Sage Journals: Discover world-class research

Abstract

Previous research has documented elevated risk of brain cancer in patients with multiple sclerosis (MS). Separately, human leukocyte antigen (HLA) has been implicated in protection or susceptibility for both conditions. The aim of the current study was to assess a possible role of shared immunogenetic influence on risk of MS and brain cancer. We first identified an immunogenetic profile for each condition based on the covariance between the population frequency of 127 high-resolution HLA alleles and the population prevalence of each condition in 14 Continental Western European countries and then evaluated the correspondence between MS and brain cancer immunogenetic profiles. Also, since each individual carries 12 HLA alleles (2 × 6 genes), we estimated HLA protection and susceptibility for MS and brain cancer at the individual level. We found that the immunogenetic profiles of MS and brain cancer were highly correlated overall (P < .001) and across all 6 HLA genes with the strongest association observed for DRB1, followed by DQB1 and HLA-A. These findings of immunogenetic overlap between MS and brain cancer are discussed in light of the role of HLA in the immune system response to viruses and other foreign antigens.

Keywords

Multiple sclerosis brain cancer human leukocyte antigen immunogenetics

Significance Statement

Multiple sclerosis and brain cancer prevalences covary with the population frequencies of human leukocyte antigen (HLA) alleles. The resulting immunogenetic profiles of multiple sclerosis and brain cancer are highly correlated. Shared immunogenetic influences may underlie elevated occurrence of brain cancer in patients with multiple sclerosis.

Introduction

The human immune system is designed to promote health via detection and elimination of pathogens and, under healthy conditions, that role is preserved; however, immune system alterations are implicated in a wide range of conditions,¹ and common mechanisms may underlie co-occurrence of immune-mediated conditions. To that end, both multiple sclerosis (MS), a chronic autoimmune disease characterized by central nervous system demyelination and axonal degradation, and cancer are associated with immune system dysfunction,² prompting several studies investigating cancer risk in patients with MS. Those studies have generally found that overall cancer risk as well as the risk of some specific cancers is reportedly lower in patients with multiple sclerosis (MS) compared to the general population^3-5; however, brain cancer is often found to be more common in MS patients^5-7 (c.f., Ref.⁸) and is associated with decreased survival.⁹ It has been suggested that the immunological profile of MS patients may be related to be increased or decreased risk of some cancers.⁷ Since human leukocyte antigens (HLA) are critically involved in the human immune response, here we evaluated the association of MS and brain cancer though the lens of shared HLA profiles.

The HLA region on chromosome 6, the most polymorphic region of the entire human genome, codes for 2 main classes of cell-surface proteins involved in the human immune system response to non-self antigens including viruses, bacteria, and cancer neoantigens. HLA Class I (HLA-A, B, C) binds with and presents intracellular antigen peptides to CD8+ cytotoxic T cells, signaling destruction of infected cells; HLA Class II (HLA-DR, DQ, and DP genes) binds with and presents endocytosed extracellular antigen peptides to CD4+ T cells, thereby promoting B-cell mediated antibody production and adaptive immunity. HLA has been implicated in a wide variety of conditions¹⁰ and subtle differences, particularly in the binding groove, contribute to variation in disease associations.^11,12

HLA has been widely implicated in autoimmune diseases including MS.^13-18 The strongest and most consistent HLA-MS effect is increased MS risk associated with HLA-DRB1*15:01, although several other HLA alleles have been associated with MS risk (eg, A*03:01) and protection (eg, DRB1*14:01 and A*02:01).¹⁴ We have recently utilized an immunogenetic epidemiological approach^19-27 to characterize an HLA profile for various conditions based on correspondence between the population frequencies of 127 HLA Class I and II alleles and the population prevalence of a given disease in Continental Western European countries. That approach results in identification of HLA alleles that are negatively associated with disease prevalence (ie, protective alleles) as well as HLA alleles that are positively associated with disease prevalence (ie, susceptibility alleles). Application of that approach to MS²³ generally corroborated several previous findings regarding HLA-MS associations and extended the literature regarding HLA-MS associations to include effects of a large number of HLA alleles, many of which were found to be protective against MS prevalence at the population level.

HLA has also been widely implicated in cancer.^25-29 Like MS, several previous studies investigating the influence of HLA on brain cancer have identified HLA associations with increased or decreased brain cancer risk; however, findings across studies have been inconsistent.^30-37 Population-differences in HLA composition^38,39 coupled with differences in sample characteristics may partially account for the inconsistent findings. For example, in Caucasian Germans, HLA-A*25, HLA-B*27, and HLA-DRB1*15 were associated with increased risk of cerebral glioma whereas HLA-DRB1*07 was associated with decreased risk.³⁷ In Italians, HLA-A*11, DQB1*06, DRB1*14, and DRB3*01 were associated with increased risk of glioma and B*07, C*04, and C*05 were associated with decreased risk.³² Still, studies in the United States including primarily individuals of European Ancestry identified protective effects of HLA-A*32^30,31 and DQB1*05³⁷ on glioma risk and increased risk associated with B*13 and B*55³² as well as DQB1*06 and DRB1*13.³⁶ In addition to sample differences, most prior HLA-brain cancer studies have relied on 1-field HLA resolution which limits HLA-disease associations to the allele level. Since even single amino acid differences in HLA affect disease associations,^11,12 2-field resolution, which specifies protein-level differences within an allele group, permits more nuanced investigation of HLA-disease effects. One large consortium study of glioma patients and controls of European ancestry reported increased risk of glioma in carriers of DRB1*15:01, DQA1*01:02, and DQB1*06:02, and the related extended haplotype.³³

In light of increased brain cancer in MS patients and some evidence of common HLA associations in both conditions (eg, DRB1*15:01), here we first identified an immunogenetic profile for MS and brain cancer based on the covariance between the population frequency of 127 high-resolution HLA alleles and the population prevalence of each condition and then evaluated the correspondence between MS and brain cancer immunogenetic profiles.

Materials and Methods

Prevalence of multiple sclerosis and brain cancer

The population prevalences of MS and brain cancer (BC) were computed for 14 countries in Continental Western Europe (Table 1). For each country we identified the total number of people with each condition in 2016 from the Global Health Data Exchange,⁴⁰ a publicly available catalog of data from the Global Burden of Disease study, and divided those values by the total population of each country in 2016.⁴¹ Life expectancy was not included in the current analyses since it is virtually identical for these countries.²⁰

Table 1.

Multiple sclerosis (MS) and brain prevalence (BC) in the 14 CWE countries studied. Counts (N) are for 2016.

