HLA-G transcripts and isoforms and their uses

Abstract

Embodiments of the disclosure pertain to the field of HLA-G molecules and their therapeutic use. The disclosure pertains to new HLA-G isoforms, that is to say new RNA transcripts and proteins deriving from the HLA-G gene, pharmaceutical composition comprising thereof, as well as primers specific of these transcripts and antibodies specific of these proteins. The disclosure further pertains to the diagnostic or therapeutic use of these molecules.

Claims

1. A method of treating ischemia, comprising administering to a subject in need thereof a therapeutically effective amount of an isolated HLA-G protein, wherein the sequence of the isolated HLA-G protein is devoid of transmembrane/cytoplasmic domain and comprises or consists of a sequence selected from the group consisting of SEQ ID NO: 8, SEQ ID NO: 10, SEQ ID NO: 12, SEQ ID NO: 14, SEQ ID NO: 16, SEQ ID NO: 18, SEQ ID NO: 20, SEQ ID NO: 21, SEQ ID NO: 23, SEQ ID NO: 25, SEQ ID NO: 27, SEQ ID NO: 28, and SEQ ID NO: 30.

2. The method of claim 1, wherein the ischemia is ischemia associated with a cardiovascular disease, a peripheral artery disease or stroke.

Description

LEGEND OF THE FIGURES

(1) FIG. 1—Schematic representation of the structure of the HLA-G gene. A. IMGT/HLA nomenclature (top) and Ensembl database (bottom). Numbers represent exons and the domains of the HLA-G protein are shown underneath. TM: transmembrane; CT: cytoplasmic tail. B Localization of primers used for the different RT-PCR strategies. Sizes, in bp, for specific amplicons and the translation initiation codons are indicated.

(2) FIG. 2—Differential morphologic and HLA-G staining patterns of eight ccRCC included in this study. A trophoblastic tissue was used as positive control for immunohistochemical study (H&E and immunoperoxidase stains).

(3) FIG. 3—Expression of HLA-G1 in ccRCC patients. RNA were subjected to RT-PCR using the HLA-G1 specific primers G257F and G526R (upper panels) and ACTB primers as controls (lower panels). Lanes 1: adjacent non-tumor region except for tumors of patients 6 and 8. Lanes 2, 3 and 4: different tumor areas. For patients 6 and 8, all regions shown correspond to tumor areas since partial nephrectomies were performed and adjacent tumor regions were not available. M: 100 bp size marker.

(4) FIG. 4—Intron retention events found in HLA-G transcripts. Only reads spanning intron-exon junctions have been considered. Reads corresponding exclusively to intron sequences were discarded.

(5) FIG. 5—Molecular validation of main intron retention events. A: Diagrammatic representation of the RT-PCR strategy developed to amplify retained introns. B: Results of the RT-PCR analysis using actin primers as control for the absence of genomic DNA (left) and Int1 and G257R primers to detect the presence of intron 1 (right). The band of 523 bp reveals the absence of intron 2, which would produce a band of 649 bp C: HLA-G transcripts that retain only intron 4 (left panel) or HLA-G transcripts that retain several introns simultaneously (middle and right panels).

(6) FIG. 6—Identification of the 5′-extended transcript HLA-G1. A. Detail of the DNA sequence showing the reduced distance between the two ATGs. The sequence was performed upwards using the G526R primer B. Schematic representation of the 106 bp-deletion; the two ATG are underlined.

(7) FIG. 7: A: Pictures of NSG mice xenografted with RCC7 cells infected with a viral vector encoding the long HLA-G1, taken on day 38 after injection. B: Pictures of NSG mice xenografted with RCC7 cells infected with a viral vector encoding the long HLA-G1L, taken on day 38 after injection.

(8) FIG. 8: A: Pictures of nude mice xenografted with RCC7 cells expressing either GFP, HLA-G1 or HLA-G1L, on day 25 after intradermal injection. B: Pictures of nude mice xenografted with RCC7 cells grown on matrigel and expressing either GFP, HLA-G1 or HLA-G1L, on day 25 after intradermal injection.

(9) FIG. 9: Pictures of NSG immunodeficient mice 8 days after injection of a control RCC7 cells (expressing GFP) in the left ear, and of RCC7 cells expressing HLA-G1L in the right ear (A and B).

(10) FIG. 10 Pictures of NSG immunodeficient mice 8 days after injection of control RCC7 cells in the left ear, and of RCC7 cells expressing HLA-G1 in the right ear (A and B).

EXAMPLES

(11) A. Detection and Analysis of New HLA-G Isoforms

(12) 1. Materials and Methods

(13) 1.1 Tumor and Patients

(14) All patients of this study underwent a radical nephrectomy for ccRCC as first therapeutic intervention in the urology department of Saint-Louis Hospital (Paris, France) from November 2014 to April 2015. The median tumor size was of 50 mm (range, 35 to 175). According to the 2010 primary tumor TNM classification, these tumors were classified as pT1a (patient 6), pT1b (patients 1, 3, and 8), and pT3a (patients 2, 4, 5, and 7). Two patients (patients 2 and 4) had visceral metastases at presentation. All these renal tumors were classified as ccRCC by an experienced uropathologist according to the WHO classification of tumors of the kidney [8]. All patients that participated to this study gave their free and informed writing consent. The study was approved by the institutional review boards of Saint-Louis Hospital, Paris.

