Upload
others
View
5
Download
0
Embed Size (px)
Citation preview
Thomson et al 1
Modification of the cyclopropyl moiety of abacavir provides insight into the structure activity
relationship between HLA-B*57:01 binding and T-cell activation.
Short title: HLA-B*57:01-restricted T-cell responses to abacavir
Authors: Paul J Thomson,1 Patricia T Illing,2 John Farrell,1 Mohammad Alhaidari,1 Catherine C Bell,1
Neil Berry,1 Paul M O’Neill,1 Anthony W Purcell2, Kevin B Park,1 Dean J Naisbitt1
1MRC Centre for Drug Safety Science, Dept Molecular & Clinical Pharmacology, University of
Liverpool, Liverpool, UK; 2Infection and Immunity Program, Monash Biomedicine Discovery Institute
and Dept of Biochemistry and Molecular Biology, Monash University, Clayton 3800, Victoria,
Australia
*Correspondence Professor Dean J. Naisbitt (The University of Liverpool, Liverpool, England
[Telephone, 0044 151 7945346; e-mail, [email protected]]).
Acknowledgements: The authors would like to thank the volunteers for agreeing to donate blood.
This work was funded by grants from the MRC (grant number MR/R009635/1). Core support was
received from the Medical Research Council Centre for Drug Safety Science (Grant MR/l006758).
AWP is supported by a Principal Research Fellowship from the Australian National Health and
Medical Research Council (NHMRC) and NHMRC Project Grant (APP1122099).
Conflict of Interest Statement: The authors declare no conflicts of interest.
Word count: 3497
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Thomson et al 2
Graphical abstract
Alteration of the HLA-B*57:01
peptide binding repertoire
Synthesis of structural variants of abacavir
AbacavirNo Clashes Clashes
Abacavir analoguesNo Clashes
CD8+ T-cell activation
No CD8+ T-cell activation
No alteration of the HLA-B*57:01
peptide binding repertoire
23
24
25
26
Thomson et al 3
Abstract
Background: Abacavir is associated with hypersensitivity reactions in individuals positive for the
HLA-B*57:01 allele. The drug binds within the peptide-binding groove of HLA-B*57:01 altering
peptides displayed on the cell surface. Presentation of these HLA-abacavir-peptide complexes to T-
cells is hypothesised to trigger a CD8+ T-cell response underpinning the hypersensitivity. Thus, the
aim of this study was to explore the relationship between the structure of abacavir with HLA-
B*57:01 binding and CD8+ T-cell activation.
Methods: Seventeen abacavir analogues were synthesised and cytokine secretion from
abacavir/abacavir analogue-responsive CD8+ T-cell clones was measured using IFN-γ ELIspot. In silico
docking studies were undertaken to assess the predicted binding poses of the abacavir analogues
within the HLA-B*57:01 peptide binding groove. In parallel, the effect of selected abacavir analogues
on the repertoire of self-peptides presented by cellular HLA-B*57:01 was characterised using mass
spectrometry.
Results: Abacavir and ten analogues stimulated CD8+ T-cell IFN-γ release. Molecular docking of
analogues that retained antiviral activity demonstrated a relationship between predicted HLA-
B*57:01 binding orientations and the ability to induce a T-cell response. Analogues that stimulated
T-cells displayed a perturbation of the natural peptides displayed by HLA-B*57:01. The antigen-
specific CD8+ T-cell response was dependent on the enantiomeric form of abacavir at both
cyclopropyl and cyclopentyl regions.
Conclusion: Alteration of the chemical constitution of abacavir generates analogues that retain a
degree of pharmacological activity, but have variable ability to activate T-cells. Modelling and
immunopeptidome analysis delineate how drug HLA-B*57:01 binding and peptide display by antigen
presenting cells relates to the activation of CD8+ T-cells.
Keywords: Drug hypersensitivity, HLA, human, mass spectrometry, T-cells.
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
Thomson et al 4
Key messages: (1) Structural modelling aids prediction of the impact of abacavir modifications on
the activation of CD8+ T cells, but cannot be used in isolation; (2) analogues that stimulated CD8+ T-
cells displayed a perturbation of natural peptides displayed by HLA-B*57:01; (3) enantiomeric forms
of abacavir predicted to adopt different HLA-B*57:01 binding conformations display divergent CD8+
response profiles.
52
53
54
55
56
57
Thomson et al 5
Introduction.
Abacavir ((1S,4R)-4-(2-amino-6-(cyclopropylamino)-9H-purin-9-yl)-2-cyclopentene-1-methanol) is a
nucleoside reverse transcriptase inhibitor, used in the treatment of HIV. Despite its potent antiviral
activity, abacavir is associated with severe hypersensitivity reactions that occur in 4-8% of patients
receiving therapy (1). A relationship has been demonstrated between hypersensitivity and carriage
of the MHC class I allele HLA-B*57:01 (2–4). The presence of this “risk allele” has a 100% negative
predictive value in patients while yielding a positive predictive value of 48-55% (5,6). Prospective
genotyping for HLA-B*57:01 and exclusion of patients positive for the allele from abacavir therapy is
a widely applied, cost-effective means of eradicating hypersensitivity reactions (5,7,8).
Proteasomal breakdown of intracellular proteins generates the immunopeptidome that binds to
HLA class I molecules prior to display at the cell surface. Under normal circumstances, the
constitutive self-peptides displayed by HLA-B*57:01 are ignored by CD8+ T-cells; however,
introduction of novel peptides (e.g. virus-derived peptides) stimulates CD8+ T-cell proliferative
responses and the secretion of cytokines and cytolytic molecules that target cells/tissues bearing the
immunogenic peptides. Abacavir binds to HLA-B*57:01 in a non-covalent manner, occupying the C-F
pocket region of the peptide binding groove which usually accommodates the C-terminal end of the
bound peptide ligand, and leads to a dramatic alteration of the HLA-B*57:01 immunopeptidome (9–
11). The prevalence of peptides bearing aromatic amino acids such as tryptophan and phenylalanine
at the PΩ position (the C-terminus of the peptide) is diminished in the presence of abacavir, whilst
those terminating in smaller aliphatic amino acids such as valine, leucine and isoleucine increases
(9,10). As these newly presented HLA-drug-self-peptide complexes are not observed in the absence
of abacavir it has been proposed that they activate the CD8+ T-cells involved in abacavir
hypersensitivity (9,10).