	Country	N (MS)	N (BC)	N total population	MSPrevalence (%)	BCPrevalence (%)
1	Austria	10 999	2467	8 800 000	0.125	0.028
2	Belgium	14 752	7058	11 300 000	0.131	0.063
3	Denmark	11 673	8660	5 700 000	0.205	0.152
4	Finland	8209	5925	5 500 000	0.149	0.108
5	France	65 467	39 188	64 600 000	0.101	0.061
6	Germany	111 970	13 147	82 600 000	0.136	0.016
7	Greece	7727	9365	10 800 000	0.072	0.087
8	Italy	72 352	18 945	60 600 000	0.119	0.031
9	Netherlands	25 197	12 204	17 000 000	0.148	0.072
10	Norway	7518	6954	5 200 000	0.145	0.134
11	Portugal	8367	2703	10 300 000	0.081	0.026
12	Spain	43 867	20 304	43 300 000	0.101	0.047
13	Sweden	20 304	8915	9 900 000	0.205	0.090
14	Switzerland	13 968	5294	8 400 000	0.166	0.063

HLA alleles

We obtained the population frequency of 127 HLA alleles from 14 Continental Western European Countries (Austria, Belgium, Denmark, Finland, France, Germany, Greece, Italy, Netherlands, Portugal, Norway, Spain, Sweden, and Switzerland), The frequencies of all reported HLA alleles of classical genes of Class I (A, B, C) and Class II (DPB1, DQB1, DRB1) for each of the 14 Continental Western European countries were retrieved from Allele Frequencies in Worldwide Populations (http://allelefrequencies.net/hla6006a.asp) on October 20, 2020. There was a total of 2746 entries of alleles from the 14 Continental Western European countries, comprising 844 distinct alleles. Of those, 127 alleles occurred with frequencies ⩾0.01 in 9 or more countries and were used in further analyses (69 Class I and 58 Class II alleles). The alleles and their mean frequencies of the individual alleles (across countries) are given in Tables 2 and 3 for HLA Class I and Class II, respectively. Alleles of genes DRB3, DRB4 and DRB5 were not used since they do not occur in all individuals. All but one (A*36:01) of the 127 alleles have previously been reported to be Common in all population groups (African/African American, Asian/Pacific Islands, European/European descent, Middle East/North Coast of Africa, South or Central America/Hispanic/Latino, Native American populations, and unknown/not asked/multiple ancestries/other, and in all groups combined (Total).⁴² Allele A*36:01 was Intermediate for American, Asian/Pacific Islands, European/European descent, but Common for all other groups and Total.⁴²

Table 2.

The 69 HLA Class I alleles used and their mean frequencies.

Class I
Gene A			Gene B			Gene C
	Allele	Frequency		Allele	Frequency		Allele	Frequency
1	A*01:01	0.1170	1	B*07:02	0.1009	1	C*01:02	0.0370
2	A*02:01	0.2715	2	B*08:01	0.0791	2	C*03:03	0.0506
3	A*02:05	0.0122	3	B*13:02	0.0238	3	C*04:01	0.1349
4	A*03:01	0.1501	4	B*14:01	0.0091	4	C*05:01	0.0716
5	A*11:01	0.0527	5	B*14:02	0.0275	5	C*06:02	0.0829
6	A*23:01	0.0237	6	B*15:01	0.0544	6	C*07:01	0.1472
7	A*24:02	0.1051	7	B*15:17	0.0104	7	C*07:02	0.1020
8	A*25:01	0.0139	8	B*15:18	0.0043	8	C*07:04	0.0146
9	A*26:01	0.0356	9	B*18:01	0.0609	9	C*12:02	0.0160
10	A*29:01	0.0058	10	B*27:02	0.0070	10	C*12:03	0.0678
11	A*29:02	0.0315	11	B*27:05	0.0435	11	C*14:02	0.0231
12	A*30:01	0.0165	12	B*35:01	0.0690	12	C*15:02	0.0370
13	A*30:02	0.0132	13	B*35:02	0.0187	13	C*16:01	0.0303
14	A*31:01	0.0295	14	B*35:03	0.0261
15	A*32:01	0.0368	15	B*35:08	0.0111
16	A*33:01	0.0116	16	B*37:01	0.0136
17	A*33:03	0.0066	17	B*38:01	0.0276
18	A*36:01	0.0063	18	B*39:01	0.0146
19	A*68:01	0.0353	19	B*39:06	0.0069
20	A*68:02	0.0220	20	B*40:01	0.0474
			21	B*40:02	0.0212
			22	B*41:01	0.0087
			23	B*41:02	0.0056
			24	B*44:02	0.0623
			25	B*44:03	0.0431
			26	B*44:05	0.0054
			27	B*45:01	0.0090
			28	B*47:01	0.0043
			29	B*49:01	0.0220
			30	B*50:01	0.0164
			31	B*51:01	0.0640
			32	B*52:01	0.0158
			33	B*55:01	0.0129
			34	B*56:01	0.0075
			35	B*57:01	0.0278
			36	B*58:01	0.0141

Table 3.

The 58 HLA Class II alleles used and their mean frequencies.

Class II
Gene DPB1			Gene DQB1			Gene DRB1
	Allele	Frequency		Allele	Frequency		Allele	Frequency
1	DPB1*01:01	0.0462	1	DQB1*02:01	0.1514	1	DRB1*01:01	0.0862
2	DPB1*02:01	0.1486	2	DQB1*02:02	0.0664	2	DRB1*01:02	0.0202
3	DPB1*02:02	0.0127	3	DQB1*03:01	0.1925	3	DRB1*01:03	0.0067
4	DPB1*03:01	0.1273	4	DQB1*03:02	0.1030	4	DRB1*03:01	0.1233
5	DPB1*04:01	0.3905	5	DQB1*03:03	0.0444	5	DRB1*04:01	0.0688
6	DPB1*04:02	0.1206	6	DQB1*04:02	0.0370	6	DRB1*04:02	0.0118
7	DPB1*05:01	0.0232	7	DQB1*05:01	0.1181	7	DRB1*04:03	0.0124
8	DPB1*06:01	0.0162	8	DQB1*05:02	0.0432	8	DRB1*04:04	0.0248
9	DPB1*09:01	0.0117	9	DQB1*05:03	0.0314	9	DRB1*04:05	0.0217
10	DPB1*10:01	0.0176	10	DQB1*06:01	0.0155	10	DRB1*04:07	0.0099
11	DPB1*11:01	0.0178	11	DQB1*06:02	0.1138	11	DRB1*04:08	0.0074
12	DPB1*13:01	0.0174	12	DQB1*06:03	0.0679	12	DRB1*07:01	0.1090
13	DPB1*14:01	0.0164	13	DQB1*06:04	0.0339	13	DRB1*08:01	0.0384
14	DPB1*17:01	0.0212	14	DQB1*06:09	0.0078	14	DRB1*08:03	0.0048
15	DPB1*19:01	0.0099				15	DRB1*09:01	0.0186
						16	DRB1*10:01	0.0124
						17	DRB1*11:01	0.0709
						18	DRB1*11:02	0.0076
						19	DRB1*11:03	0.0096
						20	DRB1*11:04	0.0498
						21	DRB1*12:01	0.0166
						22	DRB1*13:01	0.0713
						23	DRB1*13:02	0.0433
						24	DRB1*13:03	0.0130
						25	DRB1*13:05	0.0034
						26	DRB1*14:01	0.0226
						27	DRB1*15:01	0.1104
						28	DRB1*15:02	0.0115
						29	DRB1*16:01	0.0342

Immunogenetic (HLA) profiles of MS and BC

We computed the covariance between the prevalence of each disease and the population frequency of 127 HLA Class I (N = 69) and Class II (N = 58) alleles in the 14 CWE countries above. The covariance can be negative or positive, indicating a protective or susceptibility (P/S) association, respectively. We call these covariance values MS-PScov and BC-PScov scores for MS and BC, respectively. Thus the HLA profiles for each disease is a vector of 127 PScov scores. The equation for the PScov scores is:

P S cov = \frac{1}{N - 1} \sum_{i}^{i = 1, N} (f_{i} - \bar{f}) (p_{i} - \bar{p})

(1)

where $f_{i}, p_{i}$ denote allele frequency and disease prevalence for the ith country, respectively, and $\bar{f}, \bar{p}$ are their means.

