REPLICATION OF PREVIOUS ASSOCIATION FINDINGS

Studies II to IV utilized the DNA samples from the FINNAKI genetic study cohort. The demographics and clinical characteristics of the cohort grouped according to AKI severity stage are detailed in Table 11.

Table 11.ĞŵŽŐƌĂƉŚŝĐƐĂŶĚĐůŝŶŝĐĂůĐŚĂƌĂĐƚĞƌŝƐƟĐƐĨŽƌ&/EE</ŐĞŶĞƟĐĐŽŚŽƌƚ ŐƌŽƵƉĞĚĂĐĐŽƌĚŝŶŐƚŽĂĐƵƚĞŬŝĚŶĞLJŝŶũƵƌLJ;</ͿƐĞǀĞƌŝƚLJƐƚĂŐĞ͘ĂƚĂĂƌĞƉƌĞƐĞŶƚĞĚĂƐ

5 Diabetes 2842 111 (23%) 59 (25%) 151 (32%) 311 (19%) Arteriosclerosis 2813 78 (16%) 45 (19%) 70 (15%) 176 (11%) ŚƌŽŶŝĐŽďƐƚƌƵĐƟǀĞ

ƉƵůŵŽŶĂƌLJĚŝƐĞĂƐĞ ^{2824 50 (10%) 26 (11%) 29 (6%) 142 (9%)} ŚƌŽŶŝĐůŝǀĞƌĚŝƐĞĂƐĞ 2809 14 (3%) 16 (7%) 36 (8%) 53 (3%) Systolic heart failure 2820 62 (13%) 42 (18%) 60 (13%) 161 (10%)

ĂƐĞůŝŶĞƉůĂƐŵĂ

ACE inhibitor or ARB 2783 198 (42%) 103 (44%) 217 (46%) 514 (32%)

NSAID 2730 46 (10%) 28 (12%) 51 (11%) 121 (8%) ACE inhibitor or ARB 2792 140 (29%) 70 (30%) 137 (30%) 360 (22%) ŝƵƌĞƟĐƐ 2760 166 (35%) 101 (43%) 182 (40%) 395 (25%) SAPS II 24h without

ƌĞŶĂůŽƌĂŐĞ

ĐŽŵƉŽŶĞŶƚƐ ^{2812 23 (16–31) 24 (16–32) 24 (16–33) 20 (13–29)} DĞĐŚĂŶŝĐĂůǀĞŶƟůĂƟŽŶ 2846 376 (77%) 166 (70%) 316 (66%) 1069 (65%)

^ĞƉƐŝƐ 2846 211 (43%) 109 (46%) 240 (50%) 372 (23%) Chronic kidney disease 2846 48 (10%) 18 (8%) 73 (15%) 60 (4%)

ϱ͘ϯ͘ϭƉŽƉƚŽƐŝƐͲƌĞůĂƚĞĚǀĂƌŝĂŶƚƐ;//Ϳ

In Study II, analyses were performed in patients with septic shock, patients with sepsis, and in the entire cohort.

The entire cohort of Study II consisted of 2146 patients, of whom 601 had severe AKI and 1545 did not have AKI. The genetic variants had no association with AKI either in the univariate or the multivariable model.

Š‡ ϐ‹†‹‰• ’”‡˜ƒ‹Ž‡† ‹ †‘‹ƒ– ƒ† ”‡…‡••‹˜‡ ‰‡‡–‹… –‡•– ‘†‡Ž•Ǥ Differences in allele frequencies between cases and controls in the entire cohort are illustrated in Figure 8.

&ŝŐƵƌĞϴ͘DŝŶŽƌĂůůĞůĞĨƌĞƋƵĞŶĐŝĞƐĨŽƌŐŝǀĞŶǀĂƌŝĂŶƚƐ;LJͲĂdžŝƐͿĨŽƌĐĂƐĞƐ;ďůƵĞͿĂŶĚ ĐŽŶƚƌŽůƐ;ŐƌĞĞŶͿŝŶƚŚĞĞŶƟƌĞĐŽŚŽƌƚŽĨƉĂƟĞŶƚƐ͕͘ĐLJƚŽƐŝŶĞ͖'͕ŐƵĂŶŝŶĞ͖͕ĂĚĞŶŝŶĞ͖

d͕ƚŚLJŵŝŶĞ͘

There were 653 patients with sepsis in Study II, of whom 299 had

•‡˜‡”‡ƒ†͵ͷͶŠƒ†‘Ǥ‘‡‘ˆ–Š‡ϐ‹˜‡‰‡‡–‹…˜ƒ”‹ƒ–•ƒ••‘…‹ƒ–‡†

•‹‰‹ϐ‹…ƒ–Ž› ™‹–Š ‹ ƒ ƒ††‹–‹˜‡ ‰‡‡–‹… –‡•– ‘†‡ŽǤ †Œ—•–‹‰ ™‹–Š

covariates did not change the result. Figure 9 illustrates the allele

ˆ”‡“—‡…‹‡•‹…ƒ•‡•ƒ†…‘–”‘Ž•™‹–Š•‡’•‹•ǤŠ‡•‡ϐ‹†‹‰•’”‡˜ƒ‹Ž‡†‹

the recessive and dominant genetic test models.

