Profiling Human Cytomegalovirus-Specific T Cell Responses Reveals Novel Immunogenic Open Reading Frames

ABSTRACT

Despite the prevalence and medical significance of human cytomegalovirus (HCMV) infections, a systematic analysis of the targets of T cell recognition in humans that spans the entire genome and includes recently described potential novel open reading frames (ORFs) is not available. Here, we screened a library of epitopes predicted to bind HLA class II that spans over 350 different HCMV ORFs and includes ∼150 previously described and ∼200 recently described potential novel ORFs by using an ex vivo gamma interferon (IFN-γ) FluoroSpot assay. We identified 235 unique HCMV-specific epitopes derived from 100 ORFs, some previously described as immunodominant and others that were not previously described to be immunogenic. Of those, 41 belong to the set of recently reported novel ORFs, thus providing evidence that at least some of these are actually expressed in vivo in humans. These data reveal that the breadth of the human T cell response to HCMV is much greater than previously thought. The ORFs and epitopes identified will help elucidate how T cell immunity relates to HCMV pathogenesis and instruct ongoing HCMV vaccine research.

IMPORTANCE To understand the crucial role of adaptive immunity in controlling cytomegalovirus infection and disease, we systematically analyzed the CMV “ORFeome” to identify new CMV epitopes targeted primarily by CD4 T cells in humans. Our study identified >200 new T cell epitopes derived from both canonical and novel ORFs, highlighting the substantial breadth of the anti-CMV T cell response and providing new targets for vaccine design.

INTRODUCTION

Human cytomegalovirus (HCMV; human herpesvirus 5 [HHV-5]) is a betaherpesvirus that infects the majority of the world’s population. Infection in healthy individuals is characterized by a primary asymptomatic phase, followed by the establishment of lifelong persistence/latency in several cell types (1, 2). HCMV’s 236-kbp double-stranded DNA genome facilitates its persistence and reactivation when immunity is compromised, with both viral and cellular proteins controlling viral gene expression and regulating the dynamic and reversible latent-lytic cycle that develops over a lifelong infection (3, 4). Although largely persistent, its reactivation in immunocompromised populations, such as transplant recipients and AIDS patients, causes severe disease outcomes (5–11). Congenital infection in the developing fetus is also the leading infectious cause of birth defects (12–18). Moreover, the available antiviral drug therapies are insufficient and often toxic in young children (19–22). Consequently, HCMV is recognized as a major public health problem, and development of a vaccine that prevents or at least mitigates virus-induced disease is a top priority (23–25).

Although both humoral and cell-mediated immune responses protect against HCMV infection, a considerable effort has been made toward identifying HCMV targets of cytotoxic T lymphocyte (CTL) responses due to their pivotal role in controlling HCMV disease in immunocompromised individuals (26–29). However, HCMV targets of CD4⁺ T helper cells, which amplify CTL and antibody responses or may mediate direct antiviral activity themselves, remain to be explored in detail. In order to develop a successful HCMV vaccine, it is imperative to assess the large number of candidate viral proteins for their potential to induce robust CD4⁺ T cell responses.

Previous work from Sylwester et al. extensively characterized the canonical HCMV proteins that are targeted by CD4⁺ and CD8⁺ T cell responses (30), and work by many other groups have identified immunodominant epitopes derived from these that include the 65-kDA phosphoprotein (UL83/pp65), immediate early protein 1 (UL123), tegument protein pp150 (UL32), envelope glycoprotein B (UL55), viral transcription factor IE2 (UL122), and major capsid protein (UL86) (31–38). However, a comprehensive analysis of HCMV epitope-specific T cell responses has been challenging, mainly due to the large size of the virus and the evolving impact that persistent infection has on the memory pool. Stern-Ginossar et al. recently reported all the HCMV RNAs found to be associated with ribosomes in infected fibroblasts, increasing the potential number of ORFs the virus may encode by ∼3-fold (39). Here, we designed a comprehensive screening approach to assess potential T cell responses against 563 of these ORFs, which included both previously reported and potentially novel HCMV proteins. A total of 2,593 15-mer peptides were predicted using computational algorithms, and a high-throughput screen was performed using a gamma interferon (IFN-γ) FluoroSpot assay to identify epitopes targeted by both CD8⁺ and CD4⁺ T cells in healthy HCMV-infected adults. This “whole ORFeome” approach resulted in the identification of more than 200 new CD4⁺ and CD8⁺ T cell epitopes.

RESULTS

Targets of HCMV T cell reactivity.

To define the epitopes targeted by HCMV-specific T cell responses in healthy adults, we screened peripheral blood mononuclear cells (PBMCs) of 19 subjects, 10 males and 9 females, recruited from the San Diego blood bank (SDBB). The HCMV seropositivity of all the subjects was confirmed by IgG ELISA (Fig. 1A). We tested a total of 2,593 15-mer HCMV peptides covering a total of 563 ORFs (39). Removing the predicted ORFs that were located entirely within longer ORFs resulted in a set of 359 completely unique ORFs. This set consists of approximately 150 “canonical” ORFs, with an additional 200 identified by rRNA profiling (39). These 15-mer peptides corresponded to epitopes likely to be dominant based on a bioinformatics method that predicts promiscuous binding to HLA class II molecules (40). Each ORF analyzed contained a minimum of 2 predicted epitopes, with the exception of very small ORFs of less than 15 to 20 amino acid residues, in which case at least one peptide was synthetized. The 2,593 peptides were arranged in 89 pools of 28 to 30 15-mers. The PBMC reactivity of each of the 89 pools was assayed directly ex vivo using an IFN-γ FluoroSpot assay. After identifying the pools that resulted in IFN-γ production in HCMV-positive (HCMV⁺) individuals, the top 10 most reactive pools (that, on average, accounted for more than 90% of the reactivity observed within each subject) were then deconvoluted to identify the specific epitopes (Fig. 2). Representative results from the initial screening and the deconvolution of a pool in a representative subject are shown in Fig. 3A and B. In conclusion, the results shown here indicate that human T cell responses to HCMV recognize a wide breadth of different epitope specificities.

FIG 1.

Confirmation of HCMV seropositivity in donors. IgG levels in plasma samples of subjects in the screening cohort (n = 10 males, n = 9 females) (A) and the validation cohort (n = 13 males, n = 26 females) (B) determined by ELISA. The dotted line represents the cutoff for positivity (10 nephelometric turbidity units [NTU]).

FIG 2.

Total T cell response captured by the top 10 epitope pools in each subject. The response magnitude of the top 10 pools is given as a percentage of the total response magnitude observed from all positive pools. On average, the top 10 pools accounted for ∼90% (standard deviation [SD], ±7.8%) of each subject’s total response.

FIG 3.

Strategy for HCMV epitope-specific T cell identification. PBMCs from HCMV-seropositive subjects were stimulated with 2-μg/ml pools and plated on IFN-γ-coated FluoroSpot plates for 20 h. The top 10 positive pools (indicated by asterisks on the bars) were deconvoluted to identify individual epitopes. PBMCs were stimulated with 10 μg/ml of each individual peptide contained in the pool, and reactivity was measured by IFN-γ FluoroSpot assay. (A) SFC/10⁶ PBMCs for one representative subject against the 89 peptide pools; (B) deconvoluted pool representing individual peptides.

Characterization of CMV epitope-specific immune responses.

The deconvolution of the top 10 pools from each subject identified widespread reactivity directed against 235 unique epitopes (Fig. 4 and Table 1). Interestingly, females tended to show both a higher frequency and higher magnitude of epitope-specific responses than males, although this did not reach statistical significance (Fig. 5). On average, each subject recognized 25 epitopes (Fig. 6A) and all subjects recognized at least 2 (range, 2 to 57) (Fig. 6B). Specifically, 6 out of 19 donors recognized 21 to 30 epitopes. A quarter of the epitopes (58 of the 235 recognized) were recognized by three or more subjects (Fig. 6C), and these accounted for 76% of the total T cell response (Fig. 6D).

FIG 4.

Response magnitude of each epitope identified in HCMV-seropositive individuals. Each dot represents an epitope. The y axis represents the response magnitude of individual epitopes. The x axis represents each subject. The median ± interquartile range is shown.

TABLE 1.