Standard parametric (mean, standard deviation, etc.) and nonparametric statistics (median, interquartile range, etc.) were used to evaluate the distribution of scores. We used Tukey’s⁴³ fences to identify outlier and extreme values of the PScov distribution, as follows.

Interquartile range : IQR = Q 3 - Q 1

(2)

Negative (protective) fence = Q 1 - 1.5 IQR

(3)

Positive (susceptibility) fence = Q 3 + 1.5 IQR

(4)

where Q1, Q3 are the 25th and 75th percentiles, respectively. These fences demarcated outlier values, that is, values outside the fences.

Random permutations test for assessing the statistical significance of the MS and BC HLA PScov profiles

In this analysis, we tested the null hypothesis that the MS- and BC-HLA PScov profiles (vectors) may be due to chance by performing the permutation test below. Let H be the observed Disease-HLA profile and H′ be the profile obtained by random permutation, as follows. (1) For each allele, its observed frequencies were paired to randomly permuted prevalences of MS (or BC) in the 14 CWE countries above. (2) The covariance between allele frequency and corresponding (to the permuted country) disease prevalence was computed to yield a “permuted” Disease-HLA profile, H′, consisting of 127 PScov scores. (3) The observed (H) and permuted (H’) PScov profiles were ranked and the absolute differences between the H and H′ ranks computed and summed: a sum of zero would indicate that the 2 profiles are the same. (4) Finally, we carried out this procedure 1 000 000 times and counted the number of times M that that sum was equal to zero, indicating that the ranks of the observed and permuted PScov scores are the same. (5) Then, the ratio $w = \frac{M}{1, 000, 000}$ is the probability that the observed profile H could be due to chance.

Application to individuals

Since every individual carries 12 classical HLA alleles (2 of each 3 HLA Class I and 3 Class II genes), average MS PScov and BC PScov scores were calculated:

ξ = \frac{1}{12} \sum_{k}^{k = 1, 12} MS PScov (k)

(5)

ζ = \frac{1}{12} \sum_{k}^{k = 1, 12} BC PScov(K)

(6)

We obtained expected estimates of ξ and θ using a bootstrap procedure,⁴⁴ as follows. For each gene and disease, 2 scores were drawn randomly (with replacement) from the pool of available alleles of that gene and disease, and were averaged to yield bootstrap values of ξ* and ζ* for a simulated “individual.” The procedure was repeated 1000 times for a total of 1000 ξ* and ζ* values. The association between MS PScov and BC PScov scores was evaluated by computing the Pearson correlation between the individual (original) MS and BC PScov scores, and between ξ* and ζ*.

Implementation of analysis procedures

The IBM-SPSS statistical package (version 27) was used for implementing standard statistical analyses, including nonparametric exploratory data analysis,⁴³ regression analysis, and testing of proportions. All P values reported are 2-sided. The permutation test and bootstrap procedure was implemented using FORTRAN (Geany, version 1.38, built on or after 2021-10-09) and 64-bit Mersenne Twister random number generator with a large random double-precision odd seed.

Results

Prevalences

The prevalences of MS and BC (Table 1) were weakly but statistically significantly positively associated (Figure 1; Pearson r = .541, P < .046; Spearman r = .578, P = .030, N = 14).

Figure 1.

The prevalence of brain cancer is plotted against the prevalence of MS. N = 14 CWE countries. See text for details.

HLA MS- and BC PScov scores

The MS PScov scores are given in Tables 4 and 5 for HLA Class I and II alleles, respectively; the BC PScov scores are given in Tables 6 and 7 for HLA Class I and II alleles, respectively. These sets of MS-HLA and BC-HLA PScov scores could not be due to chance, since no cases were found where a random pairing of MS- or BC- prevalence with allele frequency (out of 1 000 000 runs) matched the observed MS-HLA or BC-HLA or PScov profiles, thus rejecting the null hypothesis that the observed profiles could be accounted for by chance (P < 1 × 10⁻⁶) for both MS and BC PScov scores.

Table 4.

MS PScov scores for HLA Class I alleles.

Class I
Gene A			Gene B			Gene C
	Allele	MS PScov		Allele	MS PScov		Allele	MS PScov
1	A*01:01	−0.215	1	B*07:02	1.503	1	C*01:02	0.243
2	A*02:01	0.167	2	B*08:01	0.274	2	C*03:03	0.554
3	A*02:05	−0.053	3	B*13:02	0.044	3	C*04:01	−0.142
4	A*03:01	2.085	4	B*14:01	−0.072	4	C*05:01	−0.131
5	A*11:01	−0.359	5	B*14:02	−0.195	5	C*06:02	−0.378
6	A*23:01	−0.207	6	B*15:01	1.326	6	C*07:01	0.212
7	A*24:02	0.455	7	B*15:17	−0.024	7	C*07:02	0.387
8	A*25:01	0.037	8	B*15:18	−0.039	8	C*07:04	−0.044
9	A*26:01	−0.239	9	B*18:01	−0.932	9	C*12:02	0.045
10	A*29:01	0.022	10	B*27:02	0.007	10	C*12:03	−0.304
11	A*29:02	−0.360	11	B*27:05	1.063	11	C*14:02	−0.094
12	A*30:01	−0.023	12	B*35:01	−0.064	12	C*15:02	−0.128
13	A*30:02	−0.145	13	B*35:02	−0.090	13	C*16:01	−0.305
14	A*31:01	0.499	14	B*35:03	−0.273
15	A*32:01	−0.375	15	B*35:08	−0.145
16	A*33:01	−0.082	16	B*37:01	0.238
17	A*33:03	−0.040	17	B*38:01	−0.185
18	A*36:01	−0.027	18	B*39:01	−0.028
19	A*68:01	0.027	19	B*39:06	−0.034
20	A*68:02	0.073	20	B*40:01	1.066
21	B*40:02	0.274
22	B*41:01	−0.050
23	B*41:02	−0.042
24	B*44:02	0.028
25	B*44:03	−0.430
26	B*44:05	−0.048
27	B*45:01	−0.060
28	B*47:01	−0.014
29	B*49:01	−0.405
30	B*50:01	−0.105
31	B*51:01	−1.638
32	B*52:01	−0.113
33	B*55:01	0.038
34	B*56:01	0.055
35	B*57:01	−0.138
36	B*58:01	−0.059

Table 5.

MS PScov scores for HLA Class II alleles.

Class II
Gene DPB1			Gene DQB1			Gene DRB1
	Allele	MS PScov		Allele	MS PScov		Allele	MS PScov
1	DPB1*01:01	0.343	1	DQB1*02:01	0.624	1	DRB1*01:01	0.605
2	DPB1*02:01	−0.582	2	DQB1*02:02	−0.544	2	DRB1*01:02	−0.294
3	DPB1*02:02	−0.185	3	DQB1*03:01	−1.634	3	DRB1*01:03	−0.049
4	DPB1*03:01	0.442	4	DQB1*03:02	0.625	4	DRB1*03:01	−0.267
5	DPB1*04:01	1.026	5	DQB1*03:03	0.885	5	DRB1*04:01	1.461
6	DPB1*04:02	−0.161	6	DQB1*04:02	0.687	6	DRB1*04:02	−0.146
7	DPB1*05:01	0.124	7	DQB1*05:01	0.240	7	DRB1*04:03	−0.193
8	DPB1*06:01	0.069	8	DQB1*05:02	−0.977	8	DRB1*04:04	0.172
9	DPB1*09:01	−0.025	9	DQB1*05:03	−0.342	9	DRB1*04:05	−0.245
10	DPB1*10:01	−0.203	10	DQB1*06:01	−0.082	10	DRB1*04:07	−0.087
11	DPB1*11:01	−0.107	11	DQB1*06:02	1.038	11	DRB1*04:08	0.209
12	DPB1*13:01	−0.168	12	DQB1*06:03	0.119	12	DRB1*07:01	−1.010
13	DPB1*14:01	−0.208	13	DQB1*06:04	0.252	13	DRB1*08:01	0.853
14	DPB1*17:01	−0.110	14	DQB1*06:09	−0.038	14	DRB1*08:03	−0.001
15	DPB1*19:01	0.034				15	DRB1*08:01	0.907
						16	DRB1*10:01	0.033
						17	DRB1*11:01	−0.490
						18	DRB1*11:02	−0.087
						19	DRB1*11:03	−0.105
						20	DRB1*11:04	−0.946
						21	DRB1*12:01	0.130
						22	DRB1*13:01	0.190
						23	DRB1*13:02	0.164
						24	DRB1*13:03	−0.055
						25	DRB1*13:05	−0.037
						26	DRB1*14:01	−0.172
						27	DRB1*15:01	1.054
						28	DRB1*15:02	−0.160
						29	DRB1*16:01	−0.491