(15) 1.2 Tumor Specimen Processing

(16) For each tumor and according to the tumor size, we isolated between 3 and 10 samples of 10×5×5 mm, representing the spatial extent and macroscopic intra-tumor heterogeneity. Half of each sample was snap frozen in liquid nitrogen within 1 h of clamping of the renal artery and the other half was used to perform histological analysis and was documented by photography. Regions that did not contain tumor cells on histopathological examination were also isolated as controls.

(17) 1.3 Immunohistochemistry

(18) An immunohistochemical study was performed for each tumor on 4-μm-thick, formalin-fixed and paraffin-embedded tumor tissue sections. The following murine antibodies were used: 4H84, an IgG1 recognizing an epitope located into the alpha1 domain common to all HLA-G isoforms (dilution 1/200, Santa Cruz Biotechnology, Santa Cruz, Calif.), and two antibodies 5A6G7 and 2A12 recognizing the epitope encoded by the retained intron 5 (Ensembl database) present in soluble HLA-G5 and -G6 isoforms (dilution 1/100, Exbio antibodies, Exbio Co., CR). The staining was performed on automated slide stainers from Roche (BenchMark ULTRA system, Tucson, Ariz.) using the OptiView DAB IHC Detection Kit (Roche), Cell Conditioning 1 (CC1) short or standard antigen retrieval, an antibody incubation time of 32 min at 37° C., ultraWash procedure, counterstaining with Hematoxylin II for 4 min and bluing reagent for 8 min. Positive and negative controls gave appropriate results for each procedure.

(19) The immunohistochemical analyses were performed by the uropathologist using a BX51 microscope (Olympus France S.A.S, Rungis). Each immunostaining was scored on the basis of membranous and/or cytoplasmic staining by both intensity of staining as negative, weak, moderate, or strong and distribution of staining as negative (0% of tumor area), minimal (0-10% of tumor area), focal (<50% of tumor area), or diffuse (>50% of tumor area). A trophoblastic tissue was used as the positive control and isotype-specific immunoglobulins were used for negative controls with each run.

(20) 1.4 Trophoblast Sample Preparation

(21) Trophoblastic tissues were obtained from abortions (less than three months of pregnancy). After mechanical dissociation, the samples were preserved in Trizol™ Reagent (LifeTech, ref. 15596-026) at −80° C. until RNA extraction using the protocol described below.

(22) 1.5 RNA Extraction

(23) Total RNA was isolated from tissue sections manually crushed in Trizol™ Reagent (LifeTechnologie, ref. 15596026). After chloroform separation, the RNA was purified using miRNeasy mini Kit (Qiagen, ref. 217004) according to the manufacturer's instruction, with a DNase treatment extra step (Qiagen, ref. 79254). The RNA purity and concentration was assessed using a Nanodrop spectrophotometer and the Agilent 2100 Bioanalyzer System. RNA Integrity Number (RIN) values were mostly >8.

(24) 1.6 RT-PCR

(25) Reverse transcription of RNA into cDNA was perfomed using GoScript Reverse Transcriptase kit (Promega, ref. A5001) with a thermocycler Eppendorf (MasterCycler, Pro S). The PCR reactions were carried out in a final volume of 10 μL, containing 2 μL of cDNA template, using an ampliTaq polymerase from LifeTech (Ref. N80800166). For amplification, 40 cycles (at 94° C. for 30 sec, 55 or 60° C. for 30 sec, and 72° C. for 30 sec) were conducted. HLA-G and actin (ATCB) primers are described in Table 1. ATCB amplification was performed as control in all the experiments. The PCR amplification product was mixed with 6× loading dye (Promega, ref. G1881) and analyzed on 2% agarose gel stained with 2 μL of ethidium bromide at 1 mg/mL for 100 mL of agarose gel. The molecular weight marker used was 1 Kb plus DNA ladder from Invitrogen (Ref. 10787018). Imaging was performed using a ChemiDoc XRS System (Biorad), and interpretation using ImageLab software (Biorad).

(26) TABLE-US-00001 TABLE 1 PCR primers for RT-PCR experiments SEQ  Gene ID NO: Sequence (5′ to 3′) PrPr F 77 5′-GTAACATAGTGTGGTACTTTG Ex1F 76 5′-CCTGGACTCACACGGAAACT E2 F 81 5′-GGACTCATTCTCCCCAGACG 257 F 82 5′-GGAAGAGGAGACACGGAACA 257 R 83 5′-TGTTCCGTGTCTCCTCTTCC 526 F 84 5′-CCAATGTGGCTGAACAAAGG 526 R 85 5′-CCTTTGTTCAGCCACATTGG 963 R 86 5′-GCAGCTCCAGTGACTACAGC Int1 F 89 5′-GGCCTCAAGCGTGGCTCTCA Int3 F 78 5′-CCCAAGGCGCCTTTACCAAA Int4 R 75 5′-CCACTGCCCCTGGTAC Int5 R 79 5′-AGCCCTCACCACCGACC ATCB F 87 5′-TCCTGTGGCATCCACGAAACT ATCB R 88 5′-GAAGCATTTGCGGTGGACGAT
1.7 RNA Sequencing

(27) Indexed complementary DNA libraries were prepared from 1 μg of total RNA following the Illumina TRUSEQ protocol. Average size of the AMPure XP beads (Beckman Coulter, Inc.) purified PCR products was 275 bp. The paired-end 150 bp reads sequencing of the transcriptome was performed on equimolar pools of four cDNA libraries on a NextSeq 500 (ILLUMINA).

