lncRNA GAS8-AS1 genetic alterations in papillary thyroid carcinoma and their clinical significance

Abstract

The long non-coding RNA (lncRNA) GAS8-AS1 is the second-most frequently altered gene, following the BRAF gene, in papillary thyroid carcinoma (PTC). We aimed to study the specificity and significance of genetic alterations in GAS8-AS1 in PTC. In this study, we reported the prevalence of genetic alterations of GAS8-AS1 in tissues of 48 nodular goiter, 573 papillary thyroid cancer, 95 colorectal cancer, 101 non-small cell lung cancer, 92 glioma, and 69 gastrointestinal stromal tumor patients, and in peripheral white blood cells of 286 healthy volunteers. We observed that the genomic sequence of GAS8-AS1 had a high frequency of genetic alterations in addition to the previously reported c.713A $>$ G/714T $>$ C substitution. Substitution of c.713A $>$ G was completely linked with four other loci at c.714T $>$ C, c.728A $>$ G, c.737G $>$ A, and c.752G $>$ A. Two novel substitutions at c.749G $>$ A and c.826A $>$ G were also found. Interestingly, evidence from different samples indicated that these variations were not unique variants for PTC; they were also found in other malignant tissues and white blood cells of healthy volunteers. The c.713A $>$ G substitution was associated with the T stage of PTC, while c.749G $>$ A was more likely to occur in younger patients with PTC. PTC patients carrying heterozygous variants at the c.749 and c.826 loci had a higher risk of developing multiple lesions. These associations were also observed in patients with PTC and concomitant benign thyroid disease. Notably, the rare homozygous GG at the c.826 site conferred a higher risk of developing T2 PTC without benign thyroid disease, and a lower risk of developing T2 PTC with benign thyroid disease. Alterations of c.749G $>$ A and c.826A $>$ G had higher levels of serum TSH (thyroid stimulating hormone) in PTC subjects. Our study provides evidence that the detection of GAS8-AS1 genetic alterations would be useful in diagnostic screening and prognostic assessment of PTC.

1. Introduction

Papillary thyroid carcinoma (PTC) is the most common endocrine carcinoma with an increasing global incidence [1]. The most frequently affected hot spot in PTC is the BRAF V600E (p.Val600Glu) mutation [2]. Recently, another driver gene, long non-coding RNA GAS8 antisense RNA 1 (lncRNA GAS8-AS1), has been documented as the second-most frequently altered gene, following the BRAF gene in PTC [3]. GAS8-AS1 is located on chromosome 16 and is the RNA transcript of intron 2 of GAS8, which is a tumor suppressor gene [4]. Reported decrease in expression of GAS8-AS1 in PTC tissue [3] and plasma [5] has been associated with lymph node metastasis [5].

Two nucleotide substitutions, c.713A $>$ G and c.714T $>$ C of GAS8-AS1 that are in very strong linkage disequilibrium, have been reported to be associated with the progression of PTC and reduced RNA expression [3]. Both the wild-type (713A/714T) and variant-type alleles (713G/714C) play a role in tumor suppression, but variant-type alleles exert a weaker inhibitory effect [3]. Therefore, these two variations have been suggested as potential diagnostic biomarkers for PTC and novel prognosis predictors.

In this study, we investigated whether germline alterations, including c.713A $>$ G/c.714T $>$ C and their nearby loci, such as c.826G $>$ A and c.749G $>$ A, are specific variants in PTC. In addition to the PTC tissue, we explored the frequencies of these variations in different tumor tissues and blood white cells of healthy controls. We also focused on the association of the BRAF V600E mutation and concomitant benign thyroid disease with GAS-AS1 alterations in thyroid cancer. Moreover, we explored whether these loci affected thyroid function in patients with PTC.

2. Materials and methods

2.1 Participants and tumor samples

The protocols were approved by the Ethics Committee of the Sun Yat-Sen University Cancer Center (No. GZV01S-168), Guangdong, China. Written informed consent was obtained from all patients at their first visit. All subjects were unrelated individuals of Han Chinese descent enrolled between August 2017 and September 2018. The pathological diagnosis of the enrolled subjects was independently confirmed by two pathologists. Genetic variations were examined in tissues of 48 nodular goiter, 573 papillary thyroid cancer, 95 colorectal cancer, 101 non-small cell lung cancer, 92 glioma, and 69 gastrointestinal stromal tumor patients. In addition, genetic variations were detected in the peripheral white blood cells of 286 healthy volunteers.