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
Thomson et al 6
Abacavir-responsive CD8+ T-cell clones can be generated from the peripheral blood of healthy, drug
naïve, HLA-B*57:01+ donors (12–16). Chessman et al. identified that antigen processing was a
prerequisite for presentation of abacavir-associated antigens and the activation of CD8 + T-cells, while
Bell et al. and Adam et al. additionally demonstrated activation via a direct interaction of the drug
with cell surface HLA-B*57:01 (12,14,15). More recent studies suggest that abacavir preferentially
activates memory T-cells through cross-reactivity with peptides of viral origin (6,16). It has also
recently been shown that HLA-B*57:01 transgenic mice can be used as a model for abacavir
reactivity but require CD4+ T-cell depletion prior to the administration of the drug. This leads to
induction of skin-homing abacavir-specific CD8+ T-cells that induce skin inflammation (17).
The cyclopentyl and cyclopurine groups of abacavir sit in the D and E pockets of HLA-B*57:01,
respectively. However, the cyclopropyl group extends into the F-pocket altering its shape and
chemistry, accounting for the preferred accommodation of smaller amino acid moieties (10). The
abacavir metabolite carbovir, which lacks the cycloproyl moiety, does not activate CD8+ T-cell
responses (14) and abolition of abacavir-specific T-cell responses can be achieved via modifications
to the 6-amino cyclopropyl group (13). Fifteen abacavir analogues were previously synthesised, of
which two (isopropyl amino and isopropyl methyl amino) did not activate abacavir-specific CD8+ T-
cell clones. Molecular docking experiments predicted that both compounds would bind to HLA-
B*57:01 in a manner yielding unfavourable steric clashes between their functional groups and an
abacavir stabilised HLA-B*57:01 binding peptide (HSITYLLPV). This suggested an inability of the
compounds to facilitate presentation of this peptide in the same conformation (13). Expanding on
this concept, we have synthesized seventeen analogues with modifications to the 6-amino
cyclopropyl moiety as chemical probes to explore the molecular interaction between abacavir and
HLA-B*57:01, the display of abacavir-dependent HLA-B*57:01 binding peptides on the surface of
antigen presenting cells, and CD8+ T-cell activation. Our studies demonstrate that abacavir analogues
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
Thomson et al 7
have a distinct impact on the repertoire of peptides presented to T-cells on the surface of antigen
presenting cells, while cloning in the presence of analogue-stimulated antigen presenting cells had
the potential to generate T-cell clones with specificity outside the range of abacavir responsive T-cell
clones.
108
109
110
111
112
Thomson et al 8
Materials and methods.
A description of abacavir analogue synthesis and the assessment of antiviral activity is available as
supplementary material. Four HLA-B*57:01+ donors were selected to generate abacavir and
analogue responsive clones. Due to the journals restricted word count, text describing cell culture
methods and T-cell phenotypic and functional assessments are provided as supplementary material.
In silico modelling and mass spectrometric methods described in detail in the supplementary
material were used assess predicted abacavir (analogue) HLA-B*57:01 binding and the repertoire of
peptides displayed by HLA-B*57:01 in the presence of abacavir analogues, respectively. For T-cell
assays, the Mann-Whitney test was used to compare control and test values.
113
114
115
116
117
118
119
120
121
122
123
Thomson et al 9
Results
Abacavir-responsive T-cell clones display cross-reactivity to a subset of analogues.
From a total of 2279 clones tested across four donors, 101 proliferated in the presence of abacavir at
the first testing stage (Control, 1665 ± 188.5 cpm: abacavir (35µM), 6940 ± 729 cpm; P< 0.0001,
Figure 1a). All abacavir-responsive clones were phenotyped as CD8+. An abacavir-specific increase in
IFN-γ secretion was detected with all clones, when compared with controls (Figure 1b: 4
representative clones).
Six abacavir-responsive CD8+ T-cell clones from the four donors were incubated with abacavir and all
seventeen analogues, and IFN- γ secretion was measured. Figure 1c compares the individual
analogue response of the six clones with the mean level of abacavir-specific IFN-γ secretion using the
same six clones. Analogues A, C, D, E, H, J, K, O and P all stimulated IFN-γ secretion in abacavir-
responsive clones. High levels of IFN- γ were released from abacavir-responsive clones in the
presence of analogues A, C, D, J and P, especially analogue J (methoxy azetidine) which activated T-
cells to a higher degree than abacavir. Analogue D (azetidine) the original analogue upon which
others were constructed, demonstrated a dose-dependent increase in IFN-γ secretion in all abacavir
clones. Interestingly, analogue H (azetidine fluoro) activated T-cells at the higher concentration of
50µM, with no IFN-γ secretion observed at lower concentrations. In contrast, analogues B, F, G, I, L,
M, N, Q and analogue 15 (from our previous study (13)) were completely deficient in T-cell reactivity
with abacavir-responsive clones.
Antiviral activity.
It was possible to separate the antiviral and T-cell stimulatory properties of abacavir through the
synthesis of chemical analogues (Supplementary Table 4). Analogues D and H displayed good
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
Thomson et al 10
antiviral activity at concentrations similar to abacavir. Analogues A, B, C, G, K, M and O retained
antiviral activity (i.e., inhibited viral replication at 10-60 µM), while analogue J inhibited viral
replication at 86 µM. All remaining analogues were inactive.
Abacavir analogues are predicted to bind HLA-B*57:01 in a manner distinct from abacavir.
A molecular docking protocol was used to assess how abacavir and its analogues interact within the
peptide binding groove of HLA-B*57:01. The software replicated the crystal structure of the
predicted binding pose of abacavir within the F-pocket of HLA-B*57:01 with peptide HSITYLLPV (13).
The binding orientations of the abacavir analogues D, G, H, M, O, P and Q were predicted, with
abacavir’s predicted binding conformation superimposed as a comparator. The cyclopurine group of
abacavir forms hydrogen bonds with amino acid residues Asp 114, Ser 116, Asn 77 and Ile 124
(Figure 2b); however, a diverse range of binding interactions were observed with the abacavir
analogues.
First, analogues D, O and P, which activated abacavir-responsive clones (Figure 2a), were predicted
to bind HLA-B*57:01 in a similar manner to abacavir with their functional groups protruding into the
F pocket (Figure 2c i, iii and iv), and no clashes between the drug and amino acids of the antigen
binding cleft or pep HSITYLLPV were observed. While no clashes were also apparent for T-cell
stimulatory analogue H (Figure 2c ii), the guanosine portion of the molecule was inverted when
bound to HLA-B*57:01 indicating a predicted binding pose distinct to abacavir still yielding an ability
to induce T-cell activation.