(28) 1.8 High-Throughput Analysis of HLA-G Isoforms

(29) The Ensembl nomenclature will be used throughout the text. Short reads from NGS sequencing were mapped to human Reference Genome NCBI Hg19 using BWA aligner (BWA MEM option) [20]. Low quality mapping reads were filtered out from alignment files and the reads mapping to the HLA-G locus were extracted using samtools (Li et al., 2009). Intron retained detection was performed by selecting reads overlapping an intron and one of the surrounding exons, retention for an intron was assessed only when we detected reads overlapping both 5′ and 3′ flanking exons. Exon skipping detection was performed by analyzing reads presenting split mapping, searching for discontinuity in the order of mapped exons, eg: a read that is mapped to exon the end of 4 and start of exon 6 but is not mapped to exon 5, presents a skipping of exons. Each read subset was visually validated with IGV [22]. For the retention of intron n, the percentage of reads pni supporting the event is calculated as the ratio between the reads supporting the events (reads at junction exon n/intron n, internal intronic reads on intron n and reads at junction intron n/exon n+1) and the total number of reads spanning the region where the event occurs (the region starting from the junction between exon n and intron n to the junction between intron n and exon n+1): Let R(i) be the number of reads strictly in region i (the reads are only in region i and do not overlap with other regions) and R(i, j) be the number of reads overlapping both regions i and j. Let S(i) be the number of reads supporting a skipping of exon i (reads overlapping exon n and exon m where m>n+1). The number of reads supporting the retention of intron n is thus IRn=R(exonn, intronn)+R(intronn)+R(intronn, exonn+1). The total number of reads in the region of the retention of intron n is Tn=IRn+R(exonn, exonn+1)+S(n); pni is thus given by pni=IRn/Tn.

(30) For the skipping of exon n, the percentage of reads pne supporting the event is given by pne=S(n)/Tn. Analysis of potential biases were assessed by using the TopHat2 aligner [24].

(31) 2. Results

(32) 2.1 Marked Subcellular Heterogeneity of HLA-G Isoforms Distribution in ccRCC

(33) In order to consider HLA-G as a potential target for cancer therapy, the expression of HLA-G in tumor cells derived from patients with ccRCC was assessed. To this end, 3 to 10 sections for each tumor were isolated, according to the tumor size. Microscopy analysis performed on hematoxylin and eosin (H&E) stained slides confirmed a morphologic heterogeneity (FIG. 2, left panel), classically associated with ccRCC [8]. We further dissected this heterogeneity by immunostaining with specific antibodies directed against HLA-G: 4H84, which recognizes an epitope located into the alpha1 domain common to all seven reported HLA-G isoforms and the antibody 5A6G7 that only recognizes soluble HLA-G5 and HLA-G6 isoforms. This antibody targets the amino acids encoded by the retained intron 5 (previously known as intron 4 according to the IMGT/HLA nomenclature). Trophoblastic cells, which express HLA-G at high levels, were used as positive controls.

(34) Even though all tumors expressed HLA-G in at least one area, this expression was distinct between and inside tumors. Tumors of patients 1 and 2 showed a strong immunostaining with 4H84 antibody in all regions. The staining was membranous and cytoplasmic (FIG. 2). Noteworthy, an additional very strong staining of hyaline globules located in the cytoplasm of the tumor cells was also detected. These hyaline globules were well visible on H&E slides and constituted a very uncommon aspect of tumor [9]. On the other hand, using the 5A6G7 antibody, a weak or moderate granular cytoplasmic immunostaining was noticed in the cytoplasm but not in hyaline globules. The expression of HLA-G in tumors from other patients was very different: tumors of patients 6 and 7 presented a diffuse but moderate membrane immunostaining with 4H84 antibody. These two tumors showed no (patient 6) or weak and focal (patient 7) granular intracytoplasmic immunostaining with 5A6G7 which denoted the absence of soluble proteins HLA-G5 and HLA-G6. In two other tumors (patients 4 and 5), the expression of HLA-G evaluated by 4H84 antibody was noted in small microscopic areas of only one tumor region. Of note, the only HLA-G positive area of patient 4's tumor corresponds precisely to intracytoplasmic hyaline globules. No stain was observed in any other region of the tumor.