2.2 Data collection

Medical records and surgical pathology reports were reviewed to obtain demographic parameters and pathological characteristics of the tumors for the enrolled patients. A questionnaire of basic demographic information was collected from each healthy volunteer. The pathological slides were independently reviewed by two pathologists to confirm the diagnosis. Concomitant diseases in PTC patients, such as nodular goiter and Hashimoto’s thyroiditis, were also reported. Information on primary tumor size, TNM (tumor-node-metastasis) stage, extrathyroid extension, and metastasis was assessed based on the latest version of the National Comprehensive Cancer Network on thyroid cancer recommendations (https://www.nccn.org/). The evaluation of thyroid function was performed at the Clinical Laboratory of Sun Yat-Sen University Cancer Center. Serum fT3 (free triiodothyronine 3), fT4 (free triiodothyronine 4), TSH (thyroid stimulating hormone), ATPO (anti-thyroid peroxidase antibody), PTH (parathyroid Hormone), TG (thyroglobulin), and anti-TG levels were measured before surgery.

2.3 Detection of genetic variations in GAS8-AS1, BRAF, and TERT

The genetic variations of GAS8-AS1 (c.713A $>$ G/ c.714T $>$ C, c.826G $>$ A, and c.749G $>$ A) and TERT (C228T and C250T) were detected using polymerase chain reaction (PCR) followed by Sanger sequencing according to the following conditions: for GAS8-AS1, 94 ${}^{\circ }$ C for 5 min, followed by 95 ${}^{\circ }$ C for 30 s, 64 ${}^{\circ }$ C for 30 s, and 72 ${}^{\circ }$ C for 30 s for 30 cycles, and finally, 72 ${}^{\circ }$ C for 10 min; and for TERT, 94 ${}^{\circ }$ C for 5 min, followed by 98 ${}^{\circ }$ C for 10 s, 56 ${}^{\circ }$ C for 15 s, and 68 ${}^{\circ }$ C for 1 min 30 s for 35 cycles, and finally, 72 ${}^{\circ }$ C for 10 min. The sequences of primers were as follows: for GAS8-AS1, forward primer: 5’-CCTGCCTGGACCCTGAAGA-3’; reverse primer: 5’-GCTGCTGGATGGAGAATGGA-3’; and for TERT, forward primer: 5’-GGCCGATTCGACCTC TCT-3’; reverse primer: 5’-AGCACCTCGCGGTAGTG G-3’. BRAF V600E was detected using ARMS-PCR (SINOMED, Beijing, China, #SMD-02-026) according to the manufacturer’s protocol. Amplification and analysis were performed on an ABI 7500 Real Time PCR System (Applied Biosystems, CA, USA).

Table 1

Association between GAS-AS1 genetic alterations and the risk of goiter and thyroid carcinoma

Gene variations	Healthy control		Goiter		$P$	PTC		$P$
c.713A $>$ G					0.226			0.289
AA	198	(69.2%)	29	(60.4%)		376	(64.9%)
AG	88	(30.8%)	19	(39.6%)		197	(34.0%)
c.749G $>$ A					0.249			0.035
GG	126	(63.6%)	17	(58.6%)		327	(70.8%)
GA	65	(32.8%)	9	(31.0%)		109	(23.6%)
AA	7	(3.5%)	3	(10.3%)		26	(5.6%)
c.826G $>$ A					0.034			0.394
AA	111	(56.1%)	12	(41.4%)		204	(35.2%)
GA	66	(33.3%)	9	(31.0%)		117	(20.2%)
GG	21	(10.6%)	8	(27.6%)		55	(9.5%)

Figure 1.

Substitution of c.713A $>$ G in complete linkage disequilibrium with c.714T $>$ C, c.728A $>$ G, c.737G $>$ A and c.752G $>$ A of GAS8-AS1. (A) homozygotes of c.713AA-714TT-728AA-737GG-752GG (B) heterozygotes of c.713AG-714TC-728AG-737GA-752GA.

2.4 Statistical analysis

All statistical analyses were performed using the SPSS 16.0 software (SPSS, Chicago, IL, USA). The distribution of qualitative variables was detected using a ${\chi }^2$ test. The distribution of quantitative variables between two groups was examined using Student’s $t$ -test. The corresponding variables are shown as means. Multifactorial logistic regression analysis using likelihood-ratio test was performed to assess the association strength between genotype and clinicopathological factors of PTC, adjusting for demographic parameters. A two-sided $P$ value of less than 0.05 was considered statistically significant.

The authenticity of this article has been validated by uploading the key raw data onto the Research Data Deposit public platform (www.researchdata.org.cn), with the approval RDD number RDDA2020001499.