The predicted binding poses of analogues devoid of abacavir T-cell reactivity (G, M and Q) indicated
the presence of unfavourable steric clashes between the functional groups of the respective
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
Thomson et al 11
analogues and amino acid residues present either within HLA-B*57:01 or residues of the predicted
HLA binding peptide HSITYLLPV (Figure 2d). This suggests that the analogues would either not bind
to HLA-B*57:01 or binding would occur in a manner which would require further conformational
changes within the peptide binding groove. Specifically, analogues M and Q induced steric clashes
with the amino acid residues of the binding peptide HSITYLLPV (Figure 2d i and iii) and this was not
observed with analogues that activated abacavir-specific T-cells. It is possible that these analogues
bind HLA-B*57:01 altering the conformation of co-presented peptides or favouring the binding of
different peptides when compared with abacavir, resulting in a loss of abacavir T-cell cross-reactivity.
Activation of abacavir-specific CD8+ T-cell clones is dependent upon drug chirality.
Analogues P (S-sec-butyl amino) and Q (R-sec-butyl amino) comprise the same structure, differing only
in the chirality of their functional group, yet exhibit different effects on T-cell activation (Figure 2a).
Analogue P stimulated IFN-γ release from the abacavir-responsive clones, whereas analogue Q did
not activate the clones (Figure 2a). This indicates that the enantiomeric state of a functional group is
a governing factor at the HLA peptide binding site. Indeed, when docked into HLA-B*57:01, the
predicted binding poses of analogues differ greatly. Analogue P is predicted to bind in a manner
conformationally similar to abacavir likely favouring the presentation of an array of altered self-
peptides similar to abacavir (Figure 2c iv), whilst analogue Q is predicted to bind in a more distinct
fashion (Figure 2d iii).
Given the enantiomer-specific activation of T-cell clones observed with analogues P and Q, we also
assessed whether chiral modification at the cyclopentyl group, alters HLA-B*57:01 binding and T-cell
activation. Abacavir-responsive clones were not stimulated to proliferate or secrete IFN-γ in the
presence of the alternative 1R,4S enantiomeric form of abacavir (Figure 3a and b). Incorporation of
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
Thomson et al 12
the 1R,4S enantiomer into our molecular docking protocol revealed a binding pose distinct from
abacavir (Figure 3c). This was characterised by a loss of interactions between the cyclopurine moiety
of the compound and amino acid residues Asp 114, Ser 116 and Ile 124. Furthermore, the
cyclopropyl moiety no longer protruded into the F-pocket. This would likely lead to a reduced
stability of the compound within HLA-B*57:01 and a loss of T-cell reactivity.
Abacavir analogues have distinct impacts on the HLA-B*57:01 immunopeptidome.
Analogues D (activates abacavir-responsive T-cells), H (activates abacavir-responsive T-cells at high
concentrations) and M (does not activate abacavir-responsive T-cells) were selected alongside
analogue 15 (from our previous study (13); does not activate abacavir-responsive T-cells) to explore
perturbation of the peptides displayed by HLA-B*57:01. Untreated and abacavir-treated cells were
used as controls. Peptides identified at a 5% FDR in the described dataset were used to define the
global peptide binding motif. HLA-B*57:01 peptide ligands were predominantly 9-11 amino acids in
length (Figure 5a i). 9mer ligands were biased towards Ser, Thr>Ala>Val at position 2 (P2) and
aromatic residues (Trp>Phe>Tyr) at the C-terminus (PΩ) whilst abacavir induced presentation of an
increased proportion of Ile (and Val) terminating peptides (Figure 4a ii, iii) (10). Similar changes were
induced for 10 and 11mer peptides (data not shown). This global alteration of the peptide-binding
motif was not recapitulated by either the cross-reactive or non-cross-reactive analogues.
To compare the precise peptide sequences presented across the conditions, peptides assigned with
a confidence above a 5% FDR in at least one data set were used to validate assignments below this
threshold in other data sets (Supplementary Table 3). Removal of peptides of the constitutive
repertoire (observed in untreated samples, or in our previous study (18) revealed 1065 potential
neo-epitopes induced by treatment with abacavir (778) or the analogues (analogue 15-277,
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
Thomson et al 13
analogue D-110, analogue H-83 and analogue M-150), with varying degrees of overlap (Figure 4d-g,
Supplementary table 3). Due to low peptide numbers, anchor site biases of 9-12mer peptides were
visualised collectively by aligning the first 3 (P1 to 3) and last 3 (PΩ-2 to PΩ) residues. Whilst anchor
residue enrichment in the filtered data set for analogue M (Figure 4g iii) resembled both the
unfiltered data set (Figure 4g iii) and the constitutive repertoire (Figure 4b), enrichment of PΩ Ile and
Met was observed in the filtered data sets for analogues D and H (Figure 4e and f iii). Most of these
Ile and Met terminating peptides were also present in the abacavir induced repertoire
(Supplementary Table 3), in which Ile, Leu and Met enrichment is evident at PΩ (Figure 4c). As for
abacavir, loss of Arg enrichment at the PΩ-2 secondary anchor site (19) on filtering for constitutive
ligands was also observed. These observations are consistent with the cross-recognition of
analogues D and H by abacavir primed T-cell clones, suggesting they may enable the formation of
similar HLA-drug-peptide complexes. The failure to observe a more global effect on the
immunopeptidome suggest they do so with reduced potency compared to abacavir. Strikingly, for
analogue 15 Ile enrichment was not as evident at PΩ in the filtered data set, instead enrichment of
Met and Ala was observed (Figure 4d iii). Only approximately half of these peptides were in the
overlap region with abacavir (Supplementary Table 3), suggesting analogue 15 may have a more
distinct impact on peptide binding and presentation than analogues D and H, explaining the lack of
cross-recognition of analogue 15 by the 6 abacavir T cell clones tested.
CD8+ T-cell clones can be generated to abacavir analogues, capable of cross-reacting with abacavir.
Given the distinct subset of analogue 15 induced HLA-B*57:01 ligands and its inability to activate the
abacavir-responsive T-cell clones tested, we sought to assess whether CD8+ T-cell clones could be
generated from two HLA-B*57:01+ donors after culturing their antigen presenting cells with the four
analogues (15 [does not activate abacavir clones, but displays distinct binding peptides], G [does not
activate abacavir clones], H [activates abacavir clones at high concentrations] and J [activates
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
Thomson et al 14
abacavir clones]) with PBMC. On initial testing, 24/514, and 95/282 analogue-responsive clones
were detected with analogue H and J, respectively (Supplementary figure 1; Figure 5a). Analogue H
and J dose-response studies revealed that clones secreted IFN-γ in a concentration-dependent
manner (Figure 5b). Analogue H- and J-responsive clones were also activated with abacavir (Figure
5c). All clones were phenotyped as CD8+.