(35) The immunostaining profiles of tumor cells of patients 3 and 8 were unexpected. No immunostaining was detected with the 4H84 antibody which labels all the reported HLA-G isoforms. The lack of labeling of tumor sections with this antibody normally accounts for the absence of HLA-G expression. However, a diffuse and strong granular intracytoplasmic 5A6G7 immunostaining, and a diffuse, thin and granular intracytoplasmic immunostaining were observed in tumor cells of patients 3 and 8, respectively. This was unpredictable considering our current knowledge on the structure of the seven reported HLA-G isoforms since they all contain the alpha 1 domain recognized by the 4H84 antibody. To try to better understand these differences, we have performed a similar analysis using an antibody that also recognizes the epitope encoded by the retained intron 5 (Ensembl database) present in soluble HLA-G5 and -G6 isoforms named 2A12. The results revealed different and unanticipated immune-staining patterns, notably the labeling of hyaline globules in patients 1 and 2.

(36) Together, the results of the immunohistochemical study clearly demonstrate intra- and inter-heterogeneity of HLA-G expression in ccRCC tumors. However, some immuno-staining patterns were unexpected within the boundaries of our prevailing knowledge on the structure of HLA-G isoforms.

(37) 2.2 Survey of HLA-G1 Transcripts Expressed in ccRCC

(38) To gain a better insight into the HLA-G isoforms that are expressed in ccRCC and clarify the results of the immunohistochemical analysis, a survey of HLA-G isoform diversity was further assessed by RT-PCR. The tumor sections of the eight patients studied above were amplified with the well-known G257F and G526R primers [10] schematically represented in FIG. 1B. These primers amplify a region that contains the epitope recognized by the 4H84 antibody. Amplification of actin mRNA was performed for each sample as control. A predicted band of 290 bp, specific for the amplification of HLA-G1 transcripts, was found in all tumor sections for patients 1, 2 and 6 whereas this band was only detected in one or two regions of tumors of other patients (FIG. 3). No amplification products were detected in non-tumoral adjacent tissues. Since the sequence of the different isoforms are highly similar and these RT-PCR conditions do not allow the identification of other isoforms like HLA-G2, -G3, -G6 or -G7 which lack exon 4, the target of primer G526, we undertook a large-scale study by RNAseq in order to provide a comprehensive picture of isoforms expressed in ccRCC.

(39) 2.3 RNA-Seq Reveals Unannotated HLA-G Transcripts

(40) RNAseq technology provides the most powerful method to analyze expressed isoforms, offering the opportunity to detect alternative splicing events and unannotated transcripts which are essential for understanding development and disease mechanisms in a species [25].

(41) As a first look, we have undertaken the sequencing of four representative samples at a very high depth of coverage (depth>300×). Reads were aligned and quantified according to the Ensembl 70 (GRCh37.p8) reference annotation as described in Material and Methods. Alternative spliced isoforms were mainly categorized into two major groups: exon skipping and intron retention, in which a single exon or intron is alternatively spliced or included out of the mature message.

(42) To verify whether the HLA-G expression patterns of ccRCC patients described above constitute a representative subset of general profiles found in ccRCC patients, we have compared our results to those obtained for the “Cancer Genome of the Kidney” (CAGEKID) cohort which includes a hundred ccRCC patients that were treated in four different European countries (Czech Republic, United Kingdom, Romania and Russia). The data that have been generated constitute a high-quality resource that allowed detecting alternative splicing events with high accuracy (Scelo et al., 2014). Moreover, we have deeply assessed whether common factors such as the choice of the aligner for RNAseq data or the reference sequence to study HLA-G might potentially bias our analysis by using two different aligners, BWA MEM and TopHat2. The results confirmed that the data aligned with BWA MEM or TopHat2 produce similar results (supplementary data). Further, the count of reads at the individual level showed a great similarity between the expression profiles of HLA-G transcripts found in our small cohort of ccRCC patients and that of Cagekid. These results are summarized on Tables 2 and 3 and will be discussed more thoroughly in the following sections.

(43) 2.4 Undescribed Intron Retention Events in Expressed HLA-G Transcripts

(44) Intron retention is the rarest type of alternative splicing in mammals and account for only approximately 3% of alternate transcripts [12]. So far, only the retention of intron 3 or intron 5 (previously known as intron 2 and intron 4, according to IMGT/HLA nomenclature) was reported in literature for HLA-G transcripts. Transcripts that retain intron 3 encode HLA-G7 [13] and those retaining intron 5 encode HLA-G5 and HLA-G6 [7].

(45) In our RNAseq analysis, introns subsumed by an exon were labeled as retained. The results, represented graphically on FIG. 4 and summarized in Table 2, showed that reads representing the retention of introns 3 and 5 were the most abundant. In addition, the data support a number of overall new findings that originate from the retention of four additional introns: 1, 4, 6 and 7. To validate the expression of intron-retained transcripts, we first looked for the presence of transcripts containing the intron 1. To this end, we performed RT-PCR amplifications using a strategy described in FIG. 5. First, primer that targets intron 1 (Int1F) was used in combination with G257R, the reverse primer of G257F [13]. Since the presence of introns may be due to contaminating endogenous genomic DNA, all samples were amplified in parallel with actin specific primers located in two different exons. The expected size for the amplification of cDNA derived from mRNA is 320 bp whereas that of genomic DNA is 560 bp. The results show only the amplification of a 320 bp-fragment in all samples, demonstrating the absence of genomic contamination (FIG. 5B, left panel). In view of this result, we further amplified tumor samples using primers Int1F and G257R. An amplified band of the expected size (521 bp) was obtained, consistent with the presence of intron 1 in HLA-G transcripts (FIG. 5-B, right panel). This event was not reported before in literature since the initiation of transcription of HLA-G was solely assigned to exon 2 [26]. We did not detect a PCR amplification band of 649 bp that would correspond to the concomitant retention of intron 2. This is consistent with the results of the RNAseq analysis showing that intron 2 is infrequently retained.