5

&ŝŐƵƌĞϵ͘DŝŶŽƌĂůůĞůĞĨƌĞƋƵĞŶĐŝĞƐĨŽƌŐŝǀĞŶǀĂƌŝĂŶƚƐ;LJͲĂdžŝƐͿĨŽƌĐĂƐĞƐ;ďůƵĞͿĂŶĚ ĐŽŶƚƌŽůƐ;ŐƌĞĞŶͿŝŶƉĂƟĞŶƚƐǁŝƚŚƐĞƉƐŝƐ͕͘ĐLJƚŽƐŝŶĞ͖'͕ŐƵĂŶŝŶĞ͖͕ĂĚĞŶŝŶĞ͖d͕ƚŚLJŵŝŶĞ͘

Study II included 478 patients with septic shock. Of these patients, 252 had severe AKI and 226 did not have AKI. In a multivariate analysis, with an additive genetic test model the rs2093266 A-allele in SERPINA4 and A-allele in rs1955656 in SERPINA5 that is in complete linkage disequilibrium with

”•ʹͲͻ͵ʹ͸͸ Šƒ† ‘††• ”ƒ–‹‘• ȋȌ ‘ˆ ͲǤ͸͵ ȋͻͷΨ …‘ϐ‹†‡…‡ ‹–‡”˜ƒŽ ȋȌ 0.40 to 0.98, p-value 0.043) with the outcome of severe AKI. The protective association prevailed in the dominant adjusted model (OR 0.59, p-value 0.034), but not in the recessive genetic test model. It is noteworthy that the A-alleles were found to be protective in the original study as well. The allele frequencies for cases and controls in septic shock are illustrated in Figure 10.

&ŝŐƵƌĞϭϬ͘DŝŶŽƌĂůůĞůĞĨƌĞƋƵĞŶĐŝĞƐĨŽƌŐŝǀĞŶǀĂƌŝĂŶƚƐ;LJͲĂdžŝƐͿĨŽƌĐĂƐĞƐ;ďůƵĞͿĂŶĚ ĐŽŶƚƌŽůƐ;ŐƌĞĞŶͿŝŶƉĂƟĞŶƚƐŝŶƐĞƉƟĐƐŚŽĐŬ͕͘ĐLJƚŽƐŝŶĞ͖'͕ŐƵĂŶŝŶĞ͖͕ĂĚĞŶŝŶĞ͖d͕ƚŚLJŵŝŶĞ͘

ϱ͘ϯ͘Ϯ/ƌŽŶŵĞƚĂďŽůŝƐŵͲƌĞůĂƚĞĚǀĂƌŝĂŶƚ;///Ϳ

In Study III, the GT (guanine-thymine) dinucleotide repeat polymorphism in the HMOX1 promoter region was successfully genotyped in 653 septic patients. Of these patients, 300 had severe AKI and 353 had no AKI. The most common GT repeat numbers were 30 (48.5%) or 23 (16.8%) repeats ranging from 16 to 39.

Patient distribution between genotype classes as well as AKI event rate and tests for trend are presented in Table 12. In addition, the table presents the median concentration of HO-1 enzyme within each genotype group. Samples from 601 patients were tested for HO-1 concentration. The

‡†‹ƒ Ǧͳ ‡œ›‡ …‘…‡–”ƒ–‹‘ ™ƒ• •‹‰‹ϐ‹…ƒ–Ž› Ž‘™‡” ‹ ’ƒ–‹‡–•

with at least one L2-allele (1.0ng/mL vs 1.8ng/mL, p-value <0.001).

Table 12.ZĞƐƵůƚƐƉƌĞƐĞŶƚĞĚĨŽƌƉĂƟĞŶƚƐǁŝƚŚ</ĚŝƐƚƌŝďƵƚĞĚĂĐĐŽƌĚŝŶŐƚŽƚǁŽƚLJƉĞƐ

5 Furthermore, the HO-1 concentration was elevated in patients with

AKI compared to patients without AKI (2.2ng/mL vs 1.7ng/mL, p-value 0.001). The association between lower HO-1 concentration and L2-allele

™ƒ•˜‡”‹ϐ‹‡†‹’ƒ–‹‡–•™‹–ŠȋͳǤͲ‰Ȁ˜•ʹǤͶ‰Ȁǡp-value 0.004), as well as without AKI (1.0ng/mL vs 1.8ng/mL, p-value <0.001).

In logistic regression with an additive genetic test model, the S-allele conferred risk for AKI (OR 1.29, 95% CI 1.02 to 1.64, p-value 0.034).

In an adjusted model the result prevailed (OR 1.30, 95% CI 1.01 to 1.66, pǦ˜ƒŽ—‡ ͲǤͲͶͳȌǤ Š‹• ƒ••‘…‹ƒ–‹‘ ™ƒ• •‹‰‹ϐ‹…ƒ– ‹ —‹˜ƒ”‹ƒ–‡ ƒƒŽ›•‹• ‹

a recessive genetic test model, but not after adjusting or in a dominant genetic test model.

