Cross-typic specificity and immunotherapeutic potential of a human HPV16 E7-specific CTL line

Abstract

Cervical cancer (CaCx) is strongly associated with human papillomavirus (HPV) infection, particularly HPV types 16 and 18. The constitutive expression of HPV E6 and E7 proteins in CaCx makes them attractive targets for CTL based immunotherapy. However cervical carcinomas may have features, e.g., antigen processing defects, that limit the effectiveness of HPV specific CTL. Furthermore most vaccine development has concentrated on HPV type 16, and it is not clear whether such vaccines could induce CTL able to cross-react on related oncogenic HPV types, e.g., HPV31 and 52. To investigate these potentially important parameters in vitro, we used a CTL (D4) specific for HPV16 E7(11-20). D4 was able to kill a variety of HPV16+ CaCx cell lines including those with suspected (CaSki) or known antigen processing defects (C33A), and with low HPV DNA copy number (SiHa). D4 was also able to cross react on a related peptide from HPV52 E7 but not HPV31 E7. Further analysis suggested that D4 cross reactivity against related peptides was influenced both by TCR contact residues and a certain threshold for peptide binding. The HPV cross-reactivity was confirmed at the whole protein level as D4 was also able to recognize the endogenously processed forms of HPV16 and 52 E7 but not 31 E7. These results suggest that HPV16 E7(11-20) would be a useful epitope for immunotherapy in both HPV 16 and 52 tumours. Despite this, it is difficult to generate these CTL in response to vaccination, emphasizing the need for definition of novel epitopes and more efficient vaccination strategies.

Figure 1

CTL against HPV16 E7_11–20 (D4) can kill cervical carcinoma cell lines with antigen processing defects and low levels of HPV16 unlike CTL against HPV16 E6_29–38 (7E7). The 2 CTL lines D4 (a) and 7E7 (b) were assessed for their ability to kill a variety of carcinoma cell lines in a 4 hr chromium release assay. (a) The specificity of D4 for the following targets: C1R‐A2 transfectants (closed square), C1R‐A2 transfectants pulsed with the HPV16 E7_11–20 peptide (10 μg/ml, closed diamond), the HPV16 transformed cervical carcinoma cell line CaSki (closed triangle), the cervical epithelial cell line C33A (open circle), C33A transfected with HPV16 E6 and E7 (closed circle) and SiHa that had been infected with a recombinant vaccinia virus to express HLA A*0201 (open triangle). (b) The specificity of 7E7 for CaSki cells (closed triangle), CaSki cells pulsed with the HPV16 E6_29–‐38 peptide (10 μg/ml, ‐ ‐closed triangle‐ ‐), C33A HPV16 cells (closed circle), SiHa cells transfected to express HLA A*0201 (×), SiHa cells transfected to express HLA A*0201and pulsed with the HPV16 E6_29–38 (10 μg/ml) (‐ ‐×‐ ‐). Percentage specific lysis was calculated as follows: 100 × [(specific release − spontaneous release)/(maximal release − spontaneous release)].

Figure 2

D4 CTL discriminates between the E7 peptides of related HPV types. A 4 hr chromium release assay was used to asses the ability of D4 to kill C1R‐A2 cells pulsed for 2 hr with varying doses of the 11–20 peptide from the following HPV types: HPV16 (YMLDLQPETT) closed diamond, HPV31 (YVLDLQPEAT) closed square or HPV52 (YIDLQPETT) closed triangle. In all cases an E:T of 10:1 was used.

Figure 3

Position 9 is important for the recognition of the E7_11–20 peptide. A 4 hr chromium release assay was used to assess the importance of the amino acid at position 9 in the peptide sequence of E7_11–20 peptides. D4 was used to kill C1R‐A2 cells pulsed for 2 hr with varying concentrations of the analogue peptides comprising of the HPV31 and 52 E7_11–20 sequences. The peptides and sequences were HPV31/52 analogue, YVLDLQPETT (closed circle); HPV52/31 analogue, YILDLQPEAT (open circle) and an analogue of the HPV16 E7 peptide containing an alanine at position 9, YMLDLQPEAT (open triangle). In all cases and E:T of 10:1 was used. The killing of C1R‐A2 cells pulsed with the HPV16 E7_11–20 peptide (YMLDLQPETT) at the same concentration is also shown (closed diamond).

Figure 4

Alterations in the amino acid sequence of the E7 peptide can affect binding to HLA‐A*0201. The affect of amino acid substitutions in the E7 peptides on binding to HLA‐A*0201 and the stability of this bond was assessed in a peptide binding assay using the antigen processing defective cell line T2. These cells have low levels of surface HLA‐A*0201, which increases when a peptide binds stably. Briefly cells were incubated with the various peptides at 10 μg/ml before staining with MA2.1, a monoclonal antibody specific to HLA‐A*0201. The influenza matrix peptide (M1,_58–66), which is known to bind strongly to this HLA allele, was used as a positive control. *Peptides that bind to HLA‐A*0201 stably for greater than 4 hr.

Figure 5

D4 can kill transfectants expressing the full‐length E7 of HPV types 16 and 52.The CTL D4 was assessed for its ability to kill transfectants expressing full length E7 of HPV types 16, 31, 45 and 52. C33A cells were transfected using the liposomal transfection reagent DOTAP and selected by growing in media containing 2.5 μg/ml blasticidin. The number of antigen specific T cells was analyzed by IFNγ ELISPOT. Responder cells were incubated with the following stimulators: C33A; C33A HPV16, C33A transfected in the laboratory to express HPV16 (C33AHPV16A), HPV31, HPV45 and HPV52. The ELISPOT was developed after 18 hr and cells were mixed at ratio of 1:1 with a total number of 20,000 responders/well. *Transfectants that stimulated CTL to produce a significantly higher amount of IFNγ compared to wild‐type C33A cells incubated with D4. Under the same culture conditions, incubation of D4 with the HPV16 E7_11–20 peptide at 10 μg/ml produced 404 spots and with a positive control cocktail containing PHA, PMA, concavalin A and ionomycin more than 500 spots was seen (data not shown).