Details of 235 HCMV-specific epitopes identified in the screen^a

Epitope	Peptide sequence	Peptide length (aa)	ORF(s)	No. of subjects responding	Magnitude of response (SFC/106 PBMCs)	Response type
1	NGIRWQYQELQYLVE	15	ORFL46W, ORFL46W.iORF1_(UL13), ORFL46W.iORF2	2	222	CD4/CD8
2	RYNALTVRSRDSLLL	15	ORFL46W, ORFL46W.iORF1_(UL13), ORFL46W.iORF2	2	204	ND
3	RVRTWFVQRTTLWRR	15	ORFL46W, ORFL46W.iORF1_(UL13), ORFL46W.iORF2	1	60	CD4
4	GLWVSSYLVRRPMTI	15	ORFL46W, ORFL46W.iORF1_(UL13), ORFL46W.iORF2	2	890	CD4/CD8
5	QGATYQLSIVRQAMQ	15	ORFL46W.iORF1_(UL13)	1	650	CD4
6	GAGLRQLRQQLTVRW	15	ORFL46W.iORF2	1	20	ND
7	MRTVPVTKLYTSRMV	15	ORFL49W_(UL16) (UL16P?)	1	5,677	ND
8	AITLFFFLLALRIPQ	15	ORFL49W.iORF1	1	207	ND
9	ALFTHFVGRPRHCRL	15	ORFL50W_(UL17)	1	57	CD8
10	MLGIRAMLVMLDYYW	15	ORFL53W_(UL20)	1	350	CD4
11	PSVRMDFRARRPLRR	15	ORFL55C_(UL21A)	3	1,740	CD4/CD8
12	ARRLWILSLLAVTLT	15	ORFL57W_(UL22A), ORFL57W.iORF1	1	100	CD4/CD8
13	LLAVTLTVALAAPSQ	15	ORFL57W_(UL22A)	2	6,420	CD4/CD8
14	KDRCLVIRRRWRLVR	15	ORFL64C_(UL23)	1	60	CD4/CD8
15	FVAESITEFLNIGLR	15	ORFL64C_(UL23)	1	530	ND
16	HENGIYYGTRSMRKL	15	ORFL64C_(UL23), ORFL64C.iORF1	1	827	ND
17	FCRRFFFPDRPDFFL	15	ORFL65C	1	107	CD4/CD8
18	AEDSVFTSTRARSAT	15	ORFL70W_(UL25)	1	490	CD4
19	KFVLQDFDVQHLRRL	15	ORFL70W_(UL25)	1	40	ND
20	IINYYYVAQKKARHM	15	ORFL70W_(UL25)	1	163	ND
21	ALALHFLTSRKGVTD	15	ORFL70W_(UL25)	1	40	ND
22	LMITHFQRTIRVLRC	15	ORFL70W_(UL25)	3	2,421	CD4
23	DFLRVVRQQDAFICT	15	ORFL70W_(UL25)	2	174	CD4
24	ICVARLQAQPSSRHI	15	ORFL70W_(UL25)	1	37	ND
25	GVSSVTLLKIFSQVP	15	ORFL70W_(UL25)	2	220	CD4
26	VLATLAAVRTRRRSV	15	ORFL71C, ORFL71C.iORF1 (UL24)	2	340	ND
27	EAYVRINAGQVLPVV	15	ORFL71C, ORFL71C.iORF1 (UL24)	1	1,853	CD4
28	LHCMRYLTSSLVKRY	15	ORFL71C	1	150	ND
29	KRYFRPLLRAWSLGL	15	ORFL71C, ORFL71C.iORF1 (UL24)	4	1,388	CD4
30	HLLRNIKTAFGMRVL	15	ORFL71C, ORFL71C.iORF1 (UL24)	2	1,093	CD4
31	ARNLMEFARVGLRAV	15	ORFL71C.iORF1 (UL24)	1	1,450	CD4
32	TGLVLLLLLLVVRLL	15	ORFL73C	1	1,440	CD4/CD8
33	MLFRPTISNSIPRCR	15	ORFL76C	1	47	CD4
34	LRIIRLLRASIRHEY	15	ORFL79C_(UL27)	1	810	CD8
35	RAHIQKFERLHVRRF	15	ORFL79C_(UL27)	1	2,523	CD4
36	SLQFIGLQRRDVVAL	15	ORFL92C_(UL32)	1	83	CD4/CD8
37	RDVVALVNFLRHLTQ	15	ORFL92C_(UL32)	1	1,710	CD4
38	RRTVLFNELMLWLGY	15	ORFL92C_(UL32)	1	180	ND
39	VNAVNKLVYTGRLIM	15	ORFL92C_(UL32)	1	340	ND
40	KELRMCLSFDSNYCR	15	ORFL92C_(UL32)	1	127	CD4
41	GMKTVAFDLSSPQKS	15	ORFL92C.iORF1	1	193	CD4
42	NAIVLITQLLTNRVL	15	ORFL93W_(UL33)	1	37	ND
43	STNFLTLTVLPFIVL	15	ORFL93W_(UL33)	1	63	CD4
44	VLPFIVLSNQWLLPA	15	ORFL93W_(UL33)	1	247	ND
45	FATVALIAADRYRVL	15	ORFL93W_(UL33)	4	2,310	CD4/CD8
46	SYRSTYIILLLTWFA	15	ORFL93W_(UL33)	3	3,990	CD4/CD8
47	LTLRRTIGTLSRLVP	15	ORFL93W_(UL33)	1	163	CD4
48	RRRMVSVTLFSPYSV	15	ORFL98W.iORF1, ORFL98W.iORF2	1	53	ND
49	GRLMEVRQRNGRLRR	15	ORFL100C	1	30	ND
50	WPERCFIQLRSRSAL	15	ORFL101C, ORFL101C.iORF1_(UL36)	3	313	CD4
51	GPGFMRYQLIVLIGQ	15	ORFL101C, ORFL101C.iORF1_(UL36)	1	3,167	CD8
52	IQTMELMIRTVPRIT	15	ORFL101C, ORFL101C.iORF1_(UL36)	2	384	CD4
53	EFLVRQYVLVDTFGV	15	ORFL101C	2	104	CD8
54	RREAIVRLEKTPTCQ	15	ORFL101C, ORFL101C.iORF1_(UL36)	2	407	CD8
55	RRRFKVCDVGRRHII	15	ORFL101C, ORFL101C.iORF1_(UL36)	1	63	ND
56	RHRFLWQRRRRARLL	15	ORFL103C_(vMIA), ORFL104C_(UL37)	1	1,000	CD4/CD8
57	GSFSSFYSQIARSLG	15	ORFL105C_(UL40)	1	363	ND
58	FLKKMLLCALKGRAS	15	ORFL115C_(UL45), ORFL115C.iORF1	1	930	ND
59	MPVQRLTVNVARCVF	15	ORFL115C_(UL45)	1	20	ND
60	KFIFELYRLPRLSIA	15	ORFL115C_(UL45)	1	173	ND
61	ASKIKMLETRVTLAL	15	ORFL116W_(UL47)	1	140	CD8
62	ATMLSKYTRMSSLFN	15	ORFL127C_(UL48A)	2	650	CD8
63	AFKLDLLRMIAVSRT	15	ORFL127C_(UL48A)	2	477	CD4
64	MLFFQRYAPAFVTGY	15	ORFL143C_(UL54)	1	57	ND
65	DLKYILTRLEYLYKV	15	ORFL143C_(UL54)	1	30	ND
66	DPSYVREHGVPIHAD	15	ORFL143C_(UL54)	1	123	ND
67	TDLIRFERNIVCTSM	15	ORFL145C_(UL55)	3	273	CD4
68	EGIMVVYKRNIVAHT	15	ORFL145C_(UL55)	1	637	ND
69	HTFKVRVYQKVLTFR	15	ORFL145C_(UL55)	1	423	CD8
70	YQKVLTFRRSYAYIH	15	ORFL145C_(UL55)	1	40	ND
71	RRSYAYIHTTYLLGS	15	ORFL145C_(UL55)	5	2,730	CD4/CD8
72	QLMPDDYSNTHSTRY	15	ORFL145C_(UL55)	5	45,517	CD4/CD8
73	NLNCMVTITTARSKY	15	ORFL145C_(UL55)	2	4,810	CD4
74	NADKFFIFPNYTIVS	15	ORFL145C_(UL55)	4	5,120	CD4/CD8
75	GLVVFWQGIKQKSLV	15	ORFL145C_(UL55)	1	53	CD4
76	QLQFTYDTLRGYINR	15	ORFL145C_(UL55)	3	2,236	CD4
77	LRGYINRALAQIAEA	15	ORFL145C_(UL55)	3	30,337	CD4
78	KELSKINPSAILSAI	15	ORFL145C_(UL55)	3	18,557	CD4
79	AILSAIYNKPIAARF	15	ORFL145C_(UL55)	5	11,264	CD4
80	ASCVTINQTSVKVLR	15	ORFL145C_(UL55)	5	8,941	CD4
81	YLFKRMIDLSSISTV	15	ORFL145C_(UL55)	1	67	ND
82	EQAYQMLLALARLDA	15	ORFL145C_(UL55)	4	21,807	CD4/CD8
83	LLDRLRHRKNGYRHL	15	ORFL145C.iORF1	3	12,544	CD4
84	QILWTDGLARRTRDR	15	ORFL145C.iORF2	1	37	CD8
85	RVGITIQQLNVYHQL	15	ORFL146C_(UL56)	1	963	CD4
86	TMRSVFEMQRIRHGA	15	ORFL147C	1	67	ND
87	NIFLVGFYLLVPYLG	15	ORFL147C	1	1,633	ND
88	SLLILVVLLLIYRCC	15	ORFL159W	1	23	CD4
89	LSYMKYHHLHGLPVN	15	ORFL161C_(UL69)	3	879	CD4
90	VELCLGAGAGHVVVV	15	ORFL162W	1	383	ND
91	RSSWRASCVEVPKKP	15	ORFL165W	1	370	CD4
92	MQKYFSLDNFLHDYV	15	UL70	2	20,343	CD4/CD8
93	QTIYFLGLTALLLRY	15	ORFL181C_(UL74)	1	583	CD8
94	SFYLVNAMSRNLFRV	15	ORFL181C_(UL74)	1	43	ND
95	TMRKLKRKQALVKEQ	15	ORFL181C_(UL74)	1	1,563	CD4
96	TAVSEFMKNTHVLIR	15	ORFL181C.iORF1	1	747	CD4
97	WREDVLMDRVRKRYL	15	ORFL189W_(UL77)	1	67	ND
98	IKMWFLLGAPMIAVL	15	ORFL196W_(UL78), ORFL196W.iORF1	3	2,867	CD4
99	LFIIAFFSREPTKDL	15	ORFL196W_(UL78)	1	190	CD4
100	PKSFTLTRIHPEYIV	15	ORFL202C_(UL82/pp71)	2	340	CD4
101	PEYIVQIQNAFETNQ	15	ORFL202C_(UL82/pp71)	4	527	CD4
102	GALTLVIPSWHVFAS	15	ORFL202C_(UL82/pp71)	2	190	CD4/CD8
103	CRSATSLVGNTNADV	15	ORFL203W	1	30	CD4
104	SSCAHTTCRSATSLV	15	ORFL203W	2	104	CD4
105	SWLGQMLRPVGLCTL	15	ORFL204W	1	43	ND
106	QTGIHVRVSQPSLIL	15	ORFL205C_(UL83/pp65), ORFL205C.iORF1	11	33,563	CD4/CD8
107	MSIYVYALPLKMLNI	15	ORFL205C_(UL83/pp65)	1	83	CD8
108	PLKMLNIPSINVHHY	15	ORFL205C_(UL83/pp65)	7	2,920	CD4/CD8
109	ATKMQVIGDQYVKVY	15	ORFL205C_(UL83/pp65), ORFL205C.iORF1	2	4,637	CD4/CD8
110	PKNMIIKPGKISHIM	15	ORFL205C_(UL83/pp65), ORFL205C.iORF1	5	538	CD4/CD8
111	PGKISHIMLDVAFTS	15	ORFL205C_(UL83/pp65), ORFL205C.iORF1	9	2,386	CD4/CD8
112	MNGQQIFLEVQAIRE	15	ORFL205C_(UL83/pp65), ORFL205C.iORF1	6	3,990	CD4/CD8
113	ELRQYDPVAALFFFD	15	ORFL205C_(UL83/pp65)	8	1,106	CD4/CD8
114	GILARNLVPMVATVQ	15	ORFL205C_(UL83/pp65), ORFL205C.iORF1	11	34,633	CD4/CD8
115	ALFFFDIDLLLQRGP	15	ORFL205C.iORF1	2	73	ND
116	RVTGLVFSVVFSVSL	15	ORFL206W	4	6,380	CD4/CD8
117	LTWCVIADRQPRFSV	15	ORFL206W	3	240	CD4
118	RPKRRVVAPFRVAAA	15	ORFL207W	2	283	CD4
119	APFRVAAAGETPLGR	15	ORFL207W	6	1,089	CD4/CD8