(46) TABLE-US-00002 TABLE 2 Number of reads for all observed HLA-G splicing events in ccRCC samples Patient 1 Patient 3 Patient 4 Patient 5 B00E4I3 B00E4IS #reads q30 at HLA-G 4324 1353 238 142 6216 5066 locus (mean) exon1 total reads 6 6 0 0 0 0 exon2 total reads 120 2 0 5 39 15 exon3 total reads 1344 367 11 11 390 384 exon4 total reads 1483 260 37 19 1054 1078 exon5 total reads 1397 319 28 21 2002 1375 exon6 total reads 449 47 0 11 260 187 exon7 total reads 248 10 0 9 2 6 exon8 total reads 1079 676 16 25 1934 1503 retention of intron 1 40 4 1 0 38 12 retention of intron 2 0 3 3 0 4 1 retention of intron 3 2 101 38 0 36 71 retention of intron 4 133 28 2 0 31 84 retention of intron 5 148 28 8 0 179 87 retention of intron 6 37 46 2 7 35 67 retention of intron 7 119 47 0 0 454 96 skipping of exon 4 1 0 0 0 9 7 skipping of exon 5 0 0 0 0 0 0 skipping of exon 6 2 0 0 0 0 0 skipping of exon 7 21 1 0 0 29 31 skipping of exon 4 and 2 0 0 0 3 4 5 skipping of exon 4, 5, 6 0 0 0 0 0 0 and 7 skipping of exon 4, 5 0 0 0 0 0 0 and 7 skipping of exon 5, 6 0 0 0 0 0 0 and 7 skipping of exon 6 and 2 0 0 0 12 8 7 raw count of reads 132 15 0 0 0 0 start exon2 raw count of reads 0 5 0 0 1 23 start exon3 raw count of reads 0 3 0 0 10 8 start exon4 raw count of reads 291 67 6 5 10 3 start exon5 Patients 1, 3, 4 and 5 are representative samples selected for exploring the diversity of HLA-G isoforms. B00E4I3 and B00E4IS are the two samples with the highest HLA-G expression within the CAGEKID (CAncer GEnome of the KIDney) [14].

(47) Further analysis were conducted to validate the retention of intron 4 (FIG. 5C). To this end, RT-PCR was performed using primer G257F in combination with a primer that specifically targets intron 4 (named int4R). Amplification with these primers generated a DNA fragment of 430 bp (FIG. 5C, left panel), demonstrating the presence of intron 4 in HLA-G transcripts. The size of the amplified band is also consistent with the presence of a concomitant retention of intron 3. To further assess whether the same transcript might retain several introns simultaneously, we have performed a RT-PCR amplification using primer int3F (whose sequence is complementary to a region of intron 3) in combination with primer int4R. The results reveal a DNA fragment of 380 bp, as expected for the retention of introns 3 and 4 in the same transcript (FIG. 5C, middle panel). In addition, amplification with Int3F and Int5R primers generated an amplified band of 725 pb (FIG. 5C, right panel). Of note, the size of this band corresponds to the retention of intron 3 and 5, excluding intron 4. These results clearly demonstrate that tumor samples might express transcripts that retain a single intron and others that retain several different introns which may vary from one transcript to the other. To our knowledge, these events were not previously described.

(48) 2.5 Novel HLA-G Transcripts with 5′-Extended End

(49) The RNAseq data further revealed that some of the reads aligned on either side of exon 1 (FIG. 4). Transcripts that originate from this area were not previously reported. In fact, the structure of this region is still a matter of debate since information contained in the Ensembl database suggests that HLA-G transcripts may be initiated at this exon, which is located 5′ of the exon 1 defined by IMGT/HLA nomenclature (FIG. 1A). The presence or absence of this exon may result in major modifications which include the promoter localization, the length of the 5′-untranslated region and the transcription/translation initiation site. We assess whether HLA-G transcripts may be initiated in this exon or even upstream by RT-PCR amplification. Two specific primers were designed. Primer Ex1F, whose sequence is complementary to a region located in exon1 (Ensembl database) and primer PrPr, whose sequence is complementary to a region located further upstream currently considered as the promoter region (schematically represented in FIG. 1B). RT-PCR using these two upstream primers in combination with G526R produced two bands of expected sizes: 690 bp (for Ex1F-G526R) and 725 bp (for PrPr-G526R) respectively (data not shown). To verify the specificity of these fragments, amplified DNA samples were sequenced and nucleotide similarities were searched in public databases using BLAST. The results demonstrated a high degree of similarity with HLA-G except for a deletion of 106 bp fragment. Resulting from this deletion, the distance between the ATG located at the end of exon 1 and the one located in exon 2 was reduced from 118 bp to 12 bp (FIG. 6A). As a consequence, the 106-bp deletion brings both ATG in frame. This may now allow the initiation of translation at the ATG located in the first exon and generate a protein that would have a 5′-extended end of five additional amino acids (MKTPR (SEQ ID NO: 1)). At present, the only translation initiation start site was attributed to the ATG located in exon 2 (which corresponds to exon 1 defined by IMGT/HLA nomenclature). This transcript was also found in some of the trophoblast samples tested but not all. This indicates that factors regulating it expression are still to be elucidated.