Table 6.

BC PScov scores for HLA Class I alleles.

Class I
Gene A			Gene B			Gene C
	Allele	BC PScov		Allele	BC PScov		Allele	PScov
1	A*01:01	−0.296	1	B*07:02	0.383	1	C*01:02	0.170
2	A*02:01	0.511	2	B*08:01	0.062	2	C*03:03	0.591
3	A*02:05	−0.100	3	B*13:02	−0.068	3	C*04:01	−0.309
4	A*03:01	1.139	4	B*14:01	−0.053	4	C*05:01	0.161
5	A*11:01	−0.066	5	B*14:02	−0.210	5	C*06:02	−0.126
6	A*23:01	−0.184	6	B*15:01	0.396	6	C*07:01	−0.098
7	A*24:02	0.161	7	B*15:17	−0.032	7	C*07:02	1.138
8	A*25:01	−0.101	8	B*15:18	0.018	8	C*07:04	−0.057
9	A*26:01	−0.125	9	B*18:01	0.121	9	C*12:02	−0.218
10	A*29:01	0.064	10	B*27:02	0.064	10	C*12:03	−0.611
11	A*29:02	−0.186	11	B*27:05	0.533	11	C*14:02	−0.237
12	A*30:01	−0.102	12	B*35:01	0.229	12	C*15:02	−0.195
13	A*30:02	−0.071	13	B*35:02	−0.096	13	C*16:01	−0.061
14	A*31:01	0.220	14	B*35:03	−0.083
15	A*32:01	−0.032	15	B*35:08	0.026
16	A*33:01	−0.095	16	B*37:01	0.093
17	A*33:03	0.012	17	B*38:01	−0.053
18	A*36:01	−0.015	18	B*39:01	0.022
19	A*68:01	0.047	19	B*39:06	0.043
20	A*68:02	−0.279	20	B*40:01	0.596
			21	B*40:02	0.135
			22	B*41:01	0.050
			23	B*41:02	0.001
			24	B*44:02	−0.156
			25	B*44:03	−0.344
			26	B*44:05	0.012
			27	B*45:01	−0.001
			28	B*47:01	0.006
			29	B*49:01	−0.125
			30	B*50:01	−0.129
			31	B*51:01	0.280
			32	B*52:01	0.130
			33	B*55:01	−0.038
			34	B*56:01	0.069
			35	B*57:01	−0.046
			36	B*58:01	0.037

Table 7.

BC PScov scores for HLA Class II alleles.

Class II
Gene DPB1			Gene DQB1			Gene DRB1
	Allele	BC PScov		Allele	BC PScov		Allele	BC PScov
1	DPB1*01:01	0.528	1	DQB1*02:01	0.808	1	DRB1*01:01	0.397
2	DPB1*02:01	−0.475	2	DQB1*02:02	−0.295	2	DRB1*01:02	−0.229
3	DPB1*02:02	−0.112	3	DQB1*03:01	−1.845	3	DRB1*01:03	0.001
4	DPB1*03:01	−0.627	4	DQB1*03:02	0.604	4	DRB1*03:01	−0.601
5	DPB1*04:01	0.943	5	DQB1*03:03	0.457	5	DRB1*04:01	1.519
6	DPB1*04:02	0.557	6	DQB1*04:02	0.393	6	DRB1*04:02	−0.081
7	DPB1*05:01	0.127	7	DQB1*05:01	0.135	7	DRB1*04:03	−0.005
8	DPB1*06:01	−0.035	8	DQB1*05:02	0.026	8	DRB1*04:04	0.090
9	DPB1*09:01	−0.049	9	DQB1*05:03	−0.293	9	DRB1*04:05	−0.288
10	DPB1*10:01	−0.084	10	DQB1*06:01	−0.025	10	DRB1*04:07	−0.062
11	DPB1*11:01	0.053	11	DQB1*06:02	1.273	11	DRB1*04:08	0.151
12	DPB1*13:01	−0.083	12	DQB1*06:03	0.210	12	DRB1*07:01	−0.857
13	DPB1*14:01	−0.008	13	DQB1*06:04	0.051	13	DRB1*08:01	0.481
14	DPB1*17:01	−0.036	14	DQB1*06:09	0.020	14	DRB1*08:03	−0.002
15	DPB1*19:01	0.048			0.448	15	DRB1*09:01	0.448
						16	DRB1*10:01	0.025
						17	DRB1*11:01	−0.900
						18	DRB1*11:02	−0.006
						19	DRB1*11:03	−0.082
						20	DRB1*11:04	−0.388
						21	DRB1*12:01	0.126
						22	DRB1*13:01	0.131
						23	DRB1*13:02	0.049
						24	DRB1*13:03	−0.063
						25	DRB1*13:05	0.008
						26	DRB1*14:01	−0.144
						27	DRB1*15:01	0.962
						28	DRB1*15:02	0.049
						29	DRB1*16:01	0.105

MS and BC PScov scores are plotted ranked in Figure 2A and B respectively. It can be seen that they comprised similar numbers of protection (negative) and susceptibility (positive) values. More specifically, MS PScov scores comprised 78/127 (61.4%; Z = 2.573, P = .01, Wilson score test for single proportion; null hypothesis is 50%) protection values versus 62/127 (48.8%; Z = −0.266, P = .790, same test; null hypothesis of 50%) of BC PScov. This small preponderance of 12.6% in protective scores in MS versus BC was statistically significant (Z = 2.018, P = .044, Wald H0 test of independent sample proportions).

Figure 2.

(A) Multiple sclerosis (MS) PScov scores are plotted ranked, (B) same for brain cancer (BC) scores. N = 127 HLA alleles.

Distributions of MS and BC PScov scores

Their frequency distributions and boxplots are shown in Figures 3A, B, 4A and B, respectively. The distributions were unimodal and did not differ significantly from each other (P = .071, Kolmogorov-Smirnov test), nor did the scores paired for each allele (paired t-test, t_[126], P = .764).

Figure 3.

Frequency distributions of MS (A) and BC (B) PScov scores.

Figure 4.

Boxplots of the distributions of MS (A) and BC (B) PScov scores shown in Figure 3. See text for details.