3. Results

3.1 Specificities of GAS8-AS1 alterations in different samples

We analyzed the prevalence rates of GAS8-AS1 from different sample sources. We observed that substitution of c.713A $>$ G was in complete linkage disequilibrium with four other loci at c.714T $>$ C, c.728A $>$ G, c.737G $>$ A, and c.752G $>$ A (Fig. 1). We also found two other genetic variations, c.749G $>$ A and c.826G $>$ A at nearby loci (Fig. 2). The variations mentioned above were found in tissues of different tumors, as well as in peripheral white blood cells of healthy subjects, suggesting that these genetic variations were common germline variations (Supplementary Table 1). The genotype frequency distributions of these genetic variations did not show statistical differences among different tumor tissues. When compared with healthy subjects, the genotype frequencies of c.826G $>$ A and c.749G $>$ A were statistically different. The prevalence rates of GAS8-AS1 genetic variations are summarized in Table 1.

Table 2

Association between GAS-AS1 genetic alterations and geographical factors in thyroid cancer patients

Variable	A713G				$P$	A826G				$P$	G749A				$P$
	AA		AG			AA		GA $+$ GG			GG		GA $+$ AA
Sex					0.447					0.996					0.664
Male	118	(31.4%)	68	(34.5%)		140	(68.6%)	118	(68.6%)		216	(66.1%)	92	(68.1%)
Female	258	(68.6%)	129	(65.5%)		64	(31.4%)	54	(31.4%)		111	(33.9%)	43	(31.9%)
Age (yrs)					0.821					0.276					0.035
$⩾$ 55	23	(6.1%)	13	(6.6%)		15	(7.4%)	8	(4.7%)		31	(9.5%)	5	(3.7%)
$<$ 55	353	(93.9%)	184	(93.4%)		189	(92.6%)	164	(95.3%)		296	(90.5%)	130	(96.3%)
Family history of tumor					0.083					0.590					0.739
No	321	(85.4%)	157	(79.7%)		176	(86.3%)	145	(84.3%)		272	(83.2%)	114	(84.4%)
Yes	55	(14.6%)	40	(20.3%)		28	(13.7%)	27	(15.7%)		55	(16.8%)	21	(15.6%)
Smoking					0.795					0.179					0.336
Never	349	(92.8%)	184	(93.4%)		186	(91.2%)	163	(94.8%)		299	(91.4%)	127	(94.1%)
Ever	27	(7.2%)	13	(6.6%)		18	(8.8%)	9	(5.2%)		28	(8.6%)	8	(5.9%)
Alcohol consumption					0.957					0.402					0.649
No	372	(98.9%)	195	(99.0%)		201	(98.5%)	171	(99.4%)		323	(98.8%)	134	(99.3%)
Yes	4	(1.1%)	2	(1.0%)		3	(1.5%)	1	(0.6%)		4	(1.2%)	1	(0.7%)

Figure 2.

Genotypes of c.749G $>$ A and c.826G $>$ A of GAS8-AS1. Genotypes of c.749 (A) GG, (B) GA, and (C) AA; and genotypes of c.826 (D) GG, (E) GA, and (F) AA.

Table 3

Association between GAS-AS1 genetic alterations and clinicopatholoigic characteristics in thyroid cancer patients