ϱ͘ϯ͘ϯ/ŶŇĂŵŵĂƟŽŶͲƌĞůĂƚĞĚǀĂƌŝĂŶƚƐ;/sͿ

Ž–‘‰‡–Š‡” ʹ͹ ˜ƒ”‹ƒ–• ™‡”‡ •–—†‹‡†ǡ ‹ ͳͺ ‹ϐŽƒƒ–‹‘ ”‡Žƒ–‡† ‰‡‡•

(TNFA; IL6; CXCL8, interleukin 8; IL10; NOS3; NFKB1A, nuclear factor of kappa light polypeptide gene enhancer in B-cells inhibitor, alpha; AGT;

VEGFA; EPO; SUFU; HIF1A; PNMT; MPO; COMT; HSPB1, heat shock protein family B (small) member 1; SFTPD, surfactant protein D; HAMP, hepcidin antimicrobial peptide; BBS9, Bardet-Biedl syndrome 9). Two of the tested variants in IL6 (rs2069842 and rs2069830) were not polymorphic in

–Š‹•…‘Š‘”–Ǥ‘‡‘ˆ–Š‡–‡•–‡†˜ƒ”‹ƒ–••‹‰‹ϐ‹…ƒ–Ž›ƒ••‘…‹ƒ–‡†™‹–Š

in the primary, secondary, or in tertiary analysis. In the primary analysis there were 2207 patients, 625 with severe AKI (KDIGO stages 2 or 3) and 1582 without AKI. In the secondary analysis 2647 patients were included, with 1065 patients with AKI (KDIGO stages 1, 2 or 3) and 1582 patients without AKI. In the tertiary analysis, patients with chronic kidney disease (n=199) were included (altogether 2846 patients) and the analysis was adjusted for covariates (chronic liver failure, body mass index (BMI), use of

‘•–‡”‘‹†ƒŽƒ–‹Ǧ‹ϐŽƒƒ–‘”›†”—‰•ȋȌƒ•’‡”ƒ‡–‡†‹…ƒ–‹‘ǡ use of warfarin as permanent medication, use of contrast medium, colloids or albumin prior to admission, minimum platelet count on admission,

•‹’Ž‹ϐ‹‡†ƒ…—–‡’Š›•‹‘Ž‘‰›•…‘”‡ȋȌ™‹–Š‘—–ƒ‰‡ƒ†”‡ƒŽ’‘‹–•ǡ cardiac surgery status, and sepsis status).

Ž–Š‘—‰Š‹•‹‰‹ϐ‹…ƒ–ǡ–Š‡ˆ‘ŽŽ‘™‹‰”‡•—Ž–•‘ˆ–—†›ƒ”‡’”‡•‡–‡†

”‡‰ƒ”†‹‰–Š‡†‹”‡…–‹‘‘ˆ–Š‡ϐ‹†‹‰•Ǥ

Variants rs1800629 in TNFA and rs1800896 in IL10 have been previously reported to be associated with AKI in several studies, while others have refuted the association. These SNPs did not associate with AKI in Study IV (Figure 11). No evidence of association was detected even when testing interaction in the epistasis test.

&ŝŐƵƌĞϭϭ͘KĚĚƐƌĂƟŽƐ;ďŽdžĞƐͿĂŶĚϵϱйĐŽŶĮĚĞŶĐĞŝŶƚĞƌǀĂůƐ;ǁŚŝƐŬĞƌƐͿĨŽƌŵŝŶŽƌ ĂůůĞůĞƐŽĨǀĂƌŝĂŶƚƐƌƐϭϴϬϬϲϮϵĂŶĚƌƐϭϴϬϬϴϵϲ͘/ŶĂŶĂůLJƐŝƐϭ͕ƉĂƟĞŶƚƐǁŝƚŚƐĞǀĞƌĞĂĐƵƚĞ ŬŝĚŶĞLJŝŶũƵƌLJ;</ͿǁĞƌĞĐŽŵƉĂƌĞĚƚŽƉĂƟĞŶƚƐǁŝƚŚŽƵƚ</͕ŽŵŝƫŶŐƉĂƟĞŶƚƐǁŝƚŚ ĐŚƌŽŶŝĐŬŝĚŶĞLJĚŝƐĞĂƐĞ;<Ϳ͘/ŶĂŶĂůLJƐŝƐϮ͕ĂůůƐƚĂŐĞ</ǁĂƐĐŽŵƉĂƌĞĚƚŽŶŽ</͕

ŽŵŝƫŶŐƉĂƟĞŶƚƐǁŝƚŚ<͘/ŶĂŶĂůLJƐŝƐϯ͕ƉĂƟĞŶƚƐǁŝƚŚƐĞǀĞƌĞ</ǁĞƌĞĐŽŵƉĂƌĞĚƚŽ ƉĂƟĞŶƚƐǁŝƚŚŽƵƚ</͕ŝŶĐůƵĚŝŶŐƉĂƟĞŶƚƐǁŝƚŚ<ŝŶĂŶĂĚũƵƐƚĞĚĂŶĂůLJƐŝƐ͕͘ĂĚĞŶŝŶĞ͖

'͕ŐƵĂŶŝŶĞ͖dE&͕ƚƵŵŽƌŶĞĐƌŽƐŝƐĨĂĐƚŽƌĂůƉŚĂ͖/>ϭϬ͕ŝŶƚĞƌůĞƵŬŝŶϭϬ͘

5

‘‡‘ˆ–Š‡ϐ‹†‹‰•ǡƒŽ„‡‹–‹•‹‰‹ϐ‹…ƒ–ǡ™‡”‡…‘–”ƒ†‹…–‘”›–‘’”‡˜‹‘—•

associations. When risk addition has been suggested in VEGFA rs3025039 C-allele, and COMT rs4680 A-allele, and risk protection in SUFU rs10748825 G-allele, and PNMT rs876493 A-allele. The opposite was found in Study IV (Figure 12).