Associations of MS and BC PScov scores

MS and BC PScov scores were highly significantly and positively correlated (Figure 5; r = .752, P < .001, N = 127). This significant positive correlation held for both HLA Class I and II alleles (Table 8), and all 6 HLA genes (Table 9). A similarly positive association was observed with respect to the protection/susceptibility proportions, as shown in Tables 10 to 14. It can be seen that (a) all associations were positive, as indicated by the positive Phi/Cramer’s V measure, and (b) associations were statistically significant for the whole set of 127 alleles (Table 10), and the sets for Class I (Table 11), Class II (Table 12), Class I gene A (Table 13), and Class II genes DQB1 and DRB1 (Table 14). The strongest association was observed for Class II and gene DRB1.

Figure 5.

BC PScov scores are plotted against MS PScov scores. N = 127 HLA alleles. See text for details.

Table 8.

Association between MS PScov and BC PScov scores for HLA Class I and II. Correlation is Pearson correlation coefficient.

HLA	N	Correlation	Lower 95% CI	Upper 95% CI	P value
Class I	69	.680	0.528	0.790	<.001
Class II	58	.851	0.760	0.909	<.001

Table 9.

Association between MS PScov and BC PScov scores for HLA Class I and II genes. Correlation is Pearson correlation coefficient.

HLA Class	HLA gene	N	Correlation	Lower 95% CI	Upper 95% CI	P value
I	A	20	.889	0.735	0.955	<.001
	B	36	.545	0.263	0.741	.001
	C	13	.738	0.316	0.917	.004
II	DPB1	15	.552	0.056	0.830	.033
	DQB1	14	.891	0.685	0.965	<.001
	DRB1	29	.902	0.801	0.953	<.001

Table 10.

Analysis of 2-way table of all PS scores reexpressed as binary variables.

		Brain Cancer		Total
		Protection	Susceptibility
Multiple sclerosis	Protection	54	24	78
Multiple sclerosis	Susceptibility	8	41	49
	Total	62	65	127
Test	Test value	P value
Chi-square	33.71	<.001
Fisher’s Exact		<.001
Phi/Cramer’s V	0.515	<.001

Table 11.

Analysis of 2-way table of HLA Class I PS scores reexpressed as binary variables.

HLA Class I		Brain cancer		Total
HLA Class I		Protection	Susceptibility
Multiple Sclerosis	Protection	29	16	45
Multiple Sclerosis	Susceptibility	6	18	24
	Total	35	34	69
Test	Test value	P value
Chi-square	9.743	.002
Fisher’s exact		.002
Phi/Cramer’s V	0.376	.002

Table 12.

Analysis of 2-way table of HLA Class II PS scores reexpressed as binary variables.

HLA Class II		Brain cancer		Total
HLA Class II		Protection	Susceptibility
Multiple Sclerosis	Protection	25	8	33
Multiple Sclerosis	Susceptibility	2	23	25
	Total	27	31	58
Test	Test value	P value
Chi-square	26.247	<.001
Fisher’s exact		<.001
Phi/Cramer’s V	0.673	<.001

Table 13.

Analysis of 2-way table of HLA Class I genes PS score reexpressed as binary variables.

HLA Gene A		Brain cancer		Total
HLA Gene A		Protection	Susceptibility
Multiple sclerosis	Protection	11	1	12
Multiple sclerosis	Susceptibility	2	6	8
	Total	13	7	20
Test	Test value	P value
Chi-square	9.377	.002
Fisher’s exact		.004
Phi/Cramer’s V	0.685	.002
HLA Gene B		Brain cancer		Total
HLA Gene B		Protection	Susceptibility
Multiple Sclerosis	Protection	11	13	24
Multiple Sclerosis	Susceptibility	3	9	12
	Total	14	22	36
Test	Test value	P value
Chi-square	1.461	.227
Fisher’s exact		.292
Phi/Cramer’s V	0.201	.227
HLA Gene C		Brain cancer		Total
HLA Gene C		Protection	Susceptibility
Multiple sclerosis	Protection	7	2	9
Multiple sclerosis	Susceptibility	1	3	4
	Total	8	5	13
Test	Test value	P value
Chi-square	3.259	.071
Fisher’s exact		.217
Phi/Cramer’s V	0.501	.071

Table 14.

Analysis of 2-way table of HLA Class II genes PS scores reexpressed as binary variables.

HLA Gene DPB1		Brain cancer		Total
HLA Gene DPB1		Protection	Susceptibility
Multiple sclerosis	Protection	7	2	9
Multiple sclerosis	Susceptibility	2	4	6
	Total	9	6	15
Test	Test value	P value
Chi-square	2.963	.085
Fisher’s exact		.136
Phi/Cramer’s V	0.444	.085
HLA Gene DQB1		Brain cancer		Total
HLA Gene DQB1		Protection	Susceptibility
Multiple sclerosis	Protection	4	2	6
Multiple sclerosis	Susceptibility	0	8	8
	Total	4	10	14
Test	Test value	P value
Chi-square	7.467	.006
Fisher’s exact		.015
Phi/Cramer’s V	0.730	.006
HLA Gene DRB1		Brain cancer		Total
HLA Gene DRB1		Protection	Susceptibility
Multiple sclerosis	Protection	14	4	18
Multiple sclerosis	Susceptibility	0	11	11
	Total	14	15	29
Test	Test value	P value
Chi-square	16.541	<.001
Fisher’s exact		<.001
Phi/Cramer’s V	0.755	<.001

Application to individuals

In this analysis, we generated 1000 bootstrap values of averages of MS and BC PScov scores, ξ* and ζ*, respectively, by averaging 12 PS scores, 2 per gene, randomly selected. This is a more realistic assessment of HLA protection/susceptibility for MS and BC, since an individual carries 12 HLA alleles (2 × 6 genes), each with their own PScov score. The frequency distributions of ξ* and ζ* are shown in Figure 6A and B, respectively. We found the following. (a) There was a statistically significant bias in both distributions toward susceptibility scores (only 42% protective ξ* and 43% ζ*; P < .001, Wilson score test for single proportion, against the null hypothesis of 50%). Similarly, (b) there was a significant bias toward susceptibility of the PScov score itself (mean ± SEM: ξ* 0.0346 ± 0.0047; ζ* 0.0287 ± 0.0042; P < .001, one sample t-test against the null hypothesis that mean is zero). (c) The variance of the ξ* distribution was 1.3x higher than that of ζ* (P < .001, Levene’s test for equality of variances). (d) The distributions of ξ* and ζ* did not differ significantly (P = .164, Kolmogorov-Smirnov test). (e) Paired per allele, ξ* was significantly higher than ζ*, although the difference was small (0.00594 ± 0.0021, P = .004, paired t-test). (f) The boxplots of the ξ* and ζ* distributions (Figure 7A and B) are remarkably similar, with the exact same number of protective and susceptibility outliers, the latter outnumbering the former by 8:1. (g) ξ* and ζ* were strongly, highly significantly and positively correlated (Figure 8; r = .900, P < .001, N = 1000). Similarly, their PScov counts, reexpressed as binary variables, were highly significantly and positively associated (Table 15).

Figure 6.

Frequency distributions of MS ξ* (A) and BC ζ* (B) bootstrap values. N = 1000 values per distribution. See text for details.

Figure 7.

Boxplots of the distributions of MS ξ* (A) and BC ζ* (B) PScov scores shown in Figure 6. See text for details.

Figure 8.

BC ζ* is plotted against MS ξ*. N = 1000 bootstrap values.

Table 15.

Analysis of 2-way table of HLA bootstrap values ξ* and ζ* reexpressed as binary variables.