(50) Altogether these results are consistent with the existence of a novel HLA-G transcript, named HLA-G1L, having an extended 5′-end, which might be co-expressed in trophoblasts and ccRCC tumor cells with previously reported HLA-G isoforms.

(51) 2.6 Alternatively Spliced Exons Potentially Generate Novel Soluble HLA-G Isoforms

(52) Exon skipping is one of the major forms of alternative splicing, which generates multiple mRNA isoforms differing in the precise combinations of their exon sequences. Here, we define an exon skipping event as a pairing between an exon-containing form and an exon-excluding form, occurring at the same exon and with the same flanking introns. The same exon may be involved in multiple exon skipping events.

(53) For HLA-G, only the skipping of exon 4 (HLA-G2), exon 5 (HLA-G4), or both simultaneously (HLA-G3), were reported in literature. In this study, aligned reads with BWA mem reveal the skipping of exons never uncovered before. The main skipping events are reported in Table 2. We also confirmed these results by using TopHat2.

(54) The highest read coverage was consistent with the skipping of exon 7 alone, which contains the stop codon of the protein. However, no major modifications are expected in the encoded protein lacking this exon since a supplementary in-frame stop codon is found at the beginning of exon 8. Most importantly, skipping of exon 7 concomitantly to exon 6, which encodes the transmembrane domain, is highly relevant since their absence may generate isoforms that lack the transmembrane domain and the cytoplasmic tail and therefore would constitute still unreported soluble proteins.

(55) When RT-PCR was performed with primer G963R, whose sequence is complementary to a region of exon 6, no amplification products could be obtained in combination with the forward primers G257F (exon 3) or G256F (exon 4). However, an expected 290 bp amplified fragment was generated when the primer G257F was used in combination with G526R. Together these results are consistent with HLA-G transcripts that possess exons 3 and 4 but lack exon 6. In addition, when these primers were used to analyze samples from patient 1, amplified bands were obtained using the primer combination G526F-G963R whereas no amplification was detected using G257F-G963R, consistent with the expression of transcripts that lack exon 3.

(56) 2.7 Alternative Spliced HLA-G Isoforms Lack the Alpha-1 Domain

(57) Further analysis of RNAseq data reveals that some of the reads might be initiated at exon 4. This was determined by quantifying the raw count of reads within 20 pb upstream of the exon acceptor site. The predicted N-terminal-truncated protein would lack the peptide signal and the alpha1 domain. To assess whether the translation into a protein might start in this region, we have examined the nucleotide sequence of exon 4. This analysis revealed the presence of an in-frame ATG that might serve as a translation initiation codon. Our preliminary results (not shown) reveal that transcripts that lack the alpha-1 domain may lack also the alpha-2 domain and therefore encode only the alpha-3 domain.

(58) Notably, the expression of these isoforms may now provide a hypothesis on the differences of immuno-staining patterns generated following the labeling of some tumor samples with 4H84 and antibodies that have been raised against soluble isoforms, which could not be explained previously within the boundaries of widespread knowledge on the structure of HLA-G isoforms.

(59) TABLE-US-00003 TABLE 3 Percentage of transcripts for each splicing event observed Alternative splicing % overall % overall Samples median events Samples median CAGEKID CAGEKID retention of intron 1 50 100 25.97 100 retention of intron 2 25 0 41.56 0 retention of intron 3 50 43.9 85.71 8.62 retention of intron 4 50 84.85 75.32 13.89 retention of intron 5 75 82.35 92.21 15.62 retention of intron 6 75 70.66 90.91 17.02 retention of intron 7 50 85.62 90.91 54.17 skipping of exon 4 0 0 38.96 0 skipping of exon 6 0 0 31.17 0 skipping of exon 7 50 7.66 81.82 21.23 skipping of exon 6 0 0 62.34 2.25 and 7 Percentage of overall samples is the percentage of samples presenting the event. The last two columns are the same metrics calculated for 77 CAGEKID samples expressing HLA-G.
B. Pro-Tumoral Effect of the HLA-G Isoforms
1. Materials and Methods
1.1 Production of Lentiviruses Expressing HLA-G Isoforms

(60) The HLA-G1 and HLA-G1L isoforms were introduced into the plasmid pWPXL (10510 bp), between the BamH1 (3499) and NdeI (4334) sites, just 3′ of the EF-1α promoter which directs the expression of two isoforms HLA-G.

