, D. R.;
Vajda, S. Quantifying the chameleonic properties of macrocycles and
other high-molecular-weight drugs. Drug Discovery Today 2016, 21,
712−717.
(7) Desai, P. V.; Sawada, G. A.; Watson, I. A.; Raub, T. J. Integration
of in silico and in vitro tools for scaffold optimization during drug
discovery: predicting P-glycoprotein efflux. Mol. Pharmaceutics 2013,
10, 1249−1261.
(8) Over, B.; Matsson, P.; Tyrchan, C.; Artursson, P.; Doak, B. C.;
Foley, M. A.; Hilgendorf, C.; Johnston, S. E.; Lee, M. D., IV; Lewis, R.
J.; McCarren, P.; Muncipinto, G.; Norinder, U.; Perry, M. W. D.;
Duvall, J. R.; Kihlberg, J. Structural and conformational determinants
of macrocycle cell permeability. Nat. Chem. Biol. 2016, 12, 1065−1074.
(9) Summa, V.; Ludmerer, S. W.; McCauley, J. A.; Fandozzi, C.;
Burlein, C.; Claudio, G.; Coleman, P. J.; DiMuzio, J. M.; Ferrara, M.;
Di Filippo, M.; Gates, A. T.; Graham, D. J.; Harper, S.; Hazuda, D. J.;
McHale, C.; Monteagudo, E.; Pucci, V.; Rowley, M.; Rudd, M. T.;
Soriano, A.; Stahlhut, M. W.; Vacca, J. P.; Olsen, D. B.; Liverton, N. J.;
Carroll, S. S. MK-5172, a selective inhibitor of hepatitis C virus NS3/
4a protease with broad activity across genotypes and resistant variants.
Antimicrob. Agents Chemother. 2012, 56, 4161−4167.
(10) Cesar-Razquin, A.; Snijder, B.; Frappier-Brinton, T.; Isserlin, R.;
Gyimesi, G.; Bai, X.; Reithmeier, R. A.; Hepworth, D.; Hediger, M. A.;

1664 DOI: 10.1021/acs.jmedchem.7b00237

J. Med. Chem. 2017, 60, 1662−1664