Variants	Genotype	T stage		OR (95% CI)	$P$	N stage		OR (95% CI)	$P$	No. of lesions		OR (95% CI)	$P$
		T1	T2–4			N0	N1			Single	Multiple
For A713G
Total PTC subjects	AA	283 (63.3%)	77 (75.5%)	1 (ref)		148 (68.5%)	212 (63.7%)	1 (ref)		203 (68.1%)	157 (62.5%)	1 (ref)
	AG	164 (36.7%)	25 (24.5%)	0.54 (0.33–0.88)	0.014	68 (31.5%)	121 (36.3%)	1.17 (0.81–1.70)	0.411	95 (31.9%)	94 (37.5%)	1.30 (0.91–1.85)	0.152
BRAF-wild type PTC	AA	71 (58.2%)	24 (82.8%)	1 (ref)		42 (65.6%)	53 (60.9%)	1 (ref)		59 (64.1%)	36 (61.0%)	1 (ref)
	AG	51 (41.8%)	5 (17.2%)	0.25 (0.09–0.72)	0.010	22 (34.4%)	34 (39.1%)	1.14 (0.56–2.33)	0.722	33 (35.9%)	23 (39.0%)	1.14 (0.57–2.26)	0.718
BRAF-mutant PTC	AA	202 (65.4%)	49 (71.0%0	1 (ref)		99 (69.7%)	152 (64.4%)	1 (ref)		138 (70.8%)	113 (61.7%)	1 (ref)
	AG	107 (34.6%)	20 (29.0%)	0.77 (0.43–1.37)	0.370	43 (30.3%)	84 (35.6%)	1.19 (0.76–1.88)	0.451	57 (29.2%)	70 (38.3%)	1.53 (0.99–2.36)	0.054
Without thyroid benign disease	AA	64 (64.6%)	11 (61.1%)	1 (ref)		24 (63.2%)	51 (64.6%)	1 (ref)		49 (68.1%)	26 (57.8%)	1 (ref)
	AG	35 (35.4%)	7 (38.9%)	1.17 (0.41–3.33)	0.773	14 (36.8%)	28 (35.4%)	0.92 (0.40–2.12)	0.853	23 (31.9%)	19 (42.2%)	1.68 (0.76–3.74)	0.202
With thyroid benign disease	AA	219 (62.9%)	66 (78.6%)	1 (ref)		124 (69.7%)	161 (63.4%)	1 (ref)		154 (68.1%)	131 (63.6%)	1 (ref)
	AG	129 (37.1%)	18 (21.4%)	0.44 (0.25–0.78)	0.005	54 (30.3%)	93 (36.6%)	1.24 (0.82–1.89)	0.313	72 (31.9%)	75 (36.4%)	1.25 (0.84–1.87)	0.277
For G749A
Total PTC subjects	GG	251 (70.1%)	65 (73.9%)	1 (ref)				1 (ref)				1 (ref)	$P$
	GA	86 (24.0%)	19 (21.6%)	0.86 (0.48–1.52)	0.596	132 (73.7%)	184 (68.9%)	1.17 (0.73–1.86)	0.518	194 (74.6%)	122 (65.6%)	1.64 (1.05–2.56)	0.031
	AA	21 (5.9%)	4 (4.5%)	0.75 (0.25–2.26)	0.604	40 (22.3%)	65 (24.3%)	1.85 (0.73–4.72)	0.197	52 (20.0%)	53 (28.5%)	1.27 (0.56–2.88)	0.575
BRAF-wild type PTC
	GG	71 (73.2%)	19 (70.4%)	1 (ref)				1 (ref)		62 (76.5%)	28 (65.1%)	1 (ref)
	GA	18 (18.6%)	7 (25.9%)	1.32 (0.46–3.81)	0.609	41 (75.9%)	49 (70.0%)	0.94 (0.36–2.47)	0.895	13 (16.0%)	12 (27.9%)	2.00 (0.78–5.10)	0.147
	AA	8 (8.2%)	1 (3.7%)	0.42 (0.04–4.10)	0.458	11 (20.4%)	14 (20.0%)	3.10 (0.57–6.99)	0.192	6 (7.4%)	3 (7.0%)	1.31 (0.30–5.78)	0.726
BRAF-mutant PTC
	GG	169 (68.4%)	42 (73.7%)	1 (ref)		83 (71.6%)	128 (68.1%)	1 (ref)		123 (72.8%)	88 (65.2%)	1 (ref)
	GA	65 (26.3%)	12 (21.1%)	0.76 (0.37–1.54)	0.439	28 (24.1%)	49 (26.1%0	1.15 (0.66–1.99)	0.617	38 (22.5%)	39 (28.9%)	1.44 (0.86–2.44)	0.170
	AA	13 (5.3%)	3 (5.3%)	0.92 (0.25–3.40)	0.904	5 (4.3%)	11 (5.9%)	1.47 (0.48–4.53)	0.506	8 (4.7%)	8 (5.9%)	1.40 (0.50–3.86)	0.522
Without thyroid benign disease
	GG	68 (76.4%)	12 (75.0%)	1 (ref)		30 (85.7%)	50 (71.4%)	1 (ref)		50 (76.9%)	30 (75.0%)	1 (ref)