120	IPQRLHLIKHYQLGL	15	ORFL209C_(UL85)	1	437	ND
121	IVPMPLALEINQRLL	15	ORFL209C_(UL85)	1	283	CD4
122	LASELTMTYVRKLAL	15	ORFL209C_(UL85)	1	37	CD8
123	HSILADFNSYKAHLT	15	ORFL212C_(UL86) Major Capsid Protein	1	27	ND
124	FHELRTWEIMEHMRL	15	ORFL212C_(UL86) Major Capsid Protein	1	7,343	CD4
125	PQLLFHYRNLVAVLR	15	ORFL212C_(UL86) Major Capsid Protein	1	60	ND
126	RNLVAVLRLVTRISA	15	ORFL212C_(UL86) Major Capsid Protein	1	20	ND
127	LFLAVQFVGEHVKVL	15	ORFL212C_(UL86) Major Capsid Protein	1	53	CD4
128	VRVQDLFRVFPMNVY	15	ORFL212C_(UL86) Major Capsid Protein	1	43	ND
129	LGYNSKFYSPCAQYF	15	ORFL212C_(UL86) Major Capsid Protein	1	20	ND
130	TQEALPILSTTTLAL	15	ORFL212C_(UL86) Major Capsid Protein	1	407	CD4
131	PFTVLRLSYAYRIFA	15	ORFL229W_(UL98)	1	33	ND
132	AREFLLSHDAALFRA	15	ORFL229W_(UL98)	1	23	CD8
133	MLIQQYVLSQYYIKK	15	ORFL229W_(UL98)	1	20	CD8
134	RLGTAATQIQKQTLY	15	ORFL233C	1	30	ND
135	KTQIFNKLFTNRISV	15	ORFL236C	1	1,587	CD4
136	VRSLAVDAQHAAKRV	15	ORFL238W	1	53	CD4
137	LEERDEWVRSLAVDA	15	ORFL238W	1	47	CD4
138	AAITVVPVITQSRLA	15	ORFL245C	3	450	CD4
139	PWYPITQARTLELTP	15	ORFL246C	2	996	CD4
140	MSTKRSTVPWYPITQ	15	ORFL246C	3	477	CD4
141	LRVTFHRVKPTLQRE	15	ORFL248W.iORF1	2	830	CD4
142	SGRVILWTTLRLCIL	15	ORFL249C	1	20	CD4
143	VVRKYWTFTNPNRIL	15	ORFL251W, ORFL252W, ORFL253W_(UL112), ORFL253W.iORF1, ORFL253W.iORF2	3	16,876	CD4/CD8
144	TFDVRQFVFDNARLV	15	ORFL251W, ORFL252W, ORFL253W_(UL112), ORFL253W.iORF1, ORFL253W.iORF2	6	13,736	CD4/CD8
145	VRGGIVFNKSVSSVV	15	ORFL251W, ORFL252W, ORFL253W_(UL112), ORFL253W.iORF1, ORFL253W.iORF2	5	2,691	CD4
146	GNLQVTYVRHYLKNH	15	ORFL253W_(UL112), ORFL253W.iORF1, ORFL253W.iORF2	4	655	CD4/CD8
147	AVAFLNYSSSSSAVS	15	ORFL253W_(UL112), ORFL253W.iORF1, ORFL253W.iORF2	3	2,151	CD4
148	AGLMMMRRMRRAPAE	15	ORFL253W_(UL112)	1	490	ND
149	CDLPLVSSRLLPETS	15	ORFL253W_(UL112)	1	123	CD4
150	CEIKPYVVNPVVATA	15	ORFL253W_(UL112)	3	2,583	CD4/CD8
151	DPLLRLSQVAGSGRR	15	ORFL253W_(UL112)	2	2,067	CD4
152	LPLCSTARLRLAPRR	15	ORFL253W.iORF3	1	467	CD4
153	RATGNFRSTSLYAAV	15	ORFL253W.iORF3	3	4,220	CD4
154	RCCTLRFRRRCRARC	15	ORFL253W.iORF4	2	713	CD4/CD8
155	MSATRHHRCCTLRFR	15	ORFL253W.iORF4	1	277	CD4/CD8
156	RVFCLSADWIRFLSL	15	ORFL254C_(UL114)	2	846	CD4/CD8
157	HLGWQTLSNHVIRRL	15	ORFL254C_(UL114)	1	127	CD8
158	TVVRLHVQIAGRSFT	15	ORFL258C_(UL119)	1	213	ND
159	SCTHPYVISLVTPLT	15	ORFL260C_(UL121)	1	80	ND
160	ISLVTPLTINATLRL	15	ORFL260C_(UL121)	1	317	CD4
161	CRVDADLGLLYAVCL	15	ORFL260C_(UL121)	1	673	CD4
162	VCLILSFSIVTAALW	15	ORFL260C_(UL121)	1	43	ND
163	MFFLAIRDHDTAGGI	15	ORFL261W	1	1,637	CD8
164	LQTMLRKEVNSQLSL	15	ORFL264C_(UL123) IE1, ORFL265C_(UL122) IE2	2	1,606	CD8
165	LVKQIKVRVDMVRHR	15	ORFL264C_(UL123) IE1	3	580	CD4/CD8
166	RVDMVRHRIKEHMLK	15	ORFL264C_(UL123) IE1	2	250	CD4
167	LRRKMMYMCYRNIEF	15	ORFL264C_(UL123) IE1	6	1,697	CD4/CD8
168	CSPDEIMSYAQKIFK	15	ORFL264C_(UL123) IE1	2	194	ND
169	EERDKVLTHIDHIFM	15	ORFL264C_(UL123) IE1	2	107	ND
170	VLCCYVLEETSVMLA	15	ORFL264C_(UL123) IE1	1	93	ND
171	ITKPEVISVMKRRIE	15	ORFL264C_(UL123) IE1	1	1,600	CD4
172	FAQYILGADPLRVCS	15	ORFL264C_(UL123) IE1	1	30	ND
173	EAIVAYTLATAGASS	15	ORFL264C_(UL123) IE1	3	19,333	CD4/CD8
174	TTRPFKVIIKPPVPP	15	ORFL265C_(UL122) IE2	2	424	ND
175	NKGIQIIYTRNHEVK	15	ORFL265C_(UL122) IE2, ORFL265C.iORF1, ORFL265C.iORF2, ORFL265C.iORF3,	7	3,794	CD4
176	LGSMCNLALSTPFLM	15	ORFL265C_(UL122) IE2, ORFL265C.iORF1, ORFL265C.iORF2	4	668	CD4
177	STPFLMEHTMPVTHP	15	ORFL265C_(UL122) IE2, ORFL265C.iORF1, ORFL265C.iORF2, ORFL265C.iORF3,	4	740	CD4
178	YRNMIIHAATPVDLL	15	ORFL265C.iORF3	2	130	CD4
179	VMVRIFSTNQGGFML	15	ORFL265C.iORF3	2	3,560	CD4/CD8
180	VVVGIVLCLSLASTV	15	ORFL266W_(UL124)	1	1,417	CD4/CD8
181	SPVAAELPHPSPAPM	15	ORFL267C	2	166	CD4
182	SYLAVHLRISHRYYH	15	ORFL269C	1	290	CD4/CD8
183	IAITMVMRFWQYING	15	ORFL270C	3	163	CD4
184	TALWLLLGHSRVPRV	15	UL128	1	177	ND
185	AEIRGIVTTMTHSLT	15	ORFL271C_(UL128_truncated)	1	1,713	CD4
186	NPLYLEADGRIRCGK	15	ORFL271C_(UL128_truncated), UL128	2	884	CD4
187	LHRRAAVSGRRSLLQ	15	ORFL271C.iORF1	1	87	CD4
188	MLRLLFTLVLLALYG	15	ORFL278C_(UL148)	3	4,593	CD4
189	HVRLLSYRGDPLVFK	15	ORFL278C_(UL148)	3	333	CD4/CD8
190	VVRFALYLETLSRIV	15	ORFL278C_(UL148)	2	123	ND
191	FYMNWTLRRSQTHYL	15	ORFL278C_(UL148)	6	16,883	CD4/CD8
192	QVEILKPRGVRHRAI	15	ORFL278C_(UL148)	9	3,468	CD4/CD8
193	FCVYRYNARLTRGYV	15	ORFL278C_(UL148)	3	700	CD4/CD8
194	TRGYVRYTLSPKARL	15	ORFL278C_(UL148)	7	9,337	CD4/CD8
195	SLDRFIVQYLNTLLI	15	ORFL278C_(UL148)	8	23,368	CD4/CD8
196	PTWSTTVNAHNSFLH	15	ORFL278C.iORF1	1	47	CD4
197	DRLSTLAATMCMFDY	15	ORFL279C	1	53	CD4
198	LFYRAVALGTLSALV	15	ORFL280C_(UL147A)	3	2,633	CD4/CD8
199	SSIFTSTHRGVIVAP	15	ORFL283W	1	27	CD4
200	LSVRYLSLTAYMLLA	15	ORFL284C_(UL147)	1	1,200	CD8
201	TAYKAFLWKYAKKLN	15	ORFL284C_(UL147)	1	503	CD4
202	WKYAKKLNYHYFRLR	15	ORFL284C_(UL147)	1	237	CD4
203	VYLWYVRRQLVAFCL	15	ORFL318C_(UL148A)	3	253	CD4
204	FPSARDIPKQLPEQP	15	ORFL320W	1	27	ND
205	VVAYVILERLWLAAR	15	ORFL321W.iORF1	1	23	ND
206	IRRWWISVAIVIFIG	15	ORFL321W.iORF2, ORFL321W.iORF3_(UL148D)	3	10,480	CD4
207	RWQFAVCAASKTATR	15	ORFL322W	1	50	CD4
208	PQRLLLTALAIWQRT	15	ORFL324C_(UL150)	1	983	ND
209	PWWRRLRVKRPKFPS	15	ORFS326C, ORFS326C.iORF1_(US1)	1	240	ND
210	LWYLGDYGAILKIYF	15	ORFS337C_(US10)	1	40	CD4
211	LFCGACVITRSLLLI	15	ORFS337C_(US10)	1	487	CD4
212	MNLVMLILALWAPVA	15	ORFS338C_(US11)	1	553	CD4
213	VSEYRVEYSEARCVL	15	ORFS338C_(US11)	1	263	CD4
214	MLVVTVFDTTRLFEI	15	ORFS345C_(US17)	1	840	CD4/CD8
215	VCAFCWLVLPHRLEQ	15	ORFS351C_(US21)	1	1,960	CD4/CD8
216	VSVLYFMPSEPGSAH	15	ORFS351C.iORF2	1	177	ND
217	VFQKTLSMLQGLYLR	15	ORFS352C_(US22)	2	327	ND
218	GLYLRQYDPPALRTY	15	ORFS352C_(US22)	2	633	CD4
219	WFLVMREQAAIPQIY	15	ORFS352C_(US22)	4	1,100	CD4
220	QIYARSLAADYLCCD	15	ORFS352C_(US22)	1	33	ND
221	DFRDLLNFIRQRLCC	15	ORFS352C_(US22)	1	30	CD4
222	PSQEILLLCARHLDE	15	ORFS353C_(US23)	3	110	ND
223	TDCWPFEVAPAARLA	15	ORFS353C_(US23)	2	1,637	CD4
224	LFRAGLMKVYVRRRY	15	ORFS353C_(US23)	1	870	ND
225	VVFMGRFSRVYAYDT	15	ORFS355C_(US24)	1	70	ND
226	EKYMVLVSHNLDELA	15	ORFS355C_(US24)	1	20	ND
227	PRLHCLVTTRSSTRE	15	ORFS355C.iORF1	1	1,277	CD4
228	LRYKWLIRKDRFIVR	15	ORFS361C_(US26)	1	3,247	CD4/CD8
229	TNIMLQVSNVTNHTL	15	ORFS363W_(US27)	1	58	CD4
230	IVVGLPFFLEYAKHH	15	ORFS363W_(US27)	2	1,250	ND
231	YNRMVRFIINYVGKW	15	ORFS363W_(US27)	1	23	ND
232	ITFCLYVGQFLAYVR	15	ORFS363W_(US27)	1	27	ND
233	HDPLGLTRFIMRQLM	15	ORFS370W_(US33A)	1	473	ND
234	FIMRQLMMYPLVLPF	15	ORFS370W_(US33A)	1	530	CD8
235	GLVYRELHDFYGYLQ	15	ORFS371W_(US34)	1	963	ND