&ŝŐƵƌĞϭϮ͘KĚĚƐƌĂƟŽƐ;ďŽdžĞƐͿĂŶĚϵϱйĐŽŶĮĚĞŶĐĞŝŶƚĞƌǀĂůƐ;ǁŚŝƐŬĞƌƐͿĨŽƌŵŝŶŽƌ ĂůůĞůĞƐŽĨǀĂƌŝĂŶƚƐƌƐϯϬϮϱϬϯϵ͕ƌƐϭϬϳϰϴϴϮϱ͕ƌƐϴϳϲϰϵϯ͕ĂŶĚƌƐϰϲϴϬ͘/ŶĂŶĂůLJƐŝƐϭ͕

ƉĂƟĞŶƚƐǁŝƚŚƐĞǀĞƌĞĂĐƵƚĞŬŝĚŶĞLJŝŶũƵƌLJ;</ͿǁĞƌĞĐŽŵƉĂƌĞĚƚŽƉĂƟĞŶƚƐǁŝƚŚŽƵƚ

</͕ŽŵŝƫŶŐƉĂƟĞŶƚƐǁŝƚŚĐŚƌŽŶŝĐŬŝĚŶĞLJĚŝƐĞĂƐĞ;<Ϳ͘/ŶĂŶĂůLJƐŝƐϮ͕ĂůůͲƐƚĂŐĞ</

ǁĂƐĐŽŵƉĂƌĞĚƚŽŶŽ</͕ŽŵŝƫŶŐƉĂƟĞŶƚƐǁŝƚŚ<͘/ŶĂŶĂůLJƐŝƐϯ͕ƉĂƟĞŶƚƐǁŝƚŚ ƐĞǀĞƌĞ</ǁĞƌĞĐŽŵƉĂƌĞĚƚŽƉĂƟĞŶƚƐǁŝƚŚŽƵƚ</͕ŝŶĐůƵĚŝŶŐƉĂƟĞŶƚƐǁŝƚŚ<ŝŶĂŶ ĂĚũƵƐƚĞĚĂŶĂůLJƐŝƐ͘d͕ƚŚLJŵŝŶĞ͖'͕ŐƵĂŶŝŶĞ͖͕ĐLJƚŽƐŝŶĞ͖s'&͕ǀĂƐĐƵůĂƌĞŶĚŽƚŚĞůŝĂů ŐƌŽǁƚŚĨĂĐƚŽƌ͖^h&h͕ƐƵƉƉƌĞƐƐŽƌŽĨĨƵƐĞĚŚŽŵŽůŽŐ͖WEDd͕ƉŚĞŶLJůĞƚŚĂŶŽůĂŵŝŶĞ EͲŵĞƚŚLJůƚƌĂŶƐĨĞƌĂƐĞ͖KDd͕ĐĂƚĞĐŚŽůͲKͲŵĞƚŚLJůƚƌĂŶƐĨĞƌĂƐĞ͘

In accordance with previous research, in Study IV the variants rs1800796 (C) in IL6, rs4073 (A) in CXCL8, rs2070744 (C) in NOS3, rs1050851 (A) in NFKB1A, rs2010963 (C) in VEGFA, rs7208693 (A) in MPO, rs10262995 (T) in BBS9, rs1617640 (G) in EPO, and rs2243639 (T) in SFTPD were found to have an odds ratio (OR) less than 1 in the primary analysis, and thus were assumed to confer protection rather than risk. Variants rs721917 (G) in SFTPD, rs1800795 (G) in IL6, rs699 (C) in AGT, and rs11549465 (T) in HIF1A had an OR of 1 or more in the primary analysis, and thus did not …‘–”ƒ†‹…–™‹–Š’”‡˜‹‘—•ϐ‹†‹‰••—‰‰‡•–‹‰”‹•ƒ††‹–‹‘Ǥ‘™‡˜‡”ǡƒ›

˜ƒŽ—‡• ™‡”‡ ‡š–”‡‡Ž› …Ž‘•‡ –‘ ͳǡ ƒ† ϐŽ—…–—ƒ–‡† „‡–™‡‡ ƒƒŽ›•‡•

(Figure 13).

&ŝŐƵƌĞϭϯ͘KĚĚƐƌĂƟŽƐ;ďŽdžĞƐͿĂŶĚϵϱйĐŽŶĮĚĞŶĐĞŝŶƚĞƌǀĂůƐ;ǁŚŝƐŬĞƌƐͿĨŽƌŵŝŶŽƌ ĂůůĞůĞƐŽĨǀĂƌŝĂŶƚƐƌƐϭϴϬϬϳϵϲ͕ƌƐϭϴϬϬϳϵϱ͕ƌƐϰϬϳϯ͕ƌƐϮϬϳϬϳϰϰ͕ƌƐϭϬϱϬϴϱϭ͕ƌƐϲϵϵ͕

ƌƐϮϬϭϬϵϲϯ͕ƌƐϭϲϭϳϲϰϬ͕ƌƐϭϭϱϰϵϰϲϱ͕ƌƐϳϮϬϴϲϵϯ͕ƌƐϮϮϰϯϲϯϵ͕ƌƐϳϮϭϵϭϳ͕ĂŶĚ ƌƐϭϬϮϲϮϵϵϱ͘/ŶĂŶĂůLJƐŝƐϭ͕ƉĂƟĞŶƚƐǁŝƚŚƐĞǀĞƌĞĂĐƵƚĞŬŝĚŶĞLJŝŶũƵƌLJ;</ͿǁĞƌĞ ĐŽŵƉĂƌĞĚƚŽƉĂƟĞŶƚƐǁŝƚŚŽƵƚ</͕ŽŵŝƫŶŐƉĂƟĞŶƚƐǁŝƚŚĐŚƌŽŶŝĐŬŝĚŶĞLJĚŝƐĞĂƐĞ