	GA	17 (19.1%)	3 (18.8%)	1.03 (0.26–4.16)	0.965	5 (14.3%)	15 (21.4%)	2.06 (0.64–6.61)	0.225	11 (16.9%)	9 (22.5%)	1.43 (0.51–3.99)	0.496
	AA	4 (4.5%)	1 (6.2%)	1.27 (0.13–12.72)	0.840	0 (0.0%)	5 (7.1%)	/	/	4 (6.2%)	1 (2.5%)	0.34 (0.03–3.32)	0.350
With thyroid benign disease
	GG	183 (68.0%)	53 (73.6%)	1 (ref)		102 (70.8%)	134 (68.0%)	1 (ref)		144 (73.8%)	92 (63.0%)	1 (ref)
	GA	69 (25.7%)	16 (22.2%)	0.90 (0.42–1.50)	0.481	35 (24.3%)	50 (25.4%)	1.06 (0.63–1.78)	0.819	41 (21.0%)	44 (30.1%)	1.71 (1.03–2.82)	0.037
	AA	17 (6.3%)	3 (4.2%)	0.64 (0.18–2.28)	0.491	7 (4.9%)	13 (6.6%)	1.40 (0.52–3.78)	0.502	10 (5.1%)	10 (6.8%)	1.60 (0.64–4.01)	0.315
For A826G
Total PTC subjects	AA	147 (51.9%)	47 (61.0%)	1 (ref)		81 (54.7%)	116 (53.7%)	1 (ref)		121 (59.0%)	75 (47.8%)	1 (ref)
	AG	91 (32.2%)	22 (28.6%)	0.71 (0.40–1.28)	0.257	44 (29.7%)	70 (32.4%)	1.03 (0.63–1.68)	0.896	55 (26.8%)	58 (36.9%)	1.80 (1.12–2.89)	0.016
	GG	45 (15.9%)	8 (10.4%)	0.56 (0.24–1.28)	0.170	23 (15.5%)	30 (13.9%)	0.85 (0.45–1.61)	0.614	29 (14.1%)	24 (15.3%)	1.41 (0.80–2.62)	0.277
BRAF-wild type PTC	AA	42 (59.2%)	14 (58.3%)	1 (ref)		27 (64.3%)	30 (54.5%)	1 (ref)		36 (61.0%)	20 (55.6%)	1 (ref)
	AG	18 (25.4%)	8 (33.3%)	0.58 (0.09–3.55)	0.553	11 (26.2%)	16 (29.1%)	1.17 (0.43–3.18)	0.757	14 (23.7%)	12 (33.3%)	1.69 (0.60–4.71)	0.319
	GG	11 (15.5%)	2 (8.3%)	1.18 (0.39–3.54)	0.767	4 (9.5%)	9 (16.4%)	2.29 (0.59–8.94)	0.233	9 (15.3%)	4 (11.1%)	1.20 (0.30–4.77)	0.798
BRAF-mutant PTC	AA	101 (50.0%)	30 (61.2%)	1 (ref)		51 (51.5%)	82 (53.2%)	1 (ref)		81 (57.9%)	52 (46.0%)	1 (ref)
	AG	70 (34.7%)	13 (26.5%)	0.67 (0.25–1.77)	0.418	32 (32.3%)	51 (33.1%)	0.92 (0.51–1.64)	0.766	40 (28.6%)	43 (38.1%)	1.73 (0.99–3.02)	0.055
	GG	31 (15.3%)	6 (12.2%)	0.60 (0.29–1.25)	0.603	16 (16.2%)	21 (13.6%)	0.72 (0.34–1.56)	0.410	19 (13.6%)	18 (15.9%)	1.49 (0.72–3.12)	0.286
Without thyroid benign disease	AA	38 (59.4%)	4 (36.4%)	1 (ref)		16 (66.7%)	27 (51.9%)	1 (ref)		31 (62.0%)	12 (46.2%)	1 (ref)
	AG	19 (29.7%)	3 (27.3%)	1.61 (0.31–8.41)	0.547	6 (25.0%)	16 (30.8%)	1.72 (0.54–5.45)	0.361	12 (24.0%)	10 (38.5%)	2.16 (0.71–6.53)	0.174
	GG	7 (10.9%)	4 (36.4%)	5.59 (1.09–8.65)	0.039	2 (8.3%)	9 (17.3%)	2.79 (0.50–5.48)	0.242	7 (14.0%)	4 (15.4%)	1.46 (0.35–6.05)	0.604
With thyroid benign disease	AA	109 (49.8%)	43 (65.2%)	1 (ref)		65 (52.4%)	88 (54.3%)	1 (ref)		90 (58.1%)	63 (48.1%)	1 (ref)
	AG	72 (32.9%)	19 (28.8%)	0.61 (0.33–1.16)	0.132	38 (30.6%)	53 (32.7%)	0.90 (0.52–1.56)	0.713	43 (27.7%)	48 (36.6%)	1.72 (1.01–2.93)	0.047
	GG	38 (17.4%)	4 (6.1%)	0.27 (0.09–0.81)	0.020	21 (16.9%)	21 (13.0%)	0.65 (0.31–1.33)	0.234	22 (14.2%)	20 (15.3%)	1.38 (0.69–2.76)	0.369