		Brain cancer (ζ*)		Total
		Protection	Susceptibility
Multiple sclerosis (ξ*)	Protection	361	69	430
Multiple sclerosis (ξ*)	Susceptibility	59	511	570
	Total	420	580	1000
Test	Test value	P value
Chi-square	545.1	<.001
Fisher’s exact		<.001
Phi/Cramer’s V	0.738	<.001

Discussion

Separate lines of research have documented increased risk of brain cancer in MS patients and have implicated HLA in protection or susceptibility for each condition. Here we evaluated the association between HLA profiles of MS and brain cancer to assess a possible role of shared immunogenetic influence on disease risks. We found that the immunogenetic profiles of MS and brain cancer were highly correlated overall and across all 6 HLA genes with the strongest association observed for DRB1 followed by DQB1, and HLA-A. These findings suggest not only that certain alleles may contribute to risk of both conditions but also that certain alleles confer protection against both conditions, both in terms of population prevalence and individual risk. The findings are discussed in light of the role of HLA in the immune system response to foreign antigens.

Multiple sclerosis and brain cancer immunogenetic overlap

The current study documented robust similarities in both HLA risk and protection for MS and brain cancer, especially related to Class II HLA-DRB1. Prior research has found that the highest susceptibility for MS is associated with the HLA-DRB1 gene.¹⁵ In particular, DRB1*15:01is widely linked to MS risk ¹⁴; notably, DRB1*15:01 has also been linked to brain cancer.^33,37 In our study, we similarly found strong associations between DRB1*15:01 and risk of both MS (MS PScov = 1.054) and brain cancer (BC PScov = 0.962). As shown in Tables 4 to 7, it can be seen that several other alleles also exhibited strong positive associations with both MS and brain cancer prevalences including A*03:01, DPB1*04:01, DQB1*06:02, and DRB1*04:01. Some of these associations appear to be novel; however, A*03:01 has previously been implicated in MS,⁴⁵ independent of linkage disequilibrium between DRB1*15:01 and-DQB1*06:02,⁴⁶ and DQB1*06 has been implicated in brain cancer risk.³² Of note, the present finding that DQB1*06:02 was highly associated with both MS and brain cancer risk but DQB1*06:01 was not speaks to the importance of high-resolution (2-field) genotyping to more clearly identify HLA-disease associations that may be masked by allele group-level analyses. In addition to mutual risk alleles, several alleles were negatively associated with both MS and brain cancer prevalence and are considered mutually protective including DQB1*03:01 and DRB1*07:01. Previous research has documented protective effects of DRB1*07 on glioma³⁷ and MS.⁴⁷ DRB1*07:01 has also been found to protect against autoimmune conditions including Graves disease,^48,49 rheumatoid arthritis,⁵⁰ and psoriatic arthritis.⁵¹ We are not aware of prior studies documenting protective effects of DQB1*03:01 on MS or brain cancer; however, prior studies have reported protective effects of DQB1*03:01 on neurological syndromes⁵² and Type I diabetes.⁵³ Taken together, the present findings highlight shared immunogenetic profiles for MS and brain cancer that extend beyond the well-established associations with DRB1 to include mutual robust associations with several HLA genes.

Implications for individuals

A major outcome of this study was the derivation of a unique, continuously-varying measure assessing the protection/susceptibility of an individual to multiple sclerosis (ξ) or brain cancer (ζ), based on their specific genetic HLA makeup. The allele-based MS-HLA and BC-HLA PScov scores provided the building blocks of ξ or ζ, which is the average of the 12 PScov scores of the HLA alleles carried by an individual (equations (2) and (3)). We called this approach the “crowdfunding” nature of HLA contribution to disease,⁵⁴ based on the following considerations: (1) the HLA region is the most polymorphic of the entire human genome, (2) every individual possesses 12 HLA alleles, and (3) each HLA allele may differ in terms of the direction and magnitude of its association with a given disease, as captured by the PScov score. Consequently, rather than focusing on “unique” contribution of specific alleles on disease susceptibility and protection, the primary emphasis of our crowdfunding approach is on the combined effects of relatively small contributions of individual alleles which, as a set, will contribute to overall protection/susceptibility of the disease at hand in individuals and populations. This approach is more realistic (given that individuals carry 12 HLA alleles) and different from traditional attempts to identify specific alleles in HLA-disease associations.⁵⁵ Actually, the effects of such alleles, when found, would be embedded and integrated with those of other alleles co-existing in the particular individual. This situation may explain why alleles with potentially strong effects themselves are not commonly found, since any HLA-disease association would be always based on contributions by the 12 classical HLA alleles, and of other alleles of non-classical genes (HLA-E, -G, -F), as they may be present in some individuals.

Immunogenetics and the role of viruses in multiple sclerosis and brain cancer

Considering the role of HLA in foreign antigen elimination, the highly significant correlation between immunogenetic profiles of MS and brain cancer supports an influence of exposure to foreign antigens in both conditions. Indeed, several common viruses including human herpes viruses such as Epstein-Barr virus and cytomegalovirus have been implicated in both MS and brain cancer although findings have been mixed.^56,57 Like human herpes viruses, infection with John Cunningham virus (JCV) is nearly ubiquitous,⁵⁸ and JCV similarly persists in a latent state detectable in normal human brain tissue.⁵⁹ While JCV has been primarily associated with progressive multifocal leukoencephalopathy, it has also been implicated in MS⁶⁰ and in complications related to immunosuppressive treatment for MS,^61,62 as well as in several neoplasms^63,64 including brain cancer.^64,65 Moreover, JCV DNA has been found in MS plaques and in tumor cells of a patient with concurrent MS and brain cancer.⁶⁶ Our findings concur with previous literature regarding a possible common viral contribution to MS and brain cancer and suggest their association may be partially mediated by similar immunogenetic effects. Nonetheless, definitive identification of specific virus(es) involved in both conditions remains to be determined. HLA binding and elimination of viral antigens that have been linked to MS and brain cancer is an interesting avenue for future investigation, particularly in light of mixed findings from studies evaluating effects of anti-viral therapies and preventative vaccines on both conditions.^56,57

Increased risk of brain cancer in MS patients has been partially attributed to incidental findings due to increased frequency of brain imaging in MS patients.⁹ While possible, our study documenting high correlations between the immunogenetic profiles of MS and brain cancer suggests an alternative explanation based on the human immune response to non-self antigens such as viruses and cancer neoantigens. Furthermore, the present findings identify HLA alleles that are mutually protective against both conditions, mutually susceptible, or in show discordant associations with MS and brain cancer. In the future, these findings may be leveraged to develop personalized precision adjunctive treatments for MS, brain cancer, and other immune-involved conditions based on an individual’s HLA composition.⁶⁷

Conclusion and Limitations

The present study provides a novel perspective regarding the link between MS and brain cancer; however, several limitations must be considered. First, HLA-disease associations may differ for different forms of MS and/or different types of brain cancers. Indeed, a recent study documented HLA influence on viral antigen response in glioma subtypes.⁶⁸ Here, we evaluated the immunogenetic overlap of MS and brain cancers as a whole and are unable to draw any conclusions about immunogenetic overlap between MS and brain cancer subtypes. Second, the findings are based on population-level HLA-disease associations. We utilized those population-based findings to estimate individual-level HLA protection and susceptibility for MS and brain cancer; however, those estimates remain to be confirmed. Further, since HLA varies across populations,^38,39 additional studies are warranted to determine whether the present findings extend beyond Continental Western Europe to other regions. Finally, another limitation of the study is that alleles of genes DRB3, DRB4 and DRB5 were not used since they do not occur in all individuals.