^aaa, amino acids; ND, not determined.

FIG 5.

IFN-γ response in males and females. (A) Frequency of response; (B) magnitude of response. Each dot represents a donor. Black dot/bar represents males, and red dot/bar represents females. The median with interquartile range is displayed. Two-tailed Mann-Whitney test.

FIG 6.

Breadth and dominance of HCMV T cell responses. (A) Number of epitopes recognized by each donor, mean ± range. (B) Proportion of the 19 donors that responded to the indicated number of epitopes. (C) Number of the 235 epitopes recognized by an individual donor or by multiple donors (2, 3, 4, and 5 or more). Specifically, 137 epitopes are recognized by an individual donor, and 98 epitopes are recognized by two or more donors. (D) Percentage of the total T cell response induced by all 235 epitopes that is captured by the epitopes recognized by individual or multiple donors. For instance, 13.2% of the total response is directed against epitopes recognized by only one donor, whereas 41.7% is directed against epitopes recognized by 5 or more donors.

We further characterized the phenotype of the T cell responses directed against these 58 dominant epitopes by intracellular IFN-γ staining (representative results shown in Fig. 7A, with the flow cytometry gating strategy shown in Fig. 8A). In the majority of tested subjects, the responding T cells were CD4⁺. More specifically, 68% of all responding T cells were CD4⁺, and 13% were both CD4⁺ and CD8⁺. In 18% of the cases, only CD8⁺ T cells responded to these 58 epitopes (Fig. 7B). Similarly, if the magnitude of the response was considered, 70% of the IFN-γ response was attributable to CD4⁺ T cells and only 30% emanated from CD8⁺ T cells (Fig. 7C). The fact that the responses were dominated by CD4⁺ T cells is consistent with the fact that the peptides tested were originally selected based on their predicted likelihood to bind HLA class II alleles. In turn, the occasional identification of epitope-specific CD8⁺ T cell responses in many cases likely reflects class I epitopes nested within the 15-mer epitopes tested in the screen. Overall, these results indicate that, as expected, the screening strategy employed mostly identifies targets of CD4⁺ T cell reactivity.

FIG 7.

Phenotypic characterization of HCMV T cell responses. (A) Representative fluorescence-activated cell sorting (FACS) plots for intracellular IFN-γ production by CD4⁺ and CD8⁺ T cells (gating axis in red) upon stimulation with two of the scoring peptide epitopes that induced them. (B and C) Number of events and percent response attributable to CD4⁺ and CD8⁺ T cell responses of dominant epitopes (n = 58) that demonstrated a response frequency of 0.15 (15%) (i.e., recognized by 3 or more donors).

FIG 8.

Gating strategy adopted in IFN-γ FluoroSpot and AIM assay. (A) Human PBMCs isolated from HCMV⁺ subjects were stimulated with each scoring peptide to identify HCMV-specific IFN-γ-producing CD4⁺ and CD8⁺ T cells. (B) Human PBMCs isolated from HCMV⁺ and HCMV⁻ subjects were stimulated with each megapool generated to identify HCMV-specific activation-induced marker assay-positive (OX40⁺ CD137⁺) CD4⁺ and (CD69⁺ CD137⁺) CD8⁺ T cells.

Analysis of the ORF of origin of the identified epitopes.

The 235 epitopes identified mapped to a total of 100 of the 359 unique ORFs screened. Of those, 28 ORFs contained more than three immunogenic peptides, and 18 ORFs were recognized in 15% or more of the donors (Fig. 9). Notably, the previously well-characterized immunodominant ORFs, such as envelope glycoprotein B (UL55), IE1 (UL123), tegument protein pp65 (UL83), major capsid protein UL86, IE2 (UL122), and pp150 (UL32), were among those associated with more than three immunogenic peptides.

FIG 9.

T cell epitope distribution by ORF of origin. Two hundred thirty-five epitopes mapped to 100 ORFs. The left y axis indicates the number of epitopes associated with each ORF (bars), and the right y axis indicates the response frequency associated with each ORF (dotted line). Seven canonical ORFs that were common in IEDB and the present screen are indicated in red. ORFL147C (arrow) represents a “new” rRNA-associated ORF where 2 distinct epitopes were identified that were targets of T cell responses in 2 of 19 HCMV⁺ individuals tested.

To address the novelty of our findings, we compared our results with ORFs that have already been reported and curated in the Immune Epitope Database (IEDB; https://www.iedb.org) (41) as a source of defined epitopes. Specifically, a query of the IEDB in October 2020 for previously characterized targets of T cell responses tested in at least 19 donors and with a minimum response frequency of 15% revealed 7 ORFs that match the conditions of our screening results: UL83/pp65 (ORFL205C), UL123/IE1 (ORFL264C), UL122/IE2 (ORFL265C), UL55/gB (ORFL145C), UL32/pp150 (ORFL92C), UL40 (ORFL105C), and UL98 (ORFL229W) (Fig. 9).

The same query revealed three additional ORFs that were not identified in our screen. These ORFs were associated with a limited number of literature-reported and IEDB curated epitopes: UL75/gH (ORFL184C; 1 epitope), UL44/DNA-pol (ORFL112C.iORF1; 3 epitopes), and UL138 (ORFL313C; 1 epitope). Importantly, our screen identified 93 ORFs that were not previously described as targets of T cell responses (Fig. 10).

FIG 10.

Overlap between IEDB-reported and new immunogenic ORFs identified in this T cell epitope screen. New epitopes were identified in all 100 ORFS, including 7 ORFs previously reported in the IEDB to be targets of T cell responses. Of the 93 ORFs found to be new targets of T cells, 52 were canonical and 41 were “novel” as identified by recent ribosomal profiling and mRNA analysis studies.

Notably, 52 of these 93 ORFs were already described in the “canonical HCMV” annotated genome, but not all have been described as targets of human T cell responses. Even more strikingly, 41 of these 93 ORFs corresponded to those viral mRNAs identified only by recent ribosomal profiling studies (39), providing evidence that they are translated in HCMV-infected cells. These results indicate that our approach successfully reidentified known ORFs as targets of T cell responses and, perhaps most importantly, greatly expanded the repertoire of canonical and “novel” ORFs recognized by T cells in healthy adults.

Novel identified epitope pools elicit antigen-specific CD4⁺ T cell responses.

Lastly, we wanted to explore whether the epitopes identified in the presented study could, alone or in combination with previously described epitopes, be utilized to generate epitope “megapools’” (MPs) (42–46) to allow detection of CMV-specific CD4 T cell responses. Accordingly, we generated the P235 MP encompassing the 235 CMV epitopes identified in the present study. For a comparison, we considered the commercially available CMV peptide pool (Mabtech; catalog no. 3619-1) encompassing a total of 42 CD4 and CD8 epitopes. Additionally, we synthetized an MP of known class II epitopes curated in the IEDB database encompassing a total of 187 CD4 epitopes (IEDB-II) (Table 2).

TABLE 2.