Table 4

Association between GAS-AS1 genetic alterations and pre-operative thyroid function

Variants	Genotype	N	fT3		fT4		TSH		ATPO		Anti-TG		TG		PTH
			Mean	$P$	Mean	$P$	Mean	$P$	Mean	$P$	Mean	$P$	Mean	$P$	Mean	$P$
For A713G
Total PTC subjects	AA	331	4.94		17.19		2.54		63.19		245.78		161.27		36.10
	AG	180	5.60	0.244	17.66	0.314	2.92	0.172	62.82	0.974	228.34	0.777	92.5	0.363	59.47	0.325
T1 stage	AA	262	4.90		17.07		2.53		66.01		224.69		147.08		36.33
	AG	158	5.57	0.178	17.49	0.296	3.04	0.105	62.5	0.784	227.32	0.967	62.82	0.334	62.60	0.331
T2 stage	AA	69	5.08		17.69		2.56		52.5		315.88		215.14		35.23
	AG	22	5.80	0.357	18.86	0.533	2.03	0.180	65.11	0.650	254.5	0.763	304.36	0.489	36.92	0.606
Sinlge lesion	AA	188	4.92		17.62		2.64		65.35		234.93		149.81		37.03
	AG	89	5.23	0.055	17.01	0.402	2.92	0.403	54.3	0.475	286.87	0.545	75.22	0.504	85.65	0.312
Multiple lesions	AA	143	4.96		17.43		2.40		60.35		260.05		176.33		34.86
	AG	91	5.96	0.254	17.70	0.621	2.92	0.217	71.15	0.534	171.09	0.319	109.6	0.510	33.85	0.541
For G749A
Total PTC subjects	GG	288	5.01		17.29		2.46		61.85		269.18		135.62		51.62
	GA	98	4.88	0.346	16.93	0.385	2.36	0.680	59.32	0.859	174.9	0.167	205.16	0.513	36.28	0.546
	AA	23	4.91	0.707	16.99	0.721	4.49	0.040	96.81	0.361	288.24	0.905	44.17	0.610	37.03	0.781
T1 stage	GG	232	4.92		17.02		2.44		63.59		254.85		120.35		55.04
	GA	80	4.83	0.33	16.88	0.719	2.39	0.875	62.83	0.962	161.46	0.135	169.59	0.699	36.75	0.559
	AA	19	4.92	0.995	17.00	0.982	5.14	0.021	112.7	0.285	346.81	0.588	19.94	0.638	37.90	0.790
T2 stage	GG	56	5.36		18.40		2.55		54.64		328.59		198.6		37.47
	GA	18	5.12	0.673	17.16	0.371	2.19	0.507	43.75	0.704	234.64	0.695	363.28	0.225	34.19	0.399
	AA	4	4.85	0.663	16.96	0.614	1.43	0.294	21.32	0.517	10.01	0.467	159.26	0.868	32.86	0.523
Sinlge lesion	GG	178	5.01		17.08		2.64		68.07		275.88		296.82		60.89
	GA	48	4.94	0.744	16.89	0.757	2.64	0.994	45.98	0.274	92.27	0.004	108.14	0.392	40.39	0.657
	AA	13	5.06	0.896	16.85	0.839	4.09	0.257	91.88	0.544	349.3	0.723	27.84	0.707	37.48	0.792
Multiple lesions	GG	110	5.01		17.62		2.17		51.78		258.35		180.48		36.63
	GA	50	4.83	0.205	16.96	0.24	2.09	0.740	72.13	0.310	254.22	0.974	117.17	0.657	32.32	0.041
	AA	10	4.72	0.291	17.18	0.708	5.02	0.088	103.2	0.172	208.86	0.845	65.39	0.714	36.44	0.967
For A826G
Total PTC subjects	AA	174	4.97		17.36		2.37		61.26		256.65		180.53		35.91
	AG	106	4.92	0.584	17.12	0.522	2.32	0.846	63.01	0.909	194.96	0.453	191.96	0.933	36.49	0.760
	GG	51	4.85	0.341	16.79	0.277	3.57	0.017	70.16	0.673	314.33	0.620	32.78	0.332	35.91	0.999
T1 stage	AA	134	4.95		17.20		2.27		64.08		232.11		179.56		36.05
	AG	85	4.86	0.404	17.01	0.666	2.35	0.777	62.23	0.916	181.06	0.531	159.97	0.904	37.19	0.606
	GG	43	4.84	0.434	16.75	0.452	3.69	0.015	79.49	0.537	303.86	0.543	20.42	0.392	35.49	0.828
T2 stage	AA	40	5.05		17.90		2.67		51.84		338.87		183.78		35.46
	AG	21	5.17	0.579	17.54	0.645	2.22	0.390	66.14	0.645	251.25	0.720	319.03	0.293	33.67	0.613
	GG	8	4.95	0.676	17.00	0.416	2.93	0.763	20	0.334	370.58	0.930	99.23	0.585	38.16	0.579
Sinlge lesion	AA	109	4.92		17.13		2.44		68.93		274.04		122.11		35.70
	AG	51	4.98	0.58	16.96	0.703	2.62	0.630	62.69	0.780	108.58	0.035	279.01	0.411	40.58	0.085
	GG	28	4.85	0.666	16.66	0.437	3.46	0.137	56.27	0.660	312.81	0.795	22.33	0.583	35.79	0.976
Multiple lesions	AA	65	5.07		17.75		2.24		48.41		227.49		278.49		36.27
	AG	55	4.86	0.216	17.26	0.462	2.05	0.486	63.3	0.464	275.06	0.723	110.31	0.338	32.70	0.143
	GG	23	4.86	0.339	16.95	0.400	3.71	0.054	87.06	0.318	316.18	0.640	45.51	0.384	36.05	0.948