Footnotes

Funding:

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Partial funding for this study was provided by the University of Minnesota (the Anita Kunin Chair in Women’s Healthy Brain Aging, the Brain and Genomics Fund, the McKnight Presidential Chair of Cognitive Neuroscience, and the American Legion Brain Sciences Chair) and the U.S. Department of Veterans Affairs. The sponsors had no role in the current study design, analysis or interpretation, or in the writing of this paper. The contents do not represent the views of the U.S. Department of Veterans Affairs or the United States Government.

Declaration of conflicting interests:

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Abbreviations

BC, Brain cancer

HLA, human leukocyte antigen

MS, Multiple sclerosis

Author Contributions

LMJ extracted disease prevalence data. APG extracted allele frequencies and performed data analysis. LMJ and APG wrote and reviewed the paper.

Ethical Approval

This article does not contain any studies with human participants performed by any of the authors.

ORCID iDs

Lisa M James

Apostolos P Georgopoulos

Data Availability

All data used were retrieved from freely accessible websites, as mentioned in Methods, and, as such, are publicly and freely available. Specifically:

Notes

References

Varadé

Magadán

González-Fernández

Á.

Human immunology and immunotherapy: main achievements and challenges. Cell Mol Immunol. 2021;18:805-828.

Cools

Ponsaerts

Van Tendeloo

Berneman

ZN.

Regulatory T cells and human disease. Clin Dev Immunol. 2007;2007:89195.

Ghajarzadeh

Mohammadi

Sahraian

MA.

Risk of cancer in multiple sclerosis (MS): A systematic review and meta-analysis. Autoimmun Rev. 2020;19:102650.

Kingwell

Bajdik

Phillips

, et al Cancer risk in multiple sclerosis: findings from British Columbia, Canada. Brain. 2012;135:2973-2979.

Marrie

Reider

Cohen

, et al A systematic review of the incidence and prevalence of cancer in multiple sclerosis. Mult Scler. 2015;21:294-304.

Bahmanyar

Montgomery

Hillert

Ekbom

Olsson

. Cancer risk among patients with multiple sclerosis and their parents. Neurology. 2009;72:1170-1177.

Kyritsis

Boussios

Pavlidis

Cancer specific risk in multiple sclerosis patients. Crit Rev Oncol Hematol. 2016;98:29-34.

Lebrun

Debouverie

Vermersch

, et al Cancer risk and impact of disease-modifying treatments in patients with multiple sclerosis. Mult Scler. 2008;14:399-405.

Hemminki

Liu

Forsti

, et al Subsequent brain tumors in patients with autoimmune disease. Neurooncol. 2013;15:1142-1150.

10.

Dendrou

Petersen

Rossjohn

Fugger

. HLA variation and disease. Nat Rev Immunol. 2018;18:325-339.

11.

Hov

Kosmoliaptsis

Traherne

, et al Electrostatic modifications of the human leukocyte antigen-DR P9 peptide-binding pocket and susceptibility to primary sclerosing cholangitis. Hepatology. 2011;53:1967-1976.

12.

Davenport

Quinn

Chicz

, et al Naturally processed peptides from two disease-resistance-associated HLA-DR13 alleles show related sequence motifs and the effects of the dimorphism at position 86 of the HLA-DR beta chain. Proc Natl Acad Sci USA. 1995;92:6567-6571.

13.

Gough

Simmonds

MJ.

The HLA region and autoimmune disease: associations and mechanisms of action. Curr Genomics. 2007;8:453-465.

14.

Hollenbach

Oksenberg

. The immunogenetics of multiple sclerosis: A comprehensive review. J Autoimmun. 2015;64:13-25.

15.

The International Multiple Sclerosis Genetics Consortium & The Wellcome Trust Case Control Consortium 2. Genetic risk and a primary role for cell-mediated immune mechanisms in multiple sclerosis. Nature. 2011;476:214-219.

16.

Sawcer

Franklin

Ban

. Multiple sclerosis genetics. Lancet Neurol. 2014;13:700-709.

17.

Baranzini

Oksenberg

. The genetics of multiple sclerosis: from 0 to 200 in 50 years. Trends Genet. 2017;33:960-970.

18.

Sospedra

Martin

Immunology of multiple sclerosis. Annu Rev Immunol. 2005;23:683-747.

19.

James

Georgopoulos

. Immunogenetic epidemiology of dementia and Parkinson’s disease in 14 continental European countries: shared human leukocyte antigen (HLA) profiles. J Immunol Sci. 2021;5:16-26.

20.

James

Georgopoulos

. The human leukocyte antigen (HLA) DRB1*13:02 allele protects against dementia in continental Western Europe. J Neurol -med. 2019;4:1-6.

21.

James

Georgopoulos

. Tri-allelic human leukocyte antigen (HLA) protection against dementia. J Neurol Neuromed. 2020;5:12-17.

22.

James

Georgopoulos

. Shared human leukocyte antigen (HLA) coverage in dementia and Parkinson’s disease. J Neurol Neuromed. 2020;5:45-54.

23.

James

Georgopoulos

. Immunogenetic epidemiology of Type 1 diabetes in 14 Continental Western European countries. J Immunol Sci. 2021;5:29-35.

24.

James

Georgopoulos

. Immunogenetic epidemiology of multiple sclerosis in 14 Continental Western European countries. J Immunol Sci. 2021;5:40-46.

25.

James

Georgopoulos

. Immunogenetic clustering of 30 cancers. Sci Rep. 2022;12:7235.

26.

James

Georgopoulos

. Immunogenetic profiles and associations of breast, cervical, ovarian, and uterine cancers. Cancer Inform. 2023;22:11769351221148588.

27.

Georgopoulos

James

. Immunogenetic Association of 127 Human Leukocyte Antigen (HLA) alleles with 30 cancers in continental Western European countries. J Immunol Sci. 2022;6:6-17.

28.

Rashkin

Graff

Kachuri

, et al Pan-cancer study detects genetic risk variants and shared genetic basis in two large cohorts. Nat Commun. 2020;11:4423.

29.

Sun

Lemaçon

Legault

M-A

, et al Genetic meta-analysis of cancer diagnosis following statin use identifies new associations and implicates human leukocyte antigen (HLA) in women. Pharmacogenomics. 2021;21:446-457.

30.

Song

Ruder

, et al Genetic epidemiology of glioblastoma multiforme: confirmatory and new findings from analyses of human leukocyte antigen alleles and motifs. PLoS One. 2009;4:e7157.

31.

Tang

Shao

Dorak

, et al Positive and negative associations of human leukocyte antigen variants with the onset and prognosis of adult glioblastoma multiforme. Cancer Epidemiol Biomarkers Prev. 2005;14:2040-2044.

32.

La Torre

Maugeri

Angileri

, et al Human leukocyte antigen frequency in human high-grade gliomas: A case-control study in Sicily. Neurosurg. 2009;64:1082-1089.

33.

Zhang

de Smith

Smirnov

, et al Non-additive and epistatic effects of HLA polymorphisms contributing to risk of adult glioma. J Neurooncol. 2017;135:237-244.

34.

Han

Deng

Wang

, et al Decreased human leukocyte antigen A*02:01 frequency is associated with risk of glioma and existence of human cytomegalovirus: a case–control study in northern China. Cancer Immunol Immunother. 2017;66:1265-1273.

35.

Choi

S-S

Choi

Baek

I-C

, et al HLA polymorphisms and risk of glioblastoma in Koreans. PLoS One. 2021;16:e0260618.