Details of HCMV-specific class II epitopes from IEDB

Epitope	Peptide sequence	Peptide length (aa)	ORF	Antigen name from IEDB
1	HINSHSQCYSSYSRVIA	17	ORFL145C_(UL55)	Glycoprotein B
2	SRVIAGTVFVAYHRD	15	ORFL145C_(UL55)	Glycoprotein B
3	CMVTITTARSKYPYH	15	ORFL145C_(UL55)	Glycoprotein B
4	VFETTGGLVVFWQGI	15	ORFL145C_(UL55)	Glycoprotein B
5	MQLIPDDYSNTHSTRYVTVK	20	ORFL145C_(UL55)	Glycoprotein B
6	LPLKMLNIPSINVH	14	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
7	PQYSEHPTFTSQYRIQ	16	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
8	FTSQYRIQGKLEYRHT	16	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
9	PPWQAGILARNLVPMV	16	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
10	KYQEFFWDANDIYRIF	16	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
11	GPISGHVLKAVFSRG	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
12	LLQTGIHVRVSQPSL	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
13	IYVYALPLKMLNIPS	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
14	LPLKMLNIPSINVHH	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
15	KDVALRHVVCAHELV	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
16	RHVVCAHELVCSMEN	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
17	CSMENTRATKMQVIG	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
18	TRATKMQVIGDQYVK	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
19	MQVIGDQYVKVYLES	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
20	VYLESFCEDVPSGKL	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
21	FCEDVPSGKLFMHVT	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
22	LGSDVEEDLTMTRNP	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
23	EEDLTMTRNPQPFMR	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
24	QPFMRPHERNGFTVL	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
25	KISHIMLDVAFTSHE	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
26	MLDVAFTSHEHFGLL	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
27	FTSHEHFGLLCPKSI	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
28	PQYSEHPTFTSQYRI	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
29	SQYRIQGKLEYRHTW	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
30	YRHTWDRHDEGAAQG	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
31	IHNPAVFTWPPWQAG	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
32	PWQAGILARNLVPMV	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
33	ATVQGQNLKYQEFFW	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
34	QNLKYQEFFWDANDI	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
35	QEFFWDANDIYRIFA	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
36	ELEGVWQPAAQPKRR	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
37	IFLEVQAIRETVELR	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
38	PPWQAGILARNLVPM	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
39	DVPSGKLFMHVTLGS	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
40	KLFMHVTLGSDVEED	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
41	DVEEDLTMTRNPQPF	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
42	VAFTSHEHFGLLCPK	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
43	SEHPTFTSQYRIQGK	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
44	LEYRHTWDRHDEGAA	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
45	PLKMLNIPSINVHHY	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
46	KVYLESFCEDVPSGK	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
47	TLGSDVEEDLTMTRN	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
48	ASTSAGRKRKSASSA	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
49	ACTSGVMTRGRLKAE	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
50	AGILARNLVPMVATV	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
51	EPDVYYTSAFVFPTK	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
52	QVIGDQYVKVYLESF	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
53	FFWDANDIYRIFAEL	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
54	LVSQYTPDSTPCHRG	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
55	SHIMLDVAFTSHEH	14	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
56	DEDSDNEIHNPAVFTW	16	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
57	SQYTPDSTPCHRG	13	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
58	KPGKISHIMLDVA	13	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
59	PTFTSQYRIQGKL	13	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
60	DTPVLPHETRLLQTGIHVRV	20	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
61	INVHHYPSAAERKHRHLPVA	20	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
62	LLQRGPQYSEHPTFT	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
63	ALFFFDIDLLLQRGPQYSE	19	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
64	DQYVKVYLESFCEDVPSGKL	20	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
65	MTRNPQPFMRPHERNGFTVL	20	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
66	MISVLGPISGHVLKAVFSRG	20	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
67	ASGKQMWQARLTVSGLAWTR	20	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
68	LPLKMLNIPSINVHHYPSAA	20	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
69	PHETRLLQTGIHVRVSQPSL	20	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
70	IYVYALPLKMLNIPSINVHH	20	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
71	QYDPVAALFFFDIDLLLQRG	20	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
72	RQYDPVAALFFFDIDL	16	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
73	HETRLLQTGIHVRVS	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
74	VYALPLKMLNIPSIN	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
75	VALRHVVCAHELVCS	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
76	HIMLDVAFTSHEHFG	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
77	FTSQYRIQGKLEYRH	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
78	YRIQGKLEYRHTWDR	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
79	ARNLVPMVATVQGQN	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
80	ANDIYRIFAELEGVW	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
81	TRQQNQWKEPDVYYT	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
82	TERKTPRVTGGGAMA	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
83	NLKYQEFFWDANDIY	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
84	TPRVTGGGAMAGAST	15	ORFL205C_(UL83/pp65)	65-kDa lower matrix phosphoprotein
85	DQYVKVYLESFCEDV	15	ORFL205C_(UL83/pp65)	HCMVUL83
86	GKISHIMLDVAFTSH	15	ORFL205C_(UL83/pp65)	HCMVUL83
87	EHPTFTSQYRIQGKL	15	ORFL205C_(UL83/pp65)	HCMVUL83
88	GQNLKYQEFFWDAND	15	ORFL205C_(UL83/pp65)	HCMVUL83
89	KYQEFFWDANDIYRI	15	ORFL205C_(UL83/pp65)	HCMVUL83
90	IIKPGKISHIMLDVA	15	ORFL205C_(UL83/pp65)	HCMVUL83
91	TRATKMQVIGDQYVKVYLES	20	ORFL205C_(UL83/pp65)	HCMVUL83
92	KLFMHVTLGSDVEEDLTMTR	20	ORFL205C_(UL83/pp65)	HCMVUL83
93	KPGKISHIMLDVAFTSHEHF	20	ORFL205C_(UL83/pp65)	HCMVUL83
94	LPVADAVIHASGKQMWQARL	20	ORFL205C_(UL83/pp65)	HCMVUL83
95	GSDSDEELVTTERKTPRVTG	20	ORFL205C_(UL83/pp65)	HCMVUL83
96	RHRQDALPGPCIASTPKKHR	20	ORFL205C_(UL83/pp65)	HCMVUL83
97	YQEFFWDANDIYR	13	ORFL205C_(UL83/pp65)	HCMVUL83
98	LAWTRQQNQWKEPDV	15	ORFL205C_(UL83/pp65)	HCMVUL83
99	YQEFFWDANDIYRIF	15	ORFL205C_(UL83/pp65)	HCMVUL83
100	EFFWDANDIYRIF	13	ORFL205C_(UL83/pp65)	HCMVUL83
101	VEEDLTMTRNPQPFM	15	ORFL205C_(UL83/pp65)	HCMVUL83
102	KPGKISHIMLDVAFTSH	17	ORFL205C_(UL83/pp65)	HCMVUL83
103	TSQYRIQGKLEYRHT	15	ORFL205C_(UL83/pp65)	HCMVUL83
104	MSIYVYALPLKMLNI	15	ORFL205C_(UL83/pp65)	HCMVUL83
105	VYYTSAFVFPTKDVA	15	ORFL205C_(UL83/pp65)	HCMVUL83
106	LRQYDPVAALFFFDI	15	ORFL205C_(UL83/pp65)	HCMVUL83
107	GPQYSEHPTFTSQYRI	16	ORFL205C_(UL83/pp65)	HCMVUL83
108	HPTFTSQYRIQGKLE	15	ORFL205C_(UL83/pp65)	HCMVUL83
109	TRLLQTGIHVRVSQP	15	ORFL205C_(UL83/pp65)	HCMVUL83
110	RNGFTVLCPKNMIIK	15	ORFL205C_(UL83/pp65)	HCMVUL83
111	PISGHVLKAVFSRGD	15	ORFL205C_(UL83/pp65)	HCMVUL83
112	GIHVRVSQPSLILVS	15	ORFL205C_(UL83/pp65)	HCMVUL83
113	IHASGKQMWQARLTV	15	ORFL205C_(UL83/pp65)	HCMVUL83
114	GKQMWQARLTVSGLA	15	ORFL205C_(UL83/pp65)	HCMVUL83
115	ENTRATKMQVIGDQY	15	ORFL205C_(UL83/pp65)	HCMVUL83
116	ATKMQVIGDQYVKVY	15	ORFL205C_(UL83/pp65)	HCMVUL83
117	RPHERNGFTVLCPKN	15	ORFL205C_(UL83/pp65)	HCMVUL83
118	AQGDDDVWTSGSDSD	15	ORFL205C_(UL83/pp65)	HCMVUL83
119	SSATACTSGVMTRGR	15	ORFL205C_(UL83/pp65)	HCMVUL83
120	YRIFAELEGVWQPAA	15	ORFL205C_(UL83/pp65)	HCMVUL83
121	AELEGVWQPAAQPKR	15	ORFL205C_(UL83/pp65)	HCMVUL83
122	AVFSRGDTPVLPHET	15	ORFL205C_(UL83/pp65)	Phosphorylated matrix protein (pp65)
123	ALPLKMLNIPSINVH	15	ORFL205C_(UL83/pp65)	pp65
124	HVLKAVFSRGDTPVL	15	ORFL205C_(UL83/pp65)	pp65
125	AHELVCSMENTRATKMQVIG	20	ORFL205C_(UL83/pp65)	Tegument protein pp65
126	FCEDVPSGKLFMHVTLGSDV	20	ORFL205C_(UL83/pp65)	Tegument protein pp65
127	TLGSDVEEDLTMTRNPQPF	19	ORFL205C_(UL83/pp65)	Tegument protein pp65
128	LLQTGIHVRVSQPSLILV	18	ORFL205C_(UL83/pp65)	Tegument protein pp65
129	SICPSQEPMSIYVYA	15	ORFL205C_(UL83/pp65)	Tegument protein pp65
130	SQEPMSIYVYALPLK	15	ORFL205C_(UL83/pp65)	Tegument protein pp65
131	LNIPSINVHHYPSAA	15	ORFL205C_(UL83/pp65)	Tegument protein pp65
132	HDVSKGDDNKLGGALQAKA	19	ORFL264C_(UL123) IE1	55-kDa immediate-early protein 1
133	ALQAKARDKKDELRRKMMY	19	ORFL264C_(UL123) IE1	55-kDa immediate-early protein 1
134	KEHMLKKYTQTEEKF	15	ORFL264C_(UL123) IE1	55-kDa immediate-early protein 1
135	QTEEKFTGAFNMMGGCLQN	19	ORFL264C_(UL123) IE1	55-kDa immediate-early protein 1
136	MGGCLQNALDILDKVHEPFE	20	ORFL264C_(UL123) IE1	55-kDa immediate-early protein 1
137	AIVAYTLATAGVSSSDSLV	19	ORFL264C_(UL123) IE1	55-kDa immediate-early protein 1
138	TMQSMYENYIVPEDKREMW	19	ORFL264C_(UL123) IE1	55-kDa immediate-early protein 1
139	RRKMMYMCYRNIEFFTKNS	19	ORFL264C_(UL123) IE1	55-kDa immediate-early protein 1
140	FFTKNSAFPKTTNGCSQAM	19	ORFL264C_(UL123) IE1	55-kDa immediate-early protein 1
141	CVETMCNEYKVTSDACMMT	19	ORFL264C_(UL123) IE1	55-kDa immediate-early protein 1
142	DACMMTMYGGASLLSEFCR	19	ORFL264C_(UL123) IE1	55-kDa immediate-early protein 1
143	NYIVPEDKREMWMACIKELH	20	ORFL264C_(UL123) IE1	55-kDa immediate-early protein 1
144	VRHRIKEHMLKKYTQTEEKF	20	ORFL264C_(UL123) IE1	55-kDa immediate-early protein 1
145	VRVDMVRHRIKEHML	15	ORFL264C_(UL123) IE1	55-kDa immediate-early protein 1
146	VKQIKVRVDMVRHRI	15	ORFL264C_(UL123) IE1	55-kDa immediate-early protein 1
147	VRHRIKEHMLKKYTQ	15	ORFL264C_(UL123) IE1	55-kDa immediate-early protein 1
148	EQSDEEEEEGAQEER	15	ORFL264C_(UL123) IE1	55-kDa immediate-early protein 1
149	VKSEPVSEIEEVAPE	15	ORFL264C_(UL123) IE1	55-kDa immediate-early protein 1
150	PVSEIEEVAPEEEED	15	ORFL264C_(UL123) IE1	55-kDa immediate-early protein 1
151	LQNALDILDKVHEPF	15	ORFL264C_(UL123) IE1	55-kDa immediate-early protein 1
152	EDKREMWMACIKELH	15	ORFL264C_(UL123) IE1	55-kDa immediate-early protein 1
153	THIDHIFMDILTTCV	15	ORFL264C_(UL123) IE1	55-kDa immediate-early protein 1
154	VLEETSVMLAKRPLI	15	ORFL264C_(UL123) IE1	55-kDa immediate-early protein 1
155	TKPEVISVMKRRIEE	15	ORFL264C_(UL123) IE1	55-kDa immediate-early protein 1
156	RRIEEICMKVFAQYI	15	ORFL264C_(UL123) IE1	55-kDa immediate-early protein 1
157	NIEFFTKNSAFPKTT	15	ORFL264C_(UL123) IE1	Regulatory protein IE1
158	LTHIDHIFMDILTTCVETM	19	ORFL264C_(UL123) IE1	Regulatory protein IE1
159	AIVAYTLATAGASSSDSLV	19	ORFL264C_(UL123) IE1	UL123; IE1
160	VRVDMVRHRIKEHMLKKYTQ	20	ORFL264C_(UL123) IE1	UL123; IE1
161	DKREMWMACIKELH	14	ORFL264C_(UL123) IE1	UL123; IE1
162	QSMYENYIVPEDKREMWMAC	20	ORFL264C_(UL123) IE1	UL123; IE1
163	TRRGRVKIDEVSRMF	15	ORFL265C_(UL122) IE2	45-kDa immediate-early protein 2
164	GDILAQAVNHAGIDS	15	ORFL265C_(UL122) IE2	45-kDa immediate-early protein 2
165	KTTRPFKVIIKPPVP	15	ORFL265C_(UL122) IE2	45-kDa immediate-early protein 2
166	FKVIIKPPVPPAPIM	15	ORFL265C_(UL122) IE2	45-kDa immediate-early protein 2
167	PEPDFTIQYRNKIID	15	ORFL265C_(UL122) IE2	45-kDa immediate-early protein 2
168	PFTIPSMHQVLDEAI	15	ORFL265C_(UL122) IE2	45-kDa immediate-early protein 2
169	LMQKFPKQVMVRIFS	15	ORFL265C_(UL122) IE2	45-kDa immediate-early protein 2
170	VRIFSTNQGGFMLPI	15	ORFL265C_(UL122) IE2	45-kDa immediate-early protein 2
171	PEDLDTLSLAIEAAI	15	ORFL265C_(UL122) IE2	45-kDa immediate-early protein 2
172	TLSLAIEAAIQDLRN	15	ORFL265C_(UL122) IE2	45-kDa immediate-early protein 2
173	SMHQVLDEAIKACKT	15	ORFL265C_(UL122) IE2	45-kDa immediate-early protein 2
174	KGIQIIYTRNHEVKS	15	ORFL265C_(UL122) IE2	45-kDa immediate-early protein 2
175	ALSTPFLMEHTMPVT	15	ORFL265C_(UL122) IE2	45-kDa immediate-early protein 2
176	FLMEHTMPVTHPPEV	15	ORFL265C_(UL122) IE2	45-kDa immediate-early protein 2
177	PYAVAFQPLLAYAY	14	UL57	Single-stranded DNA-binding protein
178	KTQLNRHSYLKDSDFLDAA	19	UL75	Envelope glycoprotein H
179	RQTEKHELLVLVKKAQLNRH	20	UL75	Glycoprotein H precursor
180	LDPHAFHLLLNTYGRPIR	18	UL75	Glycoprotein H precursor
181	KAQLNRHSYLKDSDFLDAA	19	UL75	Glycoprotein H precursor
182	DVLKSGRCQMLDRRTVEMA	19	UL75	Glycoprotein H precursor
183	LDKAFHLLLNTYGRPIR	17	UL75	Glycoprotein H precursor
184	KDQLNRHSYLKDPDFLDAA	19	UL75	Glycoprotein H precursor
185	SYLKDSDFLDAAL	13	UL75	HCMVUL75
186	RRIPHFYRVRREVPRTVNE	19	UL86	Major capsid protein
187	MDVNYFKIPNNPRGRASCM	19	UL86	Major capsid protein