3.2 Associations between GAS8-AS1 alterations and geographical features in thyroid cancer patients

As shown in Table 2, the c.749G $>$ A variation of GAS8-AS1 showed an association with the diagnostic age of PTC occurrence. The variant A allele was more likely to occur in younger people. A borderline association was found between c.713A $>$ G and family history of the tumor ( $P=$ 0.083). Subjects with the c.713G variant may be more likely to develop PTC. No association was observed between GAS8-AS1 c.713A $>$ G/749G $>$ A and gender, smoking, alcohol drinking, BRAF, or TERT mutations.

3.3 Associations between GAS8-AS1 alterations and pathological features in thyroid cancer patients

The c.713A $>$ G variation is closely related to the T stage of PTC patients. Subjects carrying the G allele had a reduced risk of developing advanced T stage PTC (OR $=$ 0.54, 95% CI: 0.33–0.88, $P=$ 0.014). We also focused on the association of the BRAF V600E mutation and concomitant benign thyroid disease with GAS-AS1 alterations in thyroid cancer. This risk of T stage PTC markedly reduced in PTC subpopulations without BRAF V600E mutations (OR $=$ 0.25, 95% CI: 0.09–0.72, $P=$ 0.010). This association was also observed in the subgroup of PTC patients with concomitant benign thyroid diseases.

As for c.749G $>$ A, subjects with the AG genotype carried an increased risk of developing multiple lesions of PTC (for AG vs. GG: OR $=$ 1.64, 95% CI: 1.05–2.56, $P=$ 0.031). Consistent associations were also found in the PTC subgroup with concomitant benign thyroid diseases.

As for c.826G $>$ A, subjects with the GG genotype also showed an increased risk of PTC with multiple lesions (for AG vs. GG, OR $=$ 1.80, 95% CI: 1.12–2.89, $P=$ 0.016). The GG genotype was associated with an increased risk of developing advanced T stage PTC without benign thyroid disease (for GG vs. AA: OR $=$ 5.59, 95% CI: 1.09–8.65, $P=$ 0.039). In contrast, this variant reduced the risk of PTC development at a higher T stage and with benign thyroid disease (GG vs. AA: OR $=$ 0.27, 95% CI: 0.09–0.81, $P=$ 0.020). These results are summarized in Table 3. No association was found between genetic alterations and capsular invasion or calcium (Supplementary Table 2).

3.4 Influence of GAS8-AS1 alterations on thyroid function in patients with thyroid cancer

We explored the effect of GAS8-AS1 genetic alterations on thyroid function (Table 4). Preoperative serum levels of fT3, fT4, TSH, ATPO, PTH, TG, and anti-TG from PTC patients were compared with the rare variant and common genotypes. No correlation was found between the c.713A $>$ G variation and the seven parameters of thyroid function. The variant AA genotype at the c.749G $>$ A site was correlated with higher levels of TSH in patients with PTC and in patients with T1 PTC. Similarly, the variant AA genotype at the c.826G $>$ A site was closely correlated with higher levels of TSH in patients with PTC and in subpopulations with T1 PTC or multiple lesions. Moreover, variants at the c.749G $>$ A and c.826G $>$ A loci showed lower levels of anti-TG in patients with single lesions. Variants at c.749G $>$ A showed a trend toward lower levels of PTH in patients with multiple lesions.