;<Ϳ͘/ŶĂŶĂůLJƐŝƐϮ͕ĂůůƐƚĂŐĞ</ǁĂƐĐŽŵƉĂƌĞĚƚŽŶŽ</͕ŽŵŝƫŶŐƉĂƟĞŶƚƐǁŝƚŚ

<͘/ŶĂŶĂůLJƐŝƐϯ͕ƉĂƟĞŶƚƐǁŝƚŚƐĞǀĞƌĞ</ǁĞƌĞĐŽŵƉĂƌĞĚƚŽƉĂƟĞŶƚƐǁŝƚŚŽƵƚ</͕

ŝŶĐůƵĚŝŶŐƉĂƟĞŶƚƐǁŝƚŚ<ŝŶĂŶĂĚũƵƐƚĞĚĂŶĂůLJƐŝƐ͕͘ĐLJƚŽƐŝŶĞ͖'͕ŐƵĂŶŝŶĞ͖͕ĂĚĞŶŝŶĞ͖

5

‹š˜ƒ”‹ƒ–•™‡”‡•–—†‹‡†‹ƒ••‘…‹ƒ–‹‘™‹–Šˆ‘”–Š‡ϐ‹”•––‹‡‹

Study IV. Three of the variants were within the IL6 gene (rs10499563, rs1474347, and rs13306435). A haplotype analysis performed based on

ƒŽŽϐ‹˜‡IL6 variants revealed no association between any of the haplotypes within two haploblocks and AKI. Variants rs2493133 in AGT, rs2868371 in HSPB1, and rs10421768 in HAMP had an OR less than 1. Moreover, the

G-allele of rs10421768 in HAMPŠƒ†–Š‡…‘ϐ‹†‡…‡‹–‡”˜ƒŽ‘ˆͲǤ͸ͻ–‘ͲǤͻͷ in the tertiary analysis (p-value 0.009) (Figure 14).

&ŝŐƵƌĞϭϰ͘KĚĚƐƌĂƟŽƐ;ďŽdžĞƐͿĂŶĚϵϱйĐŽŶĮĚĞŶĐĞŝŶƚĞƌǀĂůƐ;ǁŚŝƐŬĞƌƐͿĨŽƌŵŝŶŽƌ ĂůůĞůĞƐŽĨǀĂƌŝĂŶƚƐƌƐϭϬϰϵϵϱϲϯ͕ƌƐϭϰϳϰϯϰϳ͕ƌƐϭϯϯϬϲϰϯϱ͕ƌƐϮϰϵϯϭϯϯ͕ƌƐϮϴϲϴϯϳϭ͕

ĂŶĚƌƐϭϬϰϮϭϳϲϴ͘/ŶĂŶĂůLJƐŝƐϭ͕ƉĂƟĞŶƚƐǁŝƚŚƐĞǀĞƌĞĂĐƵƚĞŬŝĚŶĞLJŝŶũƵƌLJ;</ͿǁĞƌĞ ĐŽŵƉĂƌĞĚƚŽƉĂƟĞŶƚƐǁŝƚŚŽƵƚ</͕ŽŵŝƫŶŐƉĂƟĞŶƚƐǁŝƚŚĐŚƌŽŶŝĐŬŝĚŶĞLJĚŝƐĞĂƐĞ

;<Ϳ͘/ŶĂŶĂůLJƐŝƐϮ͕ĂůůƐƚĂŐĞ</ǁĂƐĐŽŵƉĂƌĞĚƚŽŶŽ</͕ŽŵŝƫŶŐƉĂƟĞŶƚƐǁŝƚŚ

<͘/ŶĂŶĂůLJƐŝƐϯ͕ƉĂƟĞŶƚƐǁŝƚŚƐĞǀĞƌĞ</ǁĞƌĞĐŽŵƉĂƌĞĚƚŽƉĂƟĞŶƚƐǁŝƚŚŽƵƚ</͕

ŝŶĐůƵĚŝŶŐƉĂƟĞŶƚƐǁŝƚŚ<ŝŶĂŶĂĚũƵƐƚĞĚĂŶĂůLJƐŝƐ͕͘ĐLJƚŽƐŝŶĞ͖͕ĂĚĞŶŝŶĞ͖'͕ŐƵĂŶŝŶĞ͖

/>ϲ͕ŝŶƚĞƌůĞƵŬŝŶϲ͖'d͕ĂŶŐŝŽƚĞŶƐŝŶŽŐĞŶ͖,^Wϭ͕ŚĞĂƚƐŚŽĐŬƉƌŽƚĞŝŶĨĂŵŝůLJ;ƐŵĂůůͿŵĞŵďĞƌϭ͖

,DW͕ŚĞƉĐŝĚŝŶĂŶƟŵŝĐƌŽďŝĂůƉĞƉƟĚĞ͘

Additional analyses with septic patients and with cardiac surgery patients are presented in Table 13 (previously unpublished data). This

†ƒ–ƒ’”‡•‡–•‘•‹‰‹ϐ‹…ƒ–ƒ••‘…‹ƒ–‹‘•„‡–™‡‡˜ƒ”‹ƒ–•™‹–Š•‡˜‡”‡

in these subgroups. However, the following variants showed suggestive evidence of increasing the risk of AKI in cardiac surgery patients: IL6 rs1800796 C–allele (OR 3.78, 95% CI 1.20–11.97, p-value 0.024) and C-alleles in rs699 and rs2493133 in AGT (OR 2.05, 95% CI 1.12–3.73, p-value 0.019, and OR 1.65, 95% CI 0.99–2.74, p-value 0.055, respectively).