Analogue 15-responsive CD8+ T-cell clones were generated in large numbers (99/401, numerous
producing a stimulation index of fifty or above) when PBMC from HLA-B*57:01+ donors were
cultured for 2 weeks the compound (Figure 5 a and b; Supplementary Figure 2). In contrast to
abacavir-responsive clones, these clones were stimulated to secrete IFN-γ in the presence of
analogue 15 and abacavir (Figure 5b and c).
None of the 568 clones generated from PBMC cultured with analogue G were activated with this
compound (Figure 5a).
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
Thomson et al 15
Discussion
X-ray crystal structures of HLA-B*57:01-abacavir-peptide complexes show that abacavir binds
directly to HLA-B*57:01, deep within the peptide binding cleft, underneath the C-terminal portion of
the peptide (9,10). In agreement with this, previous analyses of the immunopeptidomes of abacavir-
treated cells expressing HLA-B*57:01 revealed a change in the peptide repertoire (9–11), with a
decrease in the prevalence of peptides containing larger amino acids such as tryptophan and
phenylalanine at the PΩ position accompanied by an increase in peptides with smaller amino acids
such as valine, leucine and isoleucine at this location. Thus, the abacavir HLA-B*57:01 interaction
creates an altered chemical space for peptide binding and it has been proposed that the novel self-
peptide-abacavir-HLA-B*57:01 complexes are responsible for activating the T-cells involved in
abacavir hypersensitivity. Therefore, the objective of this study was to utilise synthetic analogues of
abacavir with structural substitutions in place of the 6-amino cyclopropyl moiety to probe the drug
HLA-B*57:01 binding interaction and to delineate how alteration of this functional group impacts
peptide presentation and T cell activation. The first fourteen analogues were synthesised around the
azetidine ring, a stable structure which allowed for the easy synthesis of further analogues with
additional functional groups. The final three analogues were structural variants of the amino group,
yielded by opening the azetidine ring structure. Analogues P and Q were of particular interest as
they allowed us to probe HLA binding and T-cell activation with different enantiomeric forms of the
same compound.
As described in our previous study (13), through structural modification several compounds that
retained a degree of pharmacological activity, but did not stimulate abacavir-responsive T-cells, were
identified (i.e., analogues B, G, and M; analogue H only activated clones at high concentrations).
Replacement of the cyclopropyl group with an azetidine ring per se yielded an analogue (D) that
displayed equipotent antiviral activity to abacavir, but this compound activated T-cells. However, the
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
Thomson et al 16
addition of branched carbon chains at the 3-position of the azetidine ring produced abacavir
analogues that did not activate cloned T-cells. This suggests the 3-position of the azetidine ring plays
a pivotal role in the formation of immunogenic drug HLA-B*57:01 peptide complexes. As there was
no simple chemical explanation for T-cell response profiles observed with the azetidine ring
containing analogues, molecular modelling and mass spectrometry-based assessment of HLA-
B*57:01 bound peptides were performed.
Analogue D activated abacavir-responsive T-cells and unsurprisingly was predicted to bind to HLA-
B*57:01 in a similar conformation to abacavir. Co-incubation of C1R-B*57:01 cells with analogue D
yielded an enrichment of C-terminal isoleucine in HLA-B*57:01 bound peptides, correlating with the
observed cross-reactivity with of abacavir-responsive T-cell clones. Abacavir-responsive clones were
also activated with analogue H, albeit only at high concentrations. The predicted binding pose of
analogue H within HLA-B*57:01 was distinct to abacavir and analogue D, with the guanosine portion
of the analogue inverted. Although C-terminal isoleucine enrichment was observed on HLA-B*57:01
binding peptides when C1R-B*57:01 cells were incubated with analogue H, a lower proportion of the
new peptide sequences overlapped with those observed with abacavir-treated cells. Unfavourable
binding interactions were predicted with the analogues G and M and the HLA-B*57:01 binding
peptide HSITYLLPV. Importantly, neither of these analogues activated abacavir-responsive clones.
Due to stock limitations, peptide repertoire studies were not performed with analogue G. However,
the peptides displayed by HLA-B*57:01 when C1R-B*57:01 were incubated with analogue M
mirrored the peptides constitutively expressed by HLA-B*57:01 suggesting no perturbation of the
immunopeptidome.
Analogue 15 from our previous study retained antiviral activity. It was predicted to bind in an
unfavourable conformation with HLA-B*57:01 and did not activate abacavir-responsive T-cells (13).
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
Thomson et al 17
Thus, the impact of analogue 15 on the HLA-B*57:01 peptide repertoire was studied. Minor
perturbation of the HLA-B*57:01 peptide repertoire displayed by C1R-B*57:01 cells was evident.
Unlike abacavir, which induces an increase in HLA-B*57:01 peptides terminating in isoleucine (9–11),
analogue 15 promoted an enrichment of alanine at the C-terminus. Given alanine terminating
peptides form a much smaller part of the observed abacavir-induced HLA-B*57:01 peptide
repertoire, it is not surprising that analogue 15 failed to activate abacavir-responsive T-cells.
However, the data raise the possibility that certain analogues contained within this and our previous
series of compounds (13) might activate alternative CD8+ T-cells to abacavir. To address this, PBMC
were cultured with 4 analogues (15 [does not activate abacavir clones, but displays distinct binding
peptides], G [does not activate abacavir clones], H [activates abacavir clones at high concentrations]
and J [activates abacavir clones]) prior to generation of T-cell clones and assessment of antigen
specificity. Clones responsive towards analogues 15, H and J were detected. Exposure of the clones
to antigen presenting cells and the analogue or abacavir promoted IFN-γ release. In contrast, the
clones generated from PBMC cultured with analogue G were not activated with analogue G or
abacavir. Collectively, these data show that although modelling experiments provide an effective
system to screen candidate compound HLA allele binding, it must be coupled with functional
analyses of compound-specific T-cell responses to predict immunogenicity.
The concept of enantiomeric-specific T-cell reactivity has seldom been explored in the context of
drug hypersensitivity. Thus, the different enantiomeric forms of butyl amino abacavir (analogues P
and Q) were synthesised. Enantiomers have identical chemical and physical properties, but are
mirror images that are non-superimposable. For this reason, enantiomeric forms of drugs may bind
differently to biological targets and display divergent pharmacological and toxicological properties.