4. Discussion

An increasing number of studies have demonstrated the vital role of GAS8-AS1 in PTC, as well as in other malignancies, including colorectal cancer and hepatocellular carcinoma [4, 6, 7]. This gene encodes a lncRNA that may function as a tumor suppressor. In PTC cells, GAS8-AS1 inhibited cell growth through regulation of the miR-135b-5p/CCND2 axis [8]. A whole exome sequencing study in 2016 reported an association of c.713A $>$ G/714T $>$ C GAS8-AS1 dinucleotide substitution with advanced PTC disease [3]. In contrast to this study, we observed a reduced risk of developing T2 PTC than developing T1 PTC in subjects carrying the c.713G variant allele. To explain this discrepancy, one likely reason is that PTC patient populations of the present and previous study had different T stages. In this study, early stage PTC patients were enrolled, which merely involved the T1 and T2 stages. Although our study enrolment criteria did not exclude patients at the T3 or T4 stage, we did not encounter eligible participants with T3 or T4 PTC who underwent thyroidectomy within our study period, from August 2017 to September 2018. The previous study included a large portion of T3 and T4-stage patients. Therefore, the grouping of T stage is different between the two studies. The T stage was divided into T1 and T2 subgroups in this study, and T1–2 and T3–4 in the previous study. Differences in the grouping of patients might partly account for the discrepancy in the association between the c.713A $>$ G and T stages. However, since no PTC patients at the T3 or T4 stage were included in our study, it precluded us from further exploring the link between GAS8-AS1 alterations and clinicopathological parameters in advanced PTC.

Our results highlight the prevalent involvement of genetic variations in GAS8-AS1 in PTC. The sequence of GAS8-AS1 had a high frequency of genetic alterations in addition to c.713A $>$ G/714T $>$ C, as previously reported [3]. Substitution of c.A713G was completely linked with four other loci at c.714T $>$ C, c.728A $>$ G, c.737G $>$ A, and c.752G $>$ A. Additionally, two novel alterations located at c.749 and c.826 were found in this study, which have not been reported. c.749 G $>$ A was associated with a greater risk of developing multiple lesions and was more likely to occur in younger PTC patients. PTC patients carrying the heterozygous variant at c.749 and c.826 sites had a higher risk of multiple lesions. These associations were also observed in patients with PTC and concomitant benign thyroid disease. Notably, the c.826GG variant conferred a higher risk for developing T2 PTC without benign thyroid disease, whereas the risk for developing T2 PTC with benign thyroid disease was lower when T1 PTC was set as a reference. However, there is no related functional report concerning GAS8-AS1 c.749 and c.826 sites. GAS8-AS1 may function as part of a surveillance mechanism that maintains the activation of the GAS8 promoter and transcription, preventing carcinogenesis [4]. GAS8 is a candidate tumor suppressor that can inhibit malignant activity, including proliferation, invasion, and metastasis, in hepatocellular carcinoma [4]. The genomic region of GAS8-AS1 was highly variable. These variations may have a potential effect on the secondary structure or the expression level of GAS8-AS1, which may further influence the regulation of target gene expression, such as GAS8, through certain epigenetic modulation mechanisms. The plasma level of GAS8-AS1 was found to be lower in patients with PTC than in those with nodular goiters, which was correlated with the lymph node metastasis of PTC [5]. However, the mechanism through which GAS8-AS1 c.749 and c.826 loci regulate tumor cell function remains unclear. Further studies on the molecular mechanisms underlying these variants in the development and progression of PTC are still urgently needed.

We explored whether genetic alterations in GAS8-AS1 affected thyroid function in PTC patients before surgery. The present study is the first to correlate GAS8-AS1 genetic variations with TSH levels in thyroid cancer patients. Alterations of c.749G $>$ A and c.826A $>$ G led to higher levels of TSH in all PTC subjects and in the T1 PTC subpopulation. c.826A $>$ G was also associated with higher TSH levels in PTC subpopulations with multiple lesions. Previous studies also showed an association between genetic variations and TSH expression, including somatic mutation and germline common variation [9, 10, 11, 12]. TSH normally acts as a specific ligand of the TSH receptor (TSHR) in the thyroid gland. TSH is the most extensively studied thyroid hormone in PTC and higher serum TSH concentrations have been associated with PTC progression. Increased TSH may trigger increased TSH/TSHR signaling, which in turn increases the expression of downstream enzymes, such as manganese superoxide dismutase and dual-specific phosphatase 6, eventually resulting in the production of oncogenic proteins, such as c-Myc [13]. A simultaneous increase in serum TSH levels in response to oncogenes, such as c-Myc, would lead to progression into malignancy, thus showing the importance of TSH signaling in the development of PTC [13, 14, 15]. These findings may help explain the increased risk of occurrence of multiple lesions in thyroid cancer patients with alterations at GAS8-AS1 c.749 and c.826 loci.

5. Conclusion

In addition to the frequently studied c.713/c.714 loci, two novel variations at the c.749 and c.826 loci in GAS8-AS1 were found to be associated with multiple lesions in PTC patients. Moreover, they had an effect on the increase in serum TSH levels. Detection of GAS8-AS1 genetic alterations would be useful in the screening and assessment of PTC.