36.

Bassig

Inskip

Burdette

, et al Selected human leukocyte antigen class II polymorphisms and risk of adult glioma. J Neuroimmunol. 2011;233:185-191.

37.

Machulla

Steinborn

Schaaf

Heidecke

Rainov

. Brain glioma and human leukocyte antigens (HLA)–is there an association. J Neurooncol. 2001;52:253-261.

38.

Garamszegi

. Global distribution of malaria-resistant MHC-HLA alleles: the number and frequencies of alleles and malaria risk. Malar J. 2014;13:349.

39.

Singh

Kaul

Khan

. A comparative review of HLA associations with hepatitis B and C viral infections across global populations. World J Gastroenterol. 2007;13:1770-1787.

40.

Global Burden of Disease Collaborative Network. (2016) Global Burden of Disease Study 2016 (GBD 2016) Results. Institute for Health Metrics and Evaluation (IHME), 2020. Accessed July 5, 2021. http://ghdx.healthdata.org/gbd-results-tool.

41.

Population Reference Bureau. 2016 World Population Data Sheet With a Special Focus on Human Needs and Sustainable Resources. Population Reference Bureau, 2016. Accessed February 5, 2019. https://www.prb.org/2016-world-population-data-sheet/

42.

Hurley

Kempenich

Wadsworth

, et al Common, intermediate and well-documented HLA alleles in world populations: CIWD version 3.0.0. HLA. 2020;95:516-531.

43.

Tukey

. Exploratory Data Analysis. Pearson; 1977.

44.

Efron

Tibshirani

. An Introduction to the Bootstrap. Chapman & Hall; 1993.

45.

Harbo

Lie

Sawcer

, et al Genes in the HLA class I region may contribute to the HLA class II-associated genetic susceptibility to multiple sclerosis. Tissue Antigens. 2004;63:237-247.

46.

Fogdell-Hahn

Ligers

Grønning

Hillert

Olerup

. Multiple sclerosis: a modifying influence of HLA class I genes in an HLA class II associated autoimmune disease. Tissue Antigens. 2000;55:140-148.

47.

Zhang

Lin

Dong

Wang

. Relationship between HLA-DRB1 polymorphism and susceptibility or resistance to multiple sclerosis in Caucasians: a meta-analysis of non-family-based studies. Autoimmun Rev. 2011;10:474-481.

48.

Chen

Huang

She

, et al HLA-DRB108, DRB103/DRB30101, and DRB30202 are susceptibility genes for Graves’ disease in North American Caucasians, whereas DRB107 is protective. J Clin Endocrinol Metab. 1999;84:3182-3186.

49.

Kula

Bednarczuk

Jurecka-Lubieniecka

, et al Interaction of HLA-DRB1 alleles with CTLA-4 in the predisposition to Graves' disease: the impact of DRB1*07. Thyroid. 2006;16:447-453.

50.

Castro-Santos

Olloquequi

Verdugo

, et al HLA-DRB1*07:01 and *08:02 alleles confer a protective effect against ACPA-positive rheumatoid arthritis in a Latin American admixed population. Biology. 2020;9:467.

51.

Barton

Worthington

, et al HLA-Cw6 and HLA-DRB1*07 together are associated with less severe joint disease in psoriatic arthritis. Ann Rheum Dis. 2007;66:807-811.

52.

Muñiz-Castrillo

Ambati

Dubois

, et al Primary DQ effect in the association between HLA and neurological syndromes with anti-GAD65 antibodies. J Neurol. 2020;267:1906-1911.

53.

Erlich

Valdes

Noble

, et al; Type 1 Diabetes Genetics Consortium. HLA DR-DQ haplotypes and genotypes and type 1 diabetes risk: analysis of the type 1 diabetes genetics consortium families. Diabetes. 2008;57:1084-1092.

54.

James

Charonis

Georgopoulos

. Schizophrenia, human leukocyte antigen (HLA), and herpes viruses: immunogenetic associations at the population level. Neurosci Insights. 2023;18:26331055231166411-26331055231166412.

55.

Blackwell

Jamieson

Burgner

. HLA and infectious diseases. Clin Microbiol Rev. 2009;22:370-385.

56.

Bar-Or

Pender

Khanna

, et al Epstein-Barr virus in multiple sclerosis: theory and emerging immunotherapies. Trends Mol Med. 2020;26:296-310.

57.

Akhtar

Vranic

Cyprian

Al Moustafa

AE.

Epstein-barr virus in gliomas: cause, association, or artifact?

Front Oncol. 2018;8:123.

58.

Shackelton

Rambaut

Pybus

Holmes

. JC virus evolution and its association with human populations. J Virol. 2006;80:9928-9933.

59.

White

3rd Ishaq

Stoner

Frisque

. JC virus DNA is present in many human brain samples from patients without progressive multifocal leukoencephalopathy. J Virol. 1992;66:5726-5734.

60.

Stoner

. Polyomavirus models of brain infection and the pathogenesis of multiple sclerosis. Brain Pathol. 1993;3:213-227.

61.

Outteryck

Zéphir

Salleron

, et al JC-virus seroconversion in multiple sclerosis patients receiving natalizumab. Mult Scler. 2014;20:822-829.

62.

Koendgen

Campbell

, et al Risk of natalizumab-associated progressive multifocal leukoencephalopathy in patients with multiple sclerosis: a retrospective analysis of data from four clinical studies. Lancet Neurol. 2017;16:925-933.

63.

Pinto

Dobson

. BK and JC virus: a review. J Infect. 2014;68:S2-S8.

64.

Imperiale

. The human polyomaviruses, BKV and JCV: molecular pathogenesis of acute disease and potential role in cancer. Virology. 2000;267:1-7.

65.

Goudarzi far

Tambrchi

Nahid samiei

, et al Association between human polyomavirus infection and brain cancer: A systematic review and meta-analysis. Microb Pathog. 2022;173:105797.

66.

Del Valle

Delbue

Gordon

, et al Expression of JC virus T-antigen in a patient with MS and glioblastoma multiforme. Neurology. 2002;58:895-900.

67.

Georgopoulos

James

Charonis

Sanders

Melanoma and Human Leukocyte Antigen (HLA): Immunogenicity of 69 HLA Class I alleles with 11 antigens expressed in melanoma tumors. Cancer Inform. 2023;22:1-13. doi:10.1177/11769351231172604

68.

Guerra

Kachuri

Wendt

, et al The immunogenetics of viral antigen response is associated with subtype-specific glioma risk and survival. Am J Hum Genet. 2022;109:1105-1116.

Positive Association Between the Immunogenetic Human Leukocyte Antigen (HLA) Profiles of Multiple Sclerosis and Brain Cancer

Abstract

Keywords

Significance Statement

Introduction

Materials and Methods

Prevalence of multiple sclerosis and brain cancer

HLA alleles

Immunogenetic (HLA) profiles of MS and BC

Random permutations test for assessing the statistical significance of the MS and BC HLA PScov profiles

Application to individuals

Implementation of analysis procedures

Results

Prevalences

HLA MS- and BC PScov scores

Distributions of MS and BC PScov scores

Associations of MS and BC PScov scores

Application to individuals

Discussion

Multiple sclerosis and brain cancer immunogenetic overlap

Implications for individuals

Immunogenetics and the role of viruses in multiple sclerosis and brain cancer

Conclusion and Limitations

Footnotes

Funding:

Declaration of conflicting interests:

Abbreviations

Author Contributions

Ethical Approval

ORCID iDs

Data Availability

Notes

References