These MPs were tested with PBMCs from a new cohort of 20 individuals (6 males and 14 females), which included both HCMV-seropositive and -seronegative donors (10 CMV⁺ and 10 CMV⁻) (Fig. 1B for IgG ELISA CMV confirmation). None of the PBMCs from these subjects were used in the original epitope mapping experiments. PBMCs were stimulated with the Mabtech, P235, IEDB-II, or a combination of both P235 and IEDB-II (P235–IEDB-II) MPs. CD4⁺ and CD8⁺ T cell responses were measured as a percentage of activation-induced marker assay-positive (OX40⁺/CD137⁺ for CD4⁺ and CD69⁺/CD137⁺ for CD8⁺) T cells, and results are displayed in Fig. 11 (flow cytometry gating strategy shown is in Fig. 8B).

FIG 11.

Epitope-specific CD4⁺ and CD8⁺ T cell responses in HCMV⁺ and HCMV⁻ subjects detected with different peptide pools. (A and C) Representative FACS plots showing HCMV-specific CD4⁺ and CD8⁺ T cell reactivity against different peptide pools based on activation-induced marker assays (OX40⁺/CD137⁺ coexpression for CD4⁺ responses and CD69⁺/CD137⁺ coexpression for CD8⁺ responses). PBMCs from HCMV⁺ (red circles) and HCMV⁻ donors (gray circles) were stimulated with 2 μg/ml of the Mabtech pool or IEDB-II–P235 pools for 24 h. (B and D) Epitope pool-specific CD4⁺ and CD8⁺ T cells measured as a percentage of activation-induced marker assay-positive (OX40⁺/CD137⁺) CD4⁺ and (CD69⁺/CD137⁺) CD8⁺ T cells. Each dot represents an individual subject. HCMV⁺ subjects demonstrated significantly higher CD4⁺ and CD8⁺ T cell AIM responses than HCMV⁻ subjects for all the different pools tested. For CD4⁺ T cell responses, Mabtech HCMV⁺ versus HCMV⁻, P = 0.0007; P235 HCMV⁺ versus HCMV⁻, P = 0.0065; IEDB-II HCMV⁺ versus HCMV⁻, P = 0.0009; P235/IEDB-II HCMV⁺ versus HCMV⁻, P = 0.004. For CD8⁺ T cell responses, Mabtech HCMV⁺ versus HCMV⁻, P = 0.0007; P235 HCMV⁺ versus HCMV⁻, P = 0.0031; IEDB-II HCMV⁺ versus HCMV⁻, P = 0.0001; P235/IEDB-II HCMV⁺ versus HCMV⁻, P = 0.0001 (using a two-tailed Mann-Whitney test). Comparisons across different pool formulations within the HCMV⁺ subjects were made using one-way analysis of variance (ANOVA) with Tukey’s multiple comparison test; P values are shown in the figure. Values are the geometric mean with the geometric standard deviation.

All HCMV MPs tested were associated with significantly higher CD4⁺ and CD8⁺ activation-induced marker (AIM) responses in HCMV⁺ individuals than in HCMV⁻ subjects, as shown in Fig. 11 (statistical differences are detailed in the figure legend). When comparing AIM responses between the HCMV pools, the IEDB-II and P235–IEDB-II MPs were associated with significantly higher HCMV-specific CD4 responses than the Mabtech pool. In contrast, CD8 AIM responses in HCMV⁺ subjects were comparable for all the pools. This was expected, as the Mabtech pool contains fewer epitopes, which are also mainly CD8 T cell specific. Additionally, the combination of the P235 and IEDB-II MPs elicited higher CD4 responses than either MP alone and had the highest-magnitude response of all pools tested. This indicates that the combination of known (IEDB-II MP) and novel epitopes and ORFs (P235 MP) can capture the broadest range of CD4 T cell responses in HCMV⁺ individuals, which has high potential for clinical diagnostic use.

DISCUSSION

In this study, we have identified >200 new epitopes derived from 100 HCMV ORFs that induce virus-specific T cell responses. Importantly, this demonstrates that the current HLA peptide-binding prediction algorithms that we and others have refined over the last several decades are extremely efficient (47–51) and represent an excellent alternative to synthesizing genome-wide overlapping peptides, especially for large pathogens such as HCMV. Despite the significant diversity in the human HLA repertoire, current advances in algorithm-based epitope identification take into consideration epitopes with potential binding to diverse haplotypes, which undoubtedly contributed to this success (40, 52). Together, this approach allowed us to increase the known T cell epitope landscape for HCMV by more than 10-fold by synthesizing only 2,593 peptides, illustrating both its efficiency and cost-effectiveness in deciphering immune targets of large pathogens.

We chose to use IFN-γ production as a readout for positive epitope reactivity in a FluoroSpot-based assay to identify HCMV-specific T cell epitopes in this study. As true for most viral infections, CMV drives a strong Th1-like CD4⁺ response, and most effector and memory viral CD8⁺ T cells also produce this cytokine (53). However, future studies assessing which of these 235 epitopes may elicit HCMV-specific CD4 T cells to produce other cytokines are merited. Previously, we have observed that dengue virus epitope-specific CD4⁺ T cells can produce both IFN-γ and interleukin 10 (IL-10) (54), something we have also seen during acute CMV infection in mice (55), where IL-10-producing CD4⁺ T cells enhance the duration of viral persistence (56). Recent studies by the Wills and Moss groups show that subsets of HCMV epitope-specific CD4⁺ T cells can produce IL-10 and also display cytolytic markers (57, 58). The potential CTL activity of HCMV-specific CD4⁺ T cells has been postulated for many years (59), and our recent results showing that CMV epitope-specific CD4 T cells can directly kill in vivo support this hypothesis (60). Taken together, our identification of >200 new T cell epitopes that elicit IFN-γ production in this study provide us and others in the field valuable new tools to dissect the phenotypes and effector functions of HCMV-specific CD4 T cells in cases of both healthy and immunocompromised patients and will also help instruct ongoing vaccine efforts.