Analogue P was predicted to bind to HLA-B*57:01 in an analogous manner to abacavir, while
analogue Q induced steric clashes with HLA-B*57:01 and the binding peptide. In agreement with the
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
Thomson et al 18
model, analogue P activated abacavir-responsive T-cells, while analogue Q did not. Abacavir also
exists in different enantiomeric forms at the cyclopentene ring. The pharmacological enantiomer has
a 1S, 4R confirmation. The alternative enantiomer, which is pharmacologically inactive, has a 1R, 4S
confirmation. This enantiomer displayed a predicted binding conformation distinct to abacavir
where the cyclopropyl group no longer protruded into the F-pocket and this led to a complete loss of
T-cell reactivity.
In conclusion, the coupling of modelling and analytical methods for functional analysis of compound-
specific human CD8+ T-cell responses has enhanced our understanding of how drugs interact with
HLA proteins to stimulate T-cells that participate in serious hypersensitivity reactions. The approach
offers an opportunity for safety scientists working in drug development to screen the potential
immunogenicity of second line compounds when HLA allele-restricted forms of drug hypersensitivity
are detected in the clinic.
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
Thomson et al 19
Author contributions: PJT, JF, CB and MA conducted the biological experiments. PI conducted the
peptide elution and immunopeptidome analyses. NB conducted the HLA-B*57:01 modelling
experiments. PON synthesized the abacavir analogues. DJN, BKP, AWP and PON designed the study
and supervised the project. PJT, PI, DJN and NB analysed. PJT, PI and DJN drafted the manuscript. All
authors critically reviewed the manuscript.
Abbreviations
Epstein-Barr Virus, EBV; reversed phase high performance liquid chromatography, HPLC; liquid
chromatography-tandem mass spectrometry, LC-MS/MS; stimulation index, SI; PBMC, peripheral
blood mononuclear cells.
356
357
358
359
360
361
362
363
364
365
366
Thomson et al 20
Figure legends
Figure 1. Activation of abacavir-responsive CD8+ T-cell clones with 6-amino cyclopropyl group
substituted analogues. a) Proliferative response and b) IFN-γ secretion of T-cell clones incubated
with abacavir (35µM). Data shown as mean ± SEM of all responsive clones vs control. * P<0.05, **
P<0.01, *** P<0.001. c) IFN-γ secretion from abacavir-responsive T-cell clones incubated with
analogues A-Q at concentrations of 0, 10, 20 and 50µM. Each analogue was tested against six
abacavir clones from four HLA-B*57:01+, healthy drug-naïve donors. Mean abacavir IFN-γ release
from the 6 clones (dashed line) was used as a comparator. Analogues K-Q were performed on an
independent occasion from A-N, accounting for the differences in mean abacavir value.
Figure 2. Direct comparison of the CD8+ T-cell activity of abacavir substituted analogues and the
binding orientations within the F-pocket of HLA-B*57:01. a) Representative ELIspot images from
wells containing abacavir and analogues D, G, H, M, O, P and Q at concentrations of 0, 10, 20 and
50µM. b) Crystal structure binding orientation of abacavir c) Crystal structure binding orientation of
analogues capable of activating abacavir-responsive T-cell clones i) D (azetidine), ii) H (azetidine-
fluoro), iii) O (isobutylamino) and iv) P (S-sec-butyl amino) within the F-pocket of HLA-B*57:01. d)
Crystal structure binding orientation of abacavir analogues with no cross-reactivity with abacavir
responsive T-cell clones i) G (azetidine-3-carbonitrile), ii) M (azetidine-trifluoromethyl), iv) Q (R- sec-
butyl amino) within the F-pocket of HLA-B*57:01. Crystal structure of HLA-B*57:01 (PDB:3UPR). Stick
representation of the peptide, (HSITYLLPV) shown in red. Key amino acid residues are shown as
yellow sticks. All non-polar hydrogen atoms removed. Key hydrogen bond interactions shown as
black dashes. Spheres used to illustrate the atomic radii of the atoms in the functional groups of
abacavir and substituted analogues, peptide (HSITYLLPV) and HLA-B*57:01 amino acids. Abacavir
structure superimposed over analogues shown as blue lines.
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
Thomson et al 21
Figure 3. Activation of abacavir-responsive CD8+ T-cell clones is enantiomer specific. a) Proliferation
of abacavir responsive T-cell clones cultured in the presence of abacavir and 1R, 4S abacavir
enantiomer at concentrations of 1, 5, 10, 50 and 100µM for 48 hours. b) Representative ELIspot
images from wells containing abacavir and 1R, 4S abacavir both at 50µM. c) Docking solution of
abacavir enantiomer (1R,4S) with abacavir superimposed, shown as blue lines in the F-pocket of
HLA-B*57:01. Stick representation of the peptide, (HSITYLLPV) shown in red. Amino acid protein
residues shown as yellow sticks, with key hydrogen bond interactions shown as black dashes. All
non-polar hydrogen atoms removed. Spheres used to illustrate the atomic radii of the atoms in the
functional group of the compounds.
Figure 4 a) Broad perturbation of C-terminal anchor preference is observed in the presence
abacavir but not analogues 15, D, H or M; however, potential neo-epitopes show a range of
overlap with abacavir induced ligands. a) Length (i) and primary anchor characteristics (ii. Position 2,
iii. C-terminal) of HLA-B*57:01 ligands isolated from CIR.B*57:01 grown in the absence of drug
treatment, or in the presence of 35μM abacavir (black) or analogues 15 (red), D (green), H (orange)
or M (blue). Analyses are based on non-redundant peptide identifications (by sequence,
modifications not considered) per data set made at a confidence greater than that for a 5% false
discovery rate (FDR) and filtered for ligands of endogenous HLA molecules of parental CIR cells.
Anchor residue preferences are shown for 9mers and are depicted as the proportion of peptides that
possess specific amino acids at position 2 (ii) and the C-terminus (iii). Data shown is the mean ± SD of
triplicate experiments for untreated and abacavir treated cells, duplicate experiments for analogue
15, and single experiments for the remaining analogues (due to availability of compounds). b)
iceLogos for P1 to P3 and PΩ-2 to PΩ of 9-12mer peptides in the constitutive repertoire of HLA-
B*57:01. c) iceLogos for 9-12mer HLA-B*57:01 ligands detected in the presence of abacavir in this
study either i) unfiltered or ii) filtered for constitutive ligands to identify potential neo-epitopes. d),
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
Thomson et al 22
e), f) and g) show i) icelogos for 9-12mer HLA-B*57:01 ligands detected in the presence of analogues
15, D, H and M, ii) Venn diagrams showing the numbers of potential neo-epitopes identified in both
abacavir and analogue treatments and iii. iceLogos for 9-12mer neo-epitopes identified. iceLogos
were generated using icelogo software utilising the human swiss-prot proteome as the reference
set. Letter height corresponds to % difference in frequency of the amino acid compared to presence
in the human proteome.