Table 13. ZĞƐƵůƚƐŽĨĂĚĚŝƟŽŶĂůĂŶĂůLJƐĞƐǁŝƚŚƐĞƉƟĐƉĂƟĞŶƚƐĂŶĚĐĂƌĚŝĂĐƐƵƌŐĞƌLJ ƉĂƟĞŶƚƐ;ƉƌĞǀŝŽƵƐůLJƵŶƉƵďůŝƐŚĞĚĚĂƚĂͿ͘^EW͕ƐŝŶŐůĞŶƵĐůĞŽƟĚĞƉŽůLJŵŽƌƉŚŝƐŵ͖KZ͕

5

TNFA rs1800629 A 664 1.01

(0.75–1.36) 0.96 196 0.92

(0.45–1.88) 0.82

IL10 rs1800896 G 664 0.82

(0.66–1.03) 0.082 196 0.68

(0.42–1.11) 0.12

IL6 rs10499563 C 668 1.04

(0.79–1.38) 0.76 199 0.99

(0.45–2.17) 0.98

IL6 rs1800796 C 671 0.95

(0.49–1.83) 0.87 197 3.78 (1.20–

11.97) 0.024

IL6 rs1800795 G 669 0.92

(0.74–1.13) 0.42 198 0.88

(0.55–1.40) 0.58

IL6 rs1474347 A 667 0.91

(0.74–1.13) 0.38 197 0.89

(0.55–1.42) 0.61

IL6 rs13306435 A 670 0.81

(0.42–1.56) 0.52 198 0.94

(0.34–2.55) 0.9

CXCL8 rs4073 A 666 0.99

(0.80–1.23) 0.95 198 1.02

(0.63–1.64) 0.95

NOS3 rs2070744 C 664 0.97

(0.77–1.22) 0.78 196 0.82

(0.49–1.37) 0.45

NFKB1A rs1050851 A 670 1.14

(0.85–1.53) 0.37 198 0.63

(0.32–1.25) 0.19

AGT rs699 C 333 0.79

(0.58–1.07) 0.13 113 2.05

(1.13–3.73) 0.019

AGT rs2493133 C 671 0.87

(0.70–1.09) 0.22 197 1.65

(0.99–2.74) 0.055

VEGFA rs2010963 C 661 0.85

(0.66–1.10) 0.22 195 1.07

(0.59–1.93) 0.82

VEGFA rs3025039 T 669 1.32

(0.98–1.78) 0.071 198 1.44

(0.78–2.68) 0.24

EPO rs1617640 G 663 1.00

(0.81–1.24) 1 196 1.26

(0.75–2.13) 0.38 SUFU rs10748825 G 664 1.08

(0.87–1.35) 0.49 196 0.96

(0.58–1.59) 0.88

HIF1A rs11549465 T 664 0.89

(0.53–1.50) 0.66 196 0.91

(0.29–2.87) 0.87

PNMT rs876493 C 663 1.01

(0.81–1.26) 0.94 196 0.86

(0.50–1.49) 0.59

MPO rs7208693 A 664 0.89

(0.63–1.25) 0.5 196 0.80

(0.33–1.93) 0.62

COMT rs4680 G 663 1.06

(0.86–1.30) 0.61 196 1.29

(0.75–2.21) 0.37

HSPB1 rs2868371 G 669 0.88

(0.68–1.14) 0.33 199 0.96

(0.52–1.77) 0.89

SFTPD rs2243639 T 671 0.90

(0.72–1.12) 0.33 198 1.06

(0.65–1.71) 0.82

SFTPD rs721917 G 669 1.05

(0.84–1.30) 0.67 198 1.03

(0.63–1.68) 0.91 HAMP rs10421768 G 671 0.89

(0.70–1.15) 0.15 199 0.74

(0.40–1.37) 0.34 BBS9 rs10262995 T 671 0.72

(0.44–1.18) 0.19 199 1.31

(0.34–5.06) 0.7 WK͕ĞƌLJƚŚƌŽƉŽŝĞƟŶ͖^h&h͕ƐƵƉƉƌĞƐƐŽƌŽĨĨƵƐĞĚŚŽŵŽůŽŐ͖,/&ϭ͕ŚLJƉŽdžŝĂͲŝŶĚƵĐŝďůĞĨĂĐƚŽƌ ϭͲĂůƉŚĂ͖WEDd͕ƉŚĞŶLJůĞƚŚĂŶŽůĂŵŝŶĞEͲŵĞƚŚLJůƚƌĂŶƐĨĞƌĂƐĞ͖DWK͕ŵLJĞůŽƉĞƌŽdžŝĚĂƐĞ͖KDd͕