Of the 100 ORFs which we show here to be sources of specific T cell epitopes, 41 were uniquely identified as ribosome-bound RNAs in HCMV-infected fibroblasts (39), with these 41 yielding 50 unique epitopes. Notably, of these 41 ORFs, 17 are predicted to produce proteins of <50 amino acids in length, and 7 contain non-ATG start codons. This is consistent with recent studies suggesting that the short/“cryptic” mRNAs present in both virally infected and tumor cells can be translated, proteolytically processed, and loaded onto HLA molecules, resulting in the induction of epitope-specific T cell responses (61–63). Interestingly, one of the larger 41 ORFs containing T cell epitopes (ORFL147C, 476 amino acids) has very recently been shown to regulate RNA binding/processing, and its deletion compromises CMV replication in fibroblasts (64). In turn, although >20% of the epitopes we identified were derived from these 41 ribosome-associated HCMV RNAs, no more than 2 of the 19 HCMV⁺ donors we tested induced T cell responses targeting any single one of them. The rather sparse recognition of these 41 ORFs differs from the more “classic” immunodominant HCMV targets such as UL55/gB and UL83/pp65, where both us and others have seen that the majority of people mount T cell responses to multiple epitopes derived from these ORFs, regardless of their HLA haplotype. This indicates that (i) these novel 41 ORFs may not be broad targets of T cell responses in infected persons, (ii) that specific individuals may more efficiently present epitopes derived from short/cryptic HCMV RNAs, or (iii) that minor HLA molecules may present them, with other possibilities also existing. Additionally, whether the proteins derived from the shorter subset of these 41 ORFs are stable and play a role in the HCMV life cycle remains an open question. Finally, we also identified 24 epitopes derived from 14 “canonical” HCMV ORFs, where the only historic support for their existence was the presence of their RNA in infected cells or bioinformatics analyses (4, 65). Notably, a recent comprehensive study in which 169 predicted canonical HCMV proteins (including these 14) were epitope tagged, expressed stably in infected cells, immunoprecipitated, and analyzed for interacting proteins by mass spectrometry supports our results that these ORFs are expressed as proteins (64).

Of the 59 canonical ORFs that we have identified here to contain T cell epitopes, >25% of these are known to function as immunomodulatory proteins (66). This is intriguing, as perhaps these HCMV proteins are more subject to being localized to antigen-processing or presentation compartments within infected cells. One of these epitopes is derived from the HCMV IL-10 orthologue, which is being considered as a potential HCMV vaccine candidate (67, 68). Additionally, 3 epitopes were found to be embedded within the viral UL128 protein, a critical component of the pentameric envelope protein complex (UL128-131/gH/gL) that mediates entry of HCMV into nonfibroblast cell types (69, 70). This is also of high potential interest in the context of vaccine development, as many believe that the pentamer should be included in a virus- or subunit-based approach (71). Notably, both viral IL-10 and UL128 have largely been considered only in the context of their abilities to induce antibody-based vaccine protection, but our identification of T cell epitopes derived from both of these HCMV proteins suggests that they may function to prime both humoral and cellular immunity.

MATERIALS AND METHODS

Study design.

For the initial HCMV ORF screen, the responses of 19 HCMV-seropositive subjects were evaluated. PBMCs were stimulated with 89 pools covering 563 ORFs of HCMV. Each pool comprised 28 to 30 15-mer peptides. PBMCs that were found reactive to a pool were further tested against individual peptides contained in the pool by using an IFN-γ FluoroSpot assay. Flow cytometry was then used to further characterize the epitopes recognized by PBMCs stimulated with individual peptides by detecting IFN-γ production from CD8⁺ and CD4⁺ T cells.

For the CMV-235 pool (P235) validation and comparison screen, the responses of a new cohort consisting of 10 CMV-seropositive and 10 HCMV-seronegative subjects were evaluated. PBMCs were stimulated with the CMV Mabtech peptide pool (catalog no. 3619-1), the CMV IEDB peptide pool (Table 2) (44, 46), the P235 pool, or a combination of both the CMV IEDB and P235 pools. PBMC responses were assayed using the same IFN-γ FluoroSpot assay. These studies were approved by the institutional review board committee at La Jolla Institute (protocol no. VD-112 and VD-174).

Subjects.

Nineteen subjects (10 males and 9 females) were recruited anonymously from the San Diego blood bank (SDBB) for the initial CMV ORF screens. For the CMV-235 comparison screens, samples from 20 subjects (6 males and 14 females) were obtained by the La Jolla Institute Clinical Core and Continental Services Group (Miami, FL) for prior, unrelated studies. Blood samples were collected by trained staff. At the time of enrollment in the initial studies, all individual subjects provided informed consent that any leftover sample could be used for future studies, which includes this study. These subjects were considered healthy as defined by no known history of any significant systemic diseases (not limited to autoimmune disease, diabetes, kidney, or liver disease, congestive heart failure, malignancy, coagulopathy, hepatitis B or C, or HIV). The demographics of these subjects are provided in Table 3. Only one person of African-American descent was included in the randomly selected cohort. Although this is consistent with the fact that only 5% of the San Diego population is African-American, it is not consistent with the fact that 13% of the entire U.S. population is of this ethnicity.

TABLE 3.

Demographic characteristics of HCMV⁺ or HCMV⁻ subjects analyzed in screening and validation studies

Characteristic	Value for:
	HCMV+ screening cohort	Validation cohort
		HCMV+	HCMV−
Total participants enrolled, n	19	10	10
Males/females, n	10/9	3/7	3/7
Median age (yrs) (range)	65 (28–80)	35.5 (22–55)	28.5 (19–42)
Caucasian, % (n)	68 (13)	40 (4)	40 (4)
Asian, % (n)	16 (3)	10 (1)	20 (2)
African-American, % (n)	5 (1)	10 (1)	10 (1)
More than one race, % (n)	0 (0)	30 (3)	30 (3)
Race unknown, % (n)	10 (2)	10 (1)	0 (0)

The IgG antibodies of the subjects for both cohorts were measured using a cytomegalovirus IgG ELISA kit from Genway Biotech, Inc., according to the manufacturer’s instructions. All 19 donors were also HLA typed (Table 4).

TABLE 4.

HLA typing results of the 19 CMV⁺ subjects in the screening cohort

Donor ID	HLA-A(L)	HLA-A(R)	HLA-B(L)	HLA-B(R)	HLA-C(L)	HLA-C(R)	HLA-DPB1(L)	HLA-DPB1(R)	HLA-DQA1(L)	HLA-DQA1(R)	HLA-DQB1(L)	HLA-DQB1(R)	HLA-DRB1(L)	HLA-DRB1(R)	HLA-DRB3/4/5(L)	HLA-DRB3/4/5(R)
4244	A*24:02	A*34:01	B*15:35	B*40:02	C*07:02	C*15:02	DPB1*05:01	DPB1*05:01	DQA1*01:03	DQA1*03:01	DQB1*04:02	DQB1*05:03	DRB1*04:05	DRB1*04:05	DRB4*01:01	DRB4*01:01
4245	A*30:01	A*30:02	B*35:01	B*44:03	C*04:01	C*16:01	DPB1*02:01	DPB1*02:01	DQA1*01:02	DQA1*05:01	DQB1*03:01	DQB1*06:09	DRB1*11:01	DRB1*13:02	DRB3*02:02	DRB3*03:01
4246	A*23:01	A*24:02	B*18:01	B*52:01	C*12:02	C*12:03	DPB1*04:01	DPB1*05:01	DQA1*01:03	DQA1*05:01	DQB1*03:01	DQB1*06:01	DRB1*11:04	DRB1*15:02	DRB3*02:02	DRB5*01:02
4260	A*02:01	A*29:02	B*35:01	B*44:03	C*04:01	C*16:01	DPB1*01:01	DPB1*02:01	DQA1*01:01	DQA1*02:01	DQB1*02:02	DQB1*05:01	DRB1*01:01	DRB1*07:01	DRB4*01:01
4261	A*11:01	A*24:02	B*07:02	B*40:01	C*03:04	C*07:02	DPB1*04:01	DPB1*04:01	DQA1*03:01	DQA1*03:01	DQB1*03:02	DQB1*03:02	DRB1*04:01	DRB1*04:04	DRB4*01:01	DRB4*01:01
4263	A*01:01	A*02:01	B*08:01	B*55:01	C*03:03	C*07:01	DPB1*01:01	DPB1*03:01	DQA1*05:01	DQA1*05:01	DQB1*02:01	DQB1*03:01	DRB1*03:01	DRB1*11:01	DRB3*01:01	DRB3*02:02
4264	A*02:03	A*25:01	B*38:02	B*51:01	C*07:02	C*16:01	DPB1*01:01	DPB1*03:01	DQA1*05:01	DQA1*05:01	DQB1*03:01	DQB1*03:01	DRB1*11:01	DRB1*11:02	DRB3*02:02	DRB3*02:02
4265	A*24:02	A*33:03	B*44:03	B*52:01	C*12:02	C*14:03	DPB1*02:01	DPB1*05:01	DQA1*01:02	DQA1*01:03	DQB1*06:01	DQB1*06:04	DRB1*13:02	DRB1*15:02	DRB3*03:01	DRB5*01:02
4266	A*01:01	A*11:01	B*18:01	B*35:03	C*12:03	C*12:03	DPB1*02:01	DPB1*04:01	DQA1*01:01	DQA1*03:01	DQB1*03:04	DQB1*05:03	DRB1*04:08	DRB1*14:01	DRB3*02:02	DRB4*01:01
4267	A*01:01	A*11:01	B*08:01	B*35:01	C*04:01	C*07:01	DPB1*04:01	DPB1*10:01	DQA1*01:01	DQA1*05:01	DQB1*02:01	DQB1*05:01	DRB1*01:01	DRB1*03:01	DRB3*01:01
4270	A*01:01	A*02:01	B*35:02	B*44:02	C*05:01	C*06:02	DPB1*04:01	DPB1*04:01	DQA1*01:03	DQA1*05:01	DQB1*03:01	DQB1*06:03	DRB1*11:04	DRB1*13:01	DRB3*01:01	DRB3*02:02
4271	A*02:01	A*29:02	B*08:01	B*44:03	C*07:01	C*16:01	DPB1*01:01	DPB1*01:01	DQA1*02:01	DQA1*05:01	DQB1*02:01	DQB1*02:02	DRB1*03:01	DRB1*07:01	DRB3*01:01	DRB4*01:01
4272	A*02:01	A*02:01	B*15:01	B*35:01	C*03:04	C*04:01	DPB1*04:02	DPB1*04:02	DQA1*01:01	DQA1*01:02	DQB1*05:01	DQB1*06:02	DRB1*01:01	DRB1*15:01	DRB5*01:01
4277	A*11:01	A*11:01	B*13:02	B*38:02	C*06:02	C*07:02	DPB1*01:01	DPB1*17:01	DQA1*01:02	DQA1*02:01	DQB1*02:02	DQB1*05:02	DRB1*07:01	DRB1*15:02	DRB4*01:01	DRB5*01:01
4278	A*02:01	A*29:02	B*08:01	B*44:03	C*07:01	C*16:01	DPB1*01:01	DPB1*11:01	DQA1*02:01	DQA1*05:01	DQB1*02:01	DQB1*02:02	DRB1*03:01	DRB1*07:01	DRB3*01:01	DRB4*01:01
4288	A*01:01	A*03:01	B*07:02	B*27:05	C*02:02	C*07:02	DPB1*04:01	DPB1*04:01	DQA1*01:02	DQA1*01:02	DQB1*06:02	DQB1*06:02	DRB1*15:01	DRB1*15:01	DRB5*01:01	DRB5*01:01
4289	A*11:01	A*26:01	B*13:01	B*15:01	C*03:04	C*04:01	DPB1*05:01	DPB1*05:01	DQA1*03:01	DQA1*03:01	DQB1*03:02	DQB1*03:03	DRB1*04:06	DRB1*09:01	DRB4*01:01	DRB4*01:01
4291	A*24:02	A*31:01	B*40:02	B*51:01	C*03:04	C*14:02	DPB1*02:01	DPB1*04:02	DQA1*03:01	DQA1*05:01	DQB1*03:01	DQB1*03:02	DRB1*04:11	DRB1*14:03	DRB3*01:01	DRB4*01:01
4292	A*02:01	A*30:02	B*07:02	B*27:05	C*02:02	C*07:02	DPB1*04:01	DPB1*04:01	DQA1*01:01	DQA1*01:02	DQB1*05:01	DQB1*06:02	DRB1*01:01	DRB1*15:01	DRB5*01:01