Figure 5. Generation of CD8+ T-cell clones to 6-amino substituted abacavir analogues. a) Antigen
specificity of T-cell clones generated to analogues G, H, 15 and J by way of cellular proliferation.
Clones yielding an SI<2 were considered positive. b) Representative ELIspot images from wells
containing clones from two HLA-B*57:01+ donors incubated in the presence of analogues H, 15 or J
(5-20µM). c) Representative ELIspot images from wells containing clones from two HLA-B*57:01+
donors incubated in the presence of abacavir and analogues H, 15 or J (35µM).
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
Thomson et al 23
References
1. Hetherington S, McGuirk S, Powell G, Cutrell A, Naderer O, Spreen B, et al. Hypersensitivity
reactions during therapy with the nucleoside reverse transcriptase inhibitor abacavir. Clin
Ther. 2001; 23:1603–14.
2. Mallal S, Nolan D, Witt C, Masel G, Martin AM, Moore C, et al. Association between presence
of HLA-B*5701, HLA-DR7, and HLA-DQ3 and hypersensitivity to HIV-1 reverse-transcriptase
inhibitor abacavir. Lancet. 2002; 359:727–32.
3. Hetherington S, Hughes AR, Mosteller M, Shortino D, Baker KL, Spreen W, et al. Genetic
variations in HLA-B region and hypersensitivity reactions to abacavir. Lancet. 2002; 359:1121–
2.
4. Martin AM, Nolan D, Gaudieri S, Almeida CA, Nolan R, James I, et al. Predisposition to
abacavir hypersensitivity conferred by HLA-B * 5701 and a haplotypic Hsp70-Hom variant.
Proc Natl Acad Sci U S A. 2004; 101:4180–5.
5. Mallal S, Phillips E, Carosi G, Molina J-M, Workman C, Tomažič J, et al. HLA-B*5701 Screening
for Hypersensitivity to Abacavir. N Engl J Med. 2008; 358:568–79.
6. Lucas A, Lucas M, Strhyn A, Keane NM, McKinnon E, Pavlos R, et al. Abacavir-reactive memory
T cells are present in drug naïve individuals. PLoS One. 2015; 10:e0117160.
7. Hughes DA, Vilar FJ, Ward CC, Alfirevic A, Park BK, Pirmohamed M. Cost-effectiveness analysis
of HLA B * 5701 genotyping in preventing abacavir hypersensitivity. Pharmacogenomics.
2004; 14:335–42.
8. Schackman BR, Scott C a, Walensky RP, Losina E, Freedberg K a, Sax PE. The cost-effectiveness
of HLA-B*5701 genetic screening to guide initial antiretroviral therapy for HIV. AIDS. 2008;
22:2025–33.
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
Thomson et al 24
9. Ostrov D a, Grant BJ, Pompeu Y a, Sidney J, Harndahl M, Southwood S, et al. Drug
hypersensitivity caused by alteration of the MHC-presented self-peptide repertoire. Proc Natl
Acad Sci U S A. 2012; 109:9959–64.
10. Illing PT, Vivian JP, Dudek NL, Kostenko L, Chen Z, Bharadwaj M, et al. Immune self-reactivity
triggered by drug-modified HLA-peptide repertoire. Nature. 2012; 486(7404):554–8.
11. Norcross MA, Luo S, Lu L, Boyne MT, Gomarteli M, Rennelsc AD, et al. Abacavir induces
loading of novel self-peptides into HLA-BM 57:01: an autoimmune model for HLA-associated
drug hypersensitivity. AIDS. 2012; 18:1199–216.
12. Bell CC, Faulkner L, Martinsson K, Farrell J, Al A, Tugwood J, et al. T-Cells from HLA-B*57:01+
Human Subjects Are Activated with Abacavir through Two Independent Pathways and Induce
Cell Death by Multiple Mechanisms. Chem Res Toxicol. 2013; 26:759–66.
13. Naisbitt DJ, Yang EL, Alhaidari M, Berry NG, Lawrenson AS, Farrell J, et al. Towards
depersonalized abacavir therapy : chemical modification eliminates HLA-B M 57 : 01-
restricted CD8 R T-cell activation. AIDS. 2015; 29:2385–95.
14. Chessman D, Kostenko L, Lethborg T, Purcell AW, Williamson N a, Chen Z, et al. Human
leukocyte antigen class I-restricted activation of CD8+ T cells provides the immunogenetic
basis of a systemic drug hypersensitivity. Immunity. 2008; 28:822–32.
15. Adam J, Eriksson KK, Schnyder B, Fontana S, Pichler WJ, Yerly D. Avidity determines T-cell
reactivity in abacavir hypersensitivity. Eur J Immunol. 2012; 42:1706–16.
16. Yerly D, Pompeu YA, Schutte RJ, Eriksson KK, Strhyn A, Bracey AW, et al. Structural Elements
Recognized by Abacavir-Induced T Cells. Int J Mol Sci. 2017; 18(1464):1–10.
17. Cardone M, Garcia K, Tilahun ME, Boyd LF, Gebreyohannes S, Yano M, et al. A transgenic
mouse model for HLA-B*57:01-linked abacavir drug tolerance and reactivity. J Clin Invest.
2018; 128:2819–32.
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
Thomson et al 25
18. Pymm P, Illing PT, Ramarathinam SH, O’Connor GM, Hughes VA, Hitchen C, et al. MHC-I
peptides get out of the groove and enable a novel mechanism of HIV-1 escape. Nat Struct
Mol Biol. 2017;24:387–94.
19. Illing PT, Pymm P, Croft NP, Hilton HG, Jojic V, Han AS, et al. HLA-B57 micropolymorphism
defines the sequence and conformational breadth of the immunopeptidome. Nat Commun.
2018;9(1).
480
481
482
483
484
485
486
Thomson et al 26
Supplementary methods
Culture media.
Cell culture medium for T-cells (R9) is composed of RPMI supplemented with 10% human AB serum,
HEPES (25mM), penicillin (1000 U/mL), streptomycin (0.1mg/mL), L-glutamine (2mM) and transferrin
(25µg/mL). Epstein-barr virus (EBV) -transformed B-lymphoblastoid cell lines were cultured in F1
medium composed of RPMI supplemented with 10% foetal bovine serum, HEPES (25mM), penicillin
(1000 U/mL), streptomycin (0.1mg/mL) and L-glutamine (2mM).
Synthesis of 6-amino substituted abacavir analogues.