(61) For HLA-G1:

(62) The inserted fragment of 3438 bp comprises the HLA-G1 cDNA initiated in the SEQ ID NO. 93 AGTGTGGTACTTT sequence and ending in 3′ with the SEQ ID NO. 94 TGGAAGACATGAGAACTTTCCA sequence. This fragment is followed by a “red” variant of the GFP (Aequorea victoria green fluorescent protein jellyfish), named Neptune that has been brought under control of the CMV promoter. Finally, at the 3′ end, a molecular barcode was introduced as an integration marker and for in vivo monitoring of metastases (Grosselin et al., Stem Cells, 10: 2162-71, 2013).

(63) For HLA-G1L:

(64) The inserted fragment of 3279 bp comprises the HLA-G1L cDNA initiated at the SEQ ID NO. 95 ATATAGTAACATAGTGT sequence and ending in 3′ with the SEQ ID NO. 94 TGGAAGACATGAGAACTTTCCA sequence. This fragment is followed by a “blue” (cyan) variant of GFP, the ECFP which has a bimodal excitation and emission spectrum at 433/445 nm and 475/503 nm leading to a fluorochrome with a gloss and improved photostability. ECFP was put under control of the CMV promoter. Finally, at the 3′ end, a molecular barcode was introduced as integration marker and for in vivo monitoring of metastases (Grosselin et al., Stem Cells, 10: 2162-71, 2013).

(65) These 2 plasmids were used to produce lentivirus WPXL ΔU3 SIN, envelope VSV-G, OGM group II, class 2 at 1.20E+08 TU (Transduction Unit)/ml.

(66) 2. Results

(67) Each lentivirus contains a different HLA-G isoform. The lentiviruses were transduced in a line of renal cell carcinoma clear (cells RCC7) lineage perfectly characterized, not expressing HLA-G. For each isoform, two independent transductions were performed to increase the reliability and robustness of our results.

(68) An intradermal injection was performed of each of the RCC7 cell lines transduced, into 5 NSG mice per condition. Non-transduced RCC7 cells are used as a control.

(69) After intradermal injection of the cells, tumor/metastatic growth was evaluated regularly.

(70) At the time of sacrifice of the mice, tumors metastases, and different tissues were removed, for immunohistochemical and expression (RNA) analysis. Each isoform is associated with a barcode, making it possible to ensure that the tumors and metastases obtained come from the injected cells.

(71) As can be seen in FIG. 7, the xenografted mice from the cells from the RCC7 line bearing the long HLA-G1L isoform showed at J 38, a more marked tumor growth, at least partially linked to a more developed intra and peritumoral neovascularization. No intra-tumoral necrotic reworking is observed in these tumors, unlike those resulting from the RCC7 line carrying the HLA-G1 isoform.

(72) Similar experiments were done with nude mice. FIG. 8A shows the pictures of nude mice xenografted with RCC7 cells expressing either GFP, HLA-G1 or HLA-G1L, on day 25 after injection. FIG. 8B shows the pictures of nude mice xenografted with RCC7 cells grown on matrigel and expressing either GFP, HLA-G1 or HLA-G1L, on day 25 after injection.

(73) C. Pro-Angiogenic Effect of the HLA-G Isoforms

(74) RCC7 cells expressing either GFP, HLA-G1 or HLA-G1L were prepared as disclosed above (point B).

(75) The left ear of NSG mice were injected with control (RCC7 cells expressing GFP), while their right ear were injected with RCC7 cells expressing either HLA-G1 or HLA-G1L. Pictures were taken on day 8. The results are shown in FIG. 9 (control RCC7 cells vs HLA-G1L) and in FIG. 10 (control RCC7 cells vs HLA-G1).

(76) The results demonstrate a pro-angiogenic effect of the expression of HLA-G1L, which is not reproduced by the expression of HLA-G1.