Peptide prediction.

Based on the 7-allele method as previously described (40), 2,593 peptides were predicted for 563 potential HCMV ORFs. Of the 751 ORFs predicted by ribosomal profiling (39), those smaller than 15 amino acids were excluded, and only one peptide of the ORFs of 15 to 20 amino acids in length was selected for screening.

Peptide libraries and pool preparation.

The predicted peptides were commercially synthesized as crude material by TC Peptide Lab (San Diego, CA). The peptides were solubilized in dimethyl sulfoxide (DMSO) at a concentration of 20 mg/ml and spot checked for quality by mass spectrometry. The peptides were pooled into peptide pools containing 28 to 30 peptides constituting multiple ORFs per pool. A total of 89 pools were prepared covering 563 ORFs of HCMV. The final concentration of each pool was 0.7 mg/ml.

For the IEDB-II (Table 2) and P235 (Table 1) peptide pools, peptides were synthesized by A&A Ltd., San Diego, CA, resuspended in DMSO, pooled, and sequentially lyophilized as previously described (72). The IEDB-II peptide pool was developed based on data available in the IEDB (https://www.iedb.org) (41). The major histocompatibility complex (MHC) class II restricted epitopes for HCMV were extracted from the IEDB in October of 2020 by use of the following query: organism, human herpesvirus 5 (ID:10359), positive assays only, no B cell assays; MHC restriction type, class II; host, Homo sapiens. The resulting 187 epitopes (Table 2) were filtered for size (13 to 20 amino acids) and discovered using one of the following assays: enzyme-linked immunosorbent spot assay (ELISPOT), intracellular cytokine staining (ICS), multi- or tetramers, proliferation, and “helper response.” The CMV peptide pool for human CD4 and CD8 T cells containing 42 peptides (14 MHC class II restricted and 28 MHC class I restricted) representing pp50, pp65, IE1, IE2, and envelope glycoprotein B was purchased from Mabtech.

Isolation of PBMCs by Ficoll-Paque density gradient centrifugation.

One unit of blood from each donor was processed for PBMC isolation. Briefly, blood was centrifuged, and the top layer of plasma was removed. The remaining blood was diluted and layered over 15 ml of Ficoll-Paque. Tubes were spun at room temperature in a swinging-bucket rotor without brake applied. The PBMC interface was carefully removed by pipetting and washed with phosphate-buffered saline (PBS) by centrifugation at 800 rpm for 10 min with brakes off. PBMC pellet was resuspended in RPMI medium, cell number and viability were determined by trypan blue staining, and cells were cryopreserved in liquid nitrogen in freezing medium (90% fetal bovine serum and 10% DMSO) at a density of 30 × 10⁶/ml and stored until further processing.

FluoroSpot assay.

PBMCs were thawed, washed, and counted for viability by the trypan blue exclusion method. A total of 200,000 cells were plated in triplicate and stimulated with pools (2 μg/ml) or peptides (10 μg/ml), phytohemagglutinin (PHA) (10 μg/ml), or medium containing an equivalent amount of DMSO in 96-well plates (Immobilon-P; Millipore) previously coated with anti-IFN-γ antibody (1-D1K; Mabtech, Stockholm, Sweden). After 20 h of incubation at 37°C, cells were discarded and wells were washed six times with PBS–0.05% Tween 20 using an automated plate washer and further incubated with IFN-γ antibody (7-B6-1-FS-BAM) for 2 h at room temperature. After incubation, wells were washed and incubated with fluorophore-conjugated anti-BAM-490 antibody for 1 h at room temperature. Finally, the plates were washed and incubated with fluorescence enhancer for 15 min and blotted dry, and fluorescent spots were counted by computer-assisted image analysis (IRIS FluoroSpot reader; Mabtech, Sweden).

Each pool or peptide was considered positive in comparison to the background that had an equivalent amount of DMSO based on the following criteria: (i) 20 or more spot-forming cells (SFC) per 10⁶ PBMCs after background subtraction, (ii) a stimulation index greater than 2, and (iii) a P value of <0.05 by Student's t test or Poisson distribution test when comparing the peptide or pool triplicates with the negative-control triplicate.

Intracellular cytokine assay for IFN-γ.

Intracellular staining for IFN-γ and flow cytometry was performed to detect antigen-specific T cell responses. A total of 1 × 10⁶ PBMCs suspended in RPMI medium supplemented with 1% heat-inactivated human AB serum, glutamine, and penicillin-streptomycin were plated in U-bottom 96-well plates. After overnight resting at 37°C, PBMCs were spun and replaced with fresh RPMI medium and stimulated with individual peptides at a concentration of 10 μg/ml. PHA at a concentration of 5 μg/ml was used as a positive control. After 1 h of incubation at 37°C, 2 μg/ml of brefeldin was added and cells were further incubated at 37°C for an additional 5 h. The cells were then harvested, washed with 200 μl of magnetically activated cell sorting (MACS) buffer and stained with a cocktail of antibodies that contained CD3-Af700 (eBioscience; clone UCHT1), CD4-APCef780 (eBioscience; clone RPA-T4), CD8-BV650 (Biolegend; clone RPA-T8), CD14-V500 (BD Biosciences; clone M5E2), CD19-V500 (BD Biosciences; clone HIB19), and fixable viability dye e506 for 30 min at 4°C. The cells were then washed three times with 200 μl MACS buffer, fixed using 4% PFA for 10 min at 4°C, washed with 200 μl PBS, and rested at 4°C overnight in 200 μl MACS buffer. The following day, cells were washed, permeabilized by washing with 200 μl saponin buffer (0.5% saponin in PBS), washed with blocking buffer (10% human serum prepared in saponin buffer), and stained with IFN-γ–FITC (fluorescein isothiocyanate) (eBioscience; clone 4S.B3) antibody at room temperature for 30 min. The cells were finally washed with PBS and suspended in 200 μl PBS.

The cells were acquired on a ZE5 Bio-Rad plate reader, and further analysis was done on FlowJo software. Gates were applied on live single cells for CD3⁺, CD4⁺, and CD8⁺ T cell populations. The percentage of reactive CD4⁺ or CD8⁺ IFN-γ T cells was expressed as a percentage of the total number of the parent population analyzed. Reactive populations met the following two criteria: (i) well-defined cell population positive for both IFN-γ and CD4 or CD8 constituting at least 0.02% (after subtraction of their corresponding DMSO controls) of the total number of CD4⁺ or CD8⁺ cells analyzed, and (ii) a stimulation index greater than 2.

Activation-induced marker (AIM) assay.

PBMCs were thawed, washed, and counted for viability using the trypan blue exclusion method. A total of 10⁶ cells per donor/condition were plated and cultured in the presence of the CMV-specific pools (1 μg/ml for P235 and IEDB-II pools and 2 μg/ml for Mabtech pool), PHA (10 μg/ml), or medium containing an equivalent amount of DMSO in 96-well U-bottom plates. Cells were then harvested, washed with 200 μl of MACS buffer, and stained with a cocktail of antibodies that contained CD3-Af700 (eBioscience; clone UCHT1), CD4-BV605 (eBioscience; clone RPA-T4), CD8-peridinin chlorophyll protein (PerCP)-Cy5.5 (Biolegend; clone HIT8a), CD14-V500 (BD Biosciences; clone M5E2), CD19-V500 (BD Biosciences; clone HIB19), OX40-phycoerythrin (PE)-Cy7 (Biolegend; clone Ber-ACT35), CD137-allophycocyanin (APC) (Biolegend; clone 4B4-1), CD69-PE (BD Biosciences; clone FN50), and fixable viability dye e506 for 30 min at 4°C. The cells were then washed three times with 200 μl MACS buffer, fixed using 4% PFA for 10 min at 4°C, and resuspended in 200 μl of PBS for acquisition.

Cells were acquired on a BD LSRFortessa cell analyzer, and further analysis was done using FlowJo software. As previously described (44, 73), quantification of live, singlet antigen-specific CD4⁺ and CD8⁺ T cells was determined as a percentage of their OX40⁺ CD137⁺ and CD69⁺ CD137⁺ double expression (AIM⁺), respectively. CMV-specific AIM⁺ CD4 and CD8 T cell signals were background subtracted with their corresponding negative-control DMSO samples, with the minimal DMSO level set to 0.005%. The limit of detection (LOD) for the AIM⁺ assay was calculated by multiplying the upper confidence interval of the geometric mean of all DMSO samples by 2 (0.03 for CD4⁺ and 0.05 for CD8⁺ T cell responses).

Statistical analysis.

Statistical analyses were performed using GraphPad Prism versions 8.1.1 and 8.4.3. Statistical details are provided with each figure.