The initial stage involved the synthesis of a stock of intermediate 1. From this chiral intermediate the
synthesis of seventeen target molecules (A-Q) was conducted (Supplementary Figure 1). For
analogues A-N, intermediate 1 (150.00 mg, 564.55 µmol, 1.00 eq), azetidine variants (2.00 eq) and N,
N-diisopropylethylamine (145.92 mg, 1.13 mmol, 2.00 eq) were taken up into a microwave tube in
isopropyl alcohol (2.00 mL). The sealed tube was heated at 70 °C for 2 hours under microwave. LC-
MS showed that the starting material was consumed completely. The mixture was concentrated in
vacuum to give crude product. The crude product was then purified by thin-layer chromatography.
Analogues O-Q were synthesized following the same procedure with the azetidine variants replaced
with derivatives of the amino group.
Donor characteristics and T-cell cloning.
Four HLA-B*57:01+ donors were selected from the Liverpool Centre for Drug Safety Science cell bank
containing peripheral blood mononuclear cells (PBMCs) from 1200 genotyped healthy donors.
Approval for the study was obtained from the Liverpool research ethics committee and informed
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
Thomson et al 27
consent was received from participants prior to inclusion in the study. PBMCs were incubated in the
presence of abacavir, or analogues G, H, 15 (from our previous study (1)) or J (35µM), in R9 medium
for a period of 14 days. On days 6 and 9 cells were fed with R9 medium containing recombinant
human IL-2 (100U/mL) to preserve the drug driven expansion of T-cells. On day 14, CD8+ T-cells were
positively selected using MultiSort kits (Miltenyi Biotec, Surrey UK) and T-cell clones were generated
via means of serial dilution and phytohaemagglutinin (PHA; 5µg/mL) stimulation. T-cells were fed
every 2 days with R9 medium containing IL-2 (100U/mL) and growing clones were transferred to a
new 96 well plate and expanded across 4 wells. Clones were restimulated and further expanded
every 14 days.
Drug-specific T-cell responses.
Specificity of CD8+ clones was measured by culturing T-cells with irradiated autologous EBV-
transformed B-cells (as antigen presenting cells; 1x104/well) ± abacavir (35µM). Following a 48 h
incubation, [3H] thymidine (0.5µCi) was added and cellular proliferation assessed 16 h later via
scintillation counting. CD phenotyping of clones was performed using BD FACSCanto II flow
cytometer.
IFN-γ ELIspot was used as a second readout of the drug-specific response to assess dose-dependent
T-cell activation and cross-reactivity. Drug-specific clones (5x104, 50µL) were added to ELIspot plates
with EBV transformed B-cells (1x104, 50µL) and abacavir (analogues) (10, 20, 50µM; 100µL) for a
period of 48 h. Spot forming units (cytokine-secreting T-cells) were visualised and quantified using an
AID ELIspot reader (Cadama Medical, Stourbridge, UK).
Modelling of abacavir (analogues) HLA-B*57:01 binding.
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
Thomson et al 28
HLA, HLA binding peptide (HSITYLLPV) and drug ligands (abacavir and analogues D, G, H, M, O, P and
Q) were prepared for docking in Spartan’08 (wavefunction inc. Irvine, California, USA: 1991-2009).
Native SMA ligand was exported from PDB 3UPR. For each analogue, the cyclopropyl group of
abacavir was replaced. Merk molecular force field minimization calculations were performed with all
atoms frozen except newly added substitutions. For docking studies GOLD 5.1 (CCDC Software
Limited, Cambridge, UK) was used to examine the predicted binding poses of the abacavir analogues
within the F-pocket of HLA-B*57:01, PDB code 3UPR (2). Figures of abacavir analogues predicted
binding conformations within HLA-B*57:01 were produced using PYMOL software version 2.5.
MHC purification and peptide elution.
C1R.B*57:01 are transfectants of C1R cells expressing HLA-B*57:01 under geneticin selection (3).
C1R.B*57:01 were grown in RF10 [RPMI 1640 (Life Technologies, USA) supplemented with 10%
foetal bovine serum (FBS; Sigma, St Louis, USA), 7.5 mM HEPES (MP Biomedicals, Germany), 100
U/mL Pen-Strep (benzyl-penicillin/streptomycin, Life Technologies, USA), 2 mM L-glutamine (MP
Biomedicals, Germany), 76μM β-mercaptoethanolamine (Sigma-Aldrich, USA) and 150μM non-
essential amino acids (LifeTechnologies, USA)]. HLA expression was maintained in long term culture
with 0.5mg/mL geneticin (G418; LifeTechnologies, USA).
C1R.B*57:01 cells were grown to high density in the presence or absence abacavir and analogues D
(activates abacavir-responsive T-cells), H (activates abacavir-responsive T-cells at high
concentrations), M (does not activate abacavir-responsive T-cells) and 15 (from our previous study
(1) does not activate abacavir-responsive T-cells) (35μM) for a minimum of 4 days, prior to washing
in PBS, pelleting and snap freezing in liquid nitrogen. Cell pellets of 4-5x108 cells were lysed by
mechanical and detergent based lysis, the lysates cleared by ultracentrifugation, and HLA class I
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
Thomson et al 29
complexes isolated by immunoaffinity purification using solid-phase bound pan class I antibody
W6/32 as described previously (4). Complexes were dissociated using 10 % acetic acid and
fractionated by Reversed Phase High Performance Liquid Chromatography (HPLC) on a 4.6mm
internal diameter x 100mm monolithic reversed-phase C18 HPLC column (Chromolith SpeedROD;
Merck Millipore) using an ÄKTAmicro HPLC (GE Healthcare) system. The peptide/MHC mixture was
loaded at 1mL/min onto the column in 98% Buffer A (0.1% Trifluoroacetic acid) and 2% Buffer B (80%
Acetonitrile, 0.1% trifluoroacetic acid), and bound material eluted by running a gradient of buffer B
at 2ml/min of 2-15% over 0.25 minutes, 15-30% over 4 minutes, 30-40% over 8 minutes, 40-45%
buffer B over 10 min, with collection of 500µL fractions. Peptide containing fractions were vacuum
concentrated, pooled into 9-12 pools, and reconstituted in 0.1% formic acid.
Mass spectrometric analysis.
Reconstituted fraction pools were analysed by liquid chromatography-tandem mass spectrometry
(LC-MS/MS) via a data dependent acquisition strategy using a NanoUltra cHiPLC system (Eksigent)
coupled to an SCIEX 5600+ TripleTOF mass spectrometer equipped with a Nanospray III ion source.