REFERENCES

(77) 1. Rouas-Freiss, N., et al., Direct evidence to support the role of HLA-G in protecting the fetus from maternal uterine natural killer cytolysis. Proc Natl Acad Sci USA, 1997. 94(21): p. 11520-5. 2. Ibrahim, E. C., et al., Tumor-specific up-regulation of the nonclassical class I HLA-G antigen expression in renal carcinoma. Cancer Res, 2001. 61(18): p. 6838-45. 3. Bukur, J., et al., Functional role of human leukocyte antigen-G up-regulation in renal cell carcinoma. Cancer Res, 2003. 63(14): p. 4107-11. 4. Brugarolas, J., Molecular genetics of clear-cell renal cell carcinoma. J Clin Oncol, 2014. 32(18): p. 1968-76. 5. Agaugue, S., E. D. Carosella, and N. Rouas-Freiss, Role of HLA-G in tumor escape through expansion of myeloid-derived suppressor cells and cytokinic balance in favor of Th2 versus Th1/Th17. Blood, 2011. 117(26): p. 7021-31. 6. Loumagne, L., et al., In vivo evidence that secretion of HLA-G by immunogenic tumor cells allows their evasion from immunosurveillance. Int J Cancer, 2014. 135(9): p. 2107-17. 7. Fujii, T., A. Ishitani, and D. E. Geraghty, A soluble form of the HLA-G antigen is encoded by a messenger ribonucleic acid containing intron 4. J Immunol, 1994. 153(12): p. 5516-24. 8. Moch, H., et al., The 2016 WHO Classification of Tumours of the Urinary System and Male Genital Organs-Part A: Renal, Penile, and Testicular Tumours. Eur Urol, 2016. 70(1): p. 93-105. 9. Krishnan, B. and L. D. Truong, Renal epithelial neoplasms: the diagnostic implications of electron microscopic study in 55 cases. Hum Pathol, 2002. 33(1): p. 68-79. 10. Paul, P., et al., HLA-G, -E, -F preworkshop: tools and protocols for analysis of non-classical class I genes transcription and protein expression. Hum Immunol, 2000. 61(11): p. 1177-95. 11. Woolard, J., et al., Molecular diversity of VEGF-A as a regulator of its biological activity. Microcirculation, 2009. 16(7): p. 572-92. 12. Wong, J. J., et al., Intron retention in mRNA: No longer nonsense: Known and putative roles of intron retention in normal and disease biology. Bioessays, 2016. 38(1): p. 41-9. 13. Paul, P., et al., Identification of HLA-G7 as a new splice variant of the HLA-G mRNA and expression of soluble HLA-G5, -G6, and -G7 transcripts in human transfected cells. Hum Immunol, 2000. 61(11): p. 1138-49. 14. Scelo, G., et al., Variation in genomic landscape of clear cell renal cell carcinoma across Europe. Nat Commun, 2014. 5: p. 5135. 15. Nayar, R., E. Bourtsos, and D. V. DeFrias, Hyaline globules in renal cell carcinoma and hepatocellular carcinoma. A clue or a diagnostic pitfall on fine-needle aspiration? Am J Clin Pathol, 2000. 114(4): p. 576-82. 16. Gerlinger, M., et al., Genomic architecture and evolution of clear cell renal cell carcinomas defined by multiregion sequencing. Nat Genet, 2014. 46(3): p. 225-33. 17. Carosella, E. D., et al., HLA-G and HLA-E: fundamental and pathophysiological aspects. Immunol Today, 2000. 21(11): p. 532-4. 18. Hoare, H. L., et al., Subtle changes in peptide conformation profoundly affect recognition of the non-classical MHC class I molecule HLA-E by the CD94-NKG2 natural killer cell receptors. J Mol Biol, 2008. 377(5): p. 1297-303. 19. Kraemer, T., et al., HLA-E: Presentation of a Broader Peptide Repertoire Impacts the Cellular Immune Response-Implications on HSCT Outcome. Stem Cells hit, 2015. 2015: p. 346714. 20. Li, H. and R. Durbin, Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics, 2009. 25(14): p. 1754-60. 21. Li, H., et al., The Sequence Alignment/Map format and SAMtools. Bioinformatics, 2009. 25(16): p. 2078-9. 22. Robinson, J. T., et al., Integrative genomics viewer. Nat Biotechnol, 2011. 29(1): p. 24-6. 23. Morandi, F., et al., Recent Advances in Our Understanding of HLA-G Biology: Lessons from a Wide Spectrum of Human Diseases; Journal of Immunology Research, vol. 2016, Article ID 4326495, 2016. 24. Kim, D., Pertea, G., Trapnell, C., Pimentel, H., Kelley, R., and Salzberg, S. L. (2013). TopHat2: accurate alignment of transcriptomes in the presence of insertions, deletions and gene fusions. Genome biology 14, R36. 25. Wang, E. T., Sandberg, R., Luo, S., Khrebtukova, I., Zhang, L., Mayr, C., Kingsmore, S. F., Schroth, G. P., and Burge, C. B. (2008). Alternative isoform regulation in human tissue transcriptomes. Nature 456, 470-476. 26. Geraghty, D. E., Koller, B. H., and Orr, H. T. (1987). A human major histocompatibility complex class I gene that encodes a protein with a shortened cytoplasmic segment. Proceedings of the National Academy of Sciences of the United States of America 84, 9145-9149. 27. Tatusova et al, “Blast 2 sequences—a new tool for comparing protein and nucleotide sequences”, FEMS Microbiol, 1999, Lett. 174:247-250. 28. Fons, P., Chabot, S., Cartwright, J. E., Lenfant, F., L'Faqihi, F., Giustiniani, J., Herault, J., Gueguen, G., Bono, F., Savi, P., Aguerre-Girr, M., Fournel, S., Malecaze, F., Bensussan, A., Plouët, J., & Le Bouteiller, P. (2006). Soluble HLA-G1 inhibits angiogenesis through an apoptotic pathway and by direct binding to CD160 receptor expressed by endothelial cells. Blood, 108(8), 2608-2615. 29. Tronik-Le Roux D, Renard J, Vérine J, et al. Novel landscape of HLA-G isoforms expressed in clear cell renal cell carcinoma patients. Molecular Oncology. 2017; 11(11):1561-1578.