ĐĂƚĞĐŚŽůͲKͲŵĞƚŚLJůƚƌĂŶƐĨĞƌĂƐĞ͖,^Wϭ͕ŚĞĂƚƐŚŽĐŬƉƌŽƚĞŝŶĨĂŵŝůLJ;ƐŵĂůůͿŵĞŵďĞƌϭ͖^&dW͕

ƐƵƌĨĂĐƚĂŶƚƉƌŽƚĞŝŶ͖,DW͕ŚĞƉĐŝĚŝŶĂŶƟŵŝĐƌŽďŝĂůƉĞƉƟĚĞ͖^ϵ͕ĂƌĚĞƚͲŝĞĚůƐLJŶĚƌŽŵĞϵ͘

6 DISCUSSION

AKI is a complex genetic syndrome with several distinguishable sub

’Š‡‘–›’‡•Ǥ‘”‡‘˜‡”ǡ–Š‡ˆƒ…––Šƒ–•‡˜‡”ƒŽ†‡ϐ‹‹–‹‘•‘ˆƒ”‡—•‡†‹

–ƒ†‡ǡŠƒ•Šƒ’‡”‡†‡ˆˆ‘”–•–‘…‘’ƒ”‡”‡•‡ƒ”…Šϐ‹†‹‰•ǤŠ‡‰‡‡–‹…

predisposition to AKI has been investigated in limited samples and with predetermined hypotheses. In addition, some pieces of evidence from hypothesis-free studies have been published on AKI phenotypes in septic shock¹⁴⁸ and cardiac surgery associated AKI.^198,199 Thus far, the only successful replication study was performed on variants of apoptosis-related genes in patients with septic shock.

ϲ͘ϭ'Ed/^^K/d/KE^dh/^Z'Z/E'</;/Ϳ

Literature about genetic predisposition to AKI in adults was systematically

”‡˜‹‡™‡†ǡ”‡•—Ž–‹‰‹ʹͺ‘”‹‰‹ƒŽƒ”–‹…Ž‡•ǤŠ‡•‡ƒ”–‹…Ž‡•”‡’‘”–‡†ϐ‹†‹‰•

‘ „‘–Š –Š‡ †‡˜‡Ž‘’‡– ƒ† ‘—–…‘‡ ‘ˆ Ǥ ‘™‡˜‡”ǡ †‡ϐ‹‹–‹‘• ‘ˆ

™‡”‡Š‡–‡”‘‰‡‡‘—•ǤŠ‹•…ƒ„‡†—‡–‘–Š‡™‹†‡”ƒ‰‡‘ˆ†‡ϐ‹‹–‹‘•

previously in use for deterioration in renal function. In addition, the studies report heterogeneous clinical settings. On several occasions the replication resulted in ambiguous results, as the originally reported association could not be replicated in an independent study cohort. Many of the studies were

—†‡”’‘™‡”‡†ǡ™Š‹…Š‹‰Š–‡š’Žƒ‹–Š‡‹ƒ„‹Ž‹–›–‘”‡’Ž‹…ƒ–‡ϐ‹†‹‰•Ǥ Clark and Baudouin designed a scoring system with 10 criteria that each score 1 point.¹⁴⁵ Their aim was to systematically evaluate the quality of published literature on genetic-association studies in human sepsis. The criteria was validated in a systematic review by Lu and colleagues to assess the quality of 16 studies about AKI.¹⁴⁶ Six of these studies were included in Study I that utilized the same scoring system. However, these criteria, dated back to 2006, are not well suited for modern genetic study design.

Recently it has been recognized that inadequate sample size generates

ˆƒŽ•‡’‘•‹–‹˜‡ƒ••‘…‹ƒ–‹‘•‹‰ƒŽ•ǡƒ†–Š—•ǡ•’—”‹‘—•ϐ‹†‹‰•ƒ”‡”‡’‘”–‡†Ǥ

‘‡–Š‡Ž‡••ǡƒŽ–Š‘—‰Š•‘‡‘†‹ϐ‹…ƒ–‹‘•™‡”‡’‡”ˆ‘”‡†‹–Š‡…”‹–‡”‹ƒǡ it was possible to achieve a good score relatively easily. It may be argued that the most important factors for the quality of a genetic association study are internal replication and sample size, which should be taken in to account in the assessment.

Reporting study statistics, the primer sequence, and the case group

†‡ϐ‹‹–‹‘™‡”‡–Š‡…”‹–‡”‹ƒ–Šƒ–”‡…‡‹˜‡†–Š‡„‡•–•…‘”‡‹–—†›Ǥ‘‰ƒ‹

6

ƒ••ƒ›™ƒ•–‘„‡‹†‡–‹ϐ‹‡†ǤŠ‡…‘‡”…‹ƒŽ‰‡‘–›’‹‰…‘’ƒ‹‡•‘ˆ–‡

release primer engineering software to be used in association with their genotyping facilities, thus enabling the exact replication if desired. To gain

ƒ ’‘‹– ‹ †‡ϐ‹‹–‹‘ ‘ˆ –Š‡ …ƒ•‡ ‰”‘—’ǡ –Š‡ ’ƒ–‹‡–• ƒ† ™‡”‡ –‘ „‡

†‡•…”‹„‡†•‘–Šƒ–”‡’Ž‹…ƒ–‹‘‹•’‘••‹„Ž‡Ǥ˜‡‹ˆ–Š‹•…”‹–‡”‹‘™ƒ•ˆ—Žϐ‹ŽŽ‡†

”ƒ–Š‡”‘’–‹ƒŽŽ›ǡ–Š‡†‡ϐ‹‹–‹‘•™‡”‡Š‡–‡”‘‰‡‡‘—•Ǥ

Power calculations, replication, method description, and reporting blinded-experiment were the criteria that received worst score in Study I.