Samples were loaded onto a pre-equilibrated cHiPLC trap column (3µm, ChromXP C18CL, 120 Å, 0.5
mm x 200 µm), at 5µL/min in 0.1% formic acid, 2% acetonitrile, and separated over a cHiPLC column
(3µm, ChromXP C18CL, 120 Å, 15cm x 75µm) using a linear gradient of 2-35% Buffer B (80%
acetonitrile, 0.1% formic acid)/Buffer A (0.1% formic acid) over 75 minutes at a flow rate of
300nL/min. Data acquisition occurred with the following instrument parameters: ion spray voltage,
2,400 V; curtain gas, 30 l/min; ion source gas, 20 l/min; and interface heater temperature, 150 °C.
MS/MS switch criteria selected the top 20 ions meeting the following criteria per cycle: m/z >200
amu, charge state of +2 to +5, intensity >40 counts per second. After two selections for
fragmentation, ions were ignored for 30 seconds. For assignment, MS/MS spectra were searched
against the human proteome (UniProt/Swiss-Prot accessed November 2017) using ProteinPilot™
software (version 5.0, SCIEX), considering biological modifications and utilising a decoy database for
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
Thomson et al 30
false discovery rate (FDR) calculations. Identifications were filtered for peptides seen in HLA class I
immunoaffinity purifications from parental C1R (includes contaminants and peptide binders of
endogenous HLA-B*35:03 and HLA-C*04:01), class II purifications, common contaminants of MHC
pull downs observed in the lab, as well as peptides derived from the HLA proteins themselves
(Supplementary table 1).
HLA-B*57:01 immunopeptidome analysis.
To define the global peptide binding motif under different conditions, distinct peptides assigned by
ProteinPilot with a confidence above a local 5% FDR cut-off and delta mass < 0.05 were considered
(Supplementary table 2). Non-redundant sequences were used to calculate the prevalence of
peptides of each length, and of nine amino acid peptides possessing specific residues at the primary
anchor positions (P2 and PΩ). To compare peptide presentation across the conditions, and filter for
peptides of the constitutive HLA-B*57:01 repertoire, a combined list of peptides identified with a
confidence above a 5% FDR cut-off in at least one data set was used to interrogate all data sets
(Supplementary table 3). Where assignments were made in multiple data sets at a confidence above
the 5% FDR cut-off, the mean and SD of the retention time was calculated. If a peptide assignment
was made below the 5% FDR cut-off, it was considered valid if the retention time was within 5
minutes of the mean 5% FDR retention time. Assignments where the SD of the 5% FDR retention
time was greater than 2.5 were excluded. Abacavir and analogue treated data sets (non-redundant
by sequence) were filtered (by sequence, modifications not considered) of constitutive ligands
identified in the retention time validated untreated data sets as well as those described previously
(5). Sequence features of the retention time validated and filtered data sets were visualised using
Icelogo software (6) (percentage difference to human proteome using static reference method) at P1
to 3 and PΩ-2 to PΩ of 9-12mer peptide ligands.
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
Thomson et al 31
Antiviral activity.
The antiviral activity of the abacavir analogues was measured using cell viability assays carried out by
the Wuxi App Tech company in Shanghai. The antiviral activity was measured on the basis that HIV
kills MT-4 cells meaning cell survival in the presence of an analogue was a measure of antiviral
activity. The antiviral activity of the abacavir analogues was quantified as EC50 and classified into
four categories: good antiviral activity (>10µM), retained antiviral activity (10-60µM), little antiviral
activity (60-500µM) and no antiviral activity (>500µM).
Supplementary references
1. Naisbitt DJ, Yang EL, Alhaidari M, Berry NG, Lawrenson AS, Farrell J, et al. Towards
depersonalized abacavir therapy : chemical modification eliminates HLA-B M 57 : 01-
restricted CD8 R T-cell activation. AIDS. 2015; 29:2385-95.
2. Ostrov D a, Grant BJ, Pompeu Y a, Sidney J, Harndahl M, Southwood S, et al. Drug
hypersensitivity caused by alteration of the MHC-presented self-peptide repertoire. Proc Natl
Acad Sci U S A. 2012; 109:9959–64.
3. Chessman D, Kostenko L, Lethborg T, Purcell AW, Williamson N a, Chen Z, et al. Human
leukocyte antigen class I-restricted activation of CD8+ T cells provides the immunogenetic
basis of a systemic drug hypersensitivity. Immunity. 2008; 28:822–32.
4. Dudek NL, Tan CT, Gorasia DG, Croft NP, Illing PT, Purcell AW. Constitutive and inflammatory
immunopeptidome of pancreatic β-cells. Diabetes. 2012; 61:3018–25.
5. Pymm P, Illing PT, Ramarathinam SH, O’Connor GM, Hughes VA, Hitchen C, et al. MHC-I
peptides get out of the groove and enable a novel mechanism of HIV-1 escape. Nat Struct
Mol Biol. 2017; 24:387–94.
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
Thomson et al 32
6. Colaert N, Helsens K, Martens L, Vandekerckhove J, Gevaert K. Improved visualization of
protein consensus sequences by iceLogo. Nat Methods. 2009; 6:786–7.
Supplementary data
Supplementary table 1:
Peptide contaminants removed from analyses including those seen in similar HLA peptide analyses in
CIR parental cells, class II binders, HLA derived peptides, and peptides with the HLA-C*04:01 motif
(either P2 (F/Y) and P3 (D), or P2(F/Y) and PΩ (LFVM)) of endogenous CIR HLA-C.
Supplementary table 2:
Peptides identified above a 5% FDR cut-off by ProteinPilot™ 5.0 for individual LC-MS/MS data sets.
Supplementary table 3:
Comparison of peptides identified within the data sets using retention time validation of low
confidence assignments by comparison to assignments above 5% FDR cut-off. Identifications are
shown in the IDs tab. Comparison of the data sets by sequence is shown in the Sequences tab.
Supplementary Table 4. Antiviral activity of analogues A-Q (EC50) and ability to induce T-cell
activation in the presence of abacavir-responsive T-cell clones. *T-cell activity only observed at high
concentrations of the analogue.
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
Thomson et al 33
Supplementary Figure 1. Synthesis and structure of abacavir substituted analogues. a) Synthesis of
abacavir substituted analogues i) A-N ii) O-Q. b) Table comparing respective functional groups at the
6-amino cyclopropyl moiety of the abacavir analogues.
Supplementary Figure 2. Mean proliferative response from T-cell clones incubated in the presence
of a) analogue H. b) analogue 15 c) analogue J. Data shows as mean ± SEM of all responsive clones vs
control ± SEM (* P<0.05, ** P<0.01, *** P<0.001).
653
654
655
656
657
658