The reviewed studies were inadequate in sample size, thus the non-reporting of power calculations undermines the reliability of their results.

‹’‘”–ƒ–•–‡’‹‰‡‡”ƒ–‹‰•…‹‡–‹ϐ‹…‡˜‹†‡…‡‹•–Š‡˜ƒŽ‹†ƒ–‹‘‘ˆƒ

’”‹‘”‹ϐ‹†‹‰•„›ƒ”‡’Ž‹…ƒ–‹‘•–—†›Ǥˆ–Š‡•–—†›™ƒ•—†‡”’‘™‡”‡†–‘

draw the original conclusion, this would explain the low level of replication performed. The requirements for quality in the data handling were addressed by criteria about the method description and blinding, seen perhaps too redundant to report by many.

Study I found the mean score for quality of the studies inadequate — 6.4 out of 10. In order to gain consensus between the study reviewers,

†‹•…—••‹‘•–‘‘’Žƒ…‡ƒ„‘—––Š‡•’‡…‹ϐ‹…“—ƒŽ‹–›…”‹–‡”‹ƒǡ”‡•—Ž–‹‰‹–Š‡

evolution of the criteria. The interpretation of rather subjective original …”‹–‡”‹ƒ’”‘˜‡†–‘„‡†‹ˆϐ‹…—Ž–ǤŠ‡…‘””‡Žƒ–‹‘„‡ˆ‘”‡…‘•‡•—•ǡ‡ƒ•—”‡†

‘ƒ’‡ƒ”ƒǯ•”Š‘…‘””‡Žƒ–‹‘…‘‡ˆϐ‹…‹‡–ǡ™ƒ•‰‘‘†ǡ„‘–Š‹–‡”Ǧ”‡˜‹‡™‡”

within Study I and between the systematic reviews.

Most of the articles in Study I were based on candidate gene approach.

Not uncommonly only few candidate variants within a gene, selected with versatile plausibility, are suggested to associate with the studied trait. However, the coverage of genetic variation has been poor, thus few reproducible associations have been found with this approach.²⁰⁰ The inferences made from these studies evade conclusive evidence on a genetic variant predisposing to AKI, even if some generalizations can be drawn.

Š‡‘•–‹˜‡•–‹‰ƒ–‡†˜ƒ”‹ƒ–•™‡”‡™‹–Š‹’”‘Ǧƒ†ƒ–‹Ǧ‹ϐŽƒƒ–‘”›

mediator genes. A variant in TNFA was reported in association with AKI susceptibility^194,196 and outcome,155,175,196 although the majority of studies disputed these associations.176,180,190,192,193 Another controversial candidate gene is ACE, in which a variant was reported with associations in the opposite direction,^174,197 or none at all.180,183,190,193 Angiotensin converting enzyme functions in the renin-angiotensin-aldosterone pathway and

Šƒ• ƒ ‡ˆˆ‡…– ‘ ”‡ƒŽ „Ž‘‘† ϐŽ‘™Ǥ –—†‹‡• ƒ„‘—– ”‡ƒŽ Š‡‘†›ƒ‹…• constitute a common phenotype of interest. In the phenotype of cardiac-surgery-associated AKI, a common hypothesis was the protective role of APOEǦɂͶȂƒŽŽ‡Ž‡ǤŠ‹•’‘Ž›‘”’Š‹•‹•”‡…‘‰‹œ‡†–‘Šƒ˜‡–Š”‡‡’‘••‹„Ž‡

ƒŽŽ‡Ž‡•ǡɂʹǡɂ͵ǡƒ†ɂͶǤŠ‡ƒ’‘Ž‹’‘’”‘–‡‹‹•‘™–‘Šƒ˜‡’Ž‡‹‘–”‘’‹…

roles in addition to lipid transport. The kidney protective role that is familiar from the phenotype of chronic kidney disease²⁰¹ was reported in AKI development.174,179,195 The contradictory role of APOEǦɂͶ ƒ• ƒ ”‹•

increasing allele has been presented in association with cardiovascular disease²⁰² and Alzheimer’s disease.^203,204

Frank and colleagues¹⁴⁸ performed a large-scale genotyping study that was in many regards superior to the other articles in Study I. Although not truly hypothesis free, due to the choice of genotyping panel within 2100 …ƒ”†‹‘˜ƒ•…—Žƒ”ǡ ‡–ƒ„‘Ž‹…ǡ ƒ† ‹ϐŽƒƒ–‘”› ‰‡‡•ǡ –Š‡‹” ”‡•—Ž–• ™‡”‡

…‘ϐ‹”‡† —•‹‰ ƒ ‹–‡”ƒŽ ”‡’Ž‹…ƒ–‹‘Ǥ Š‡ ϐ‹˜‡ ˜ƒ”‹ƒ–• –Šƒ– ”‡ƒ‹‡†

•‹‰‹ϐ‹…ƒ– ‹ –Š‡ ˜ƒŽ‹†ƒ–‹‘ …‘Š‘”– ™‡”‡ Ž‘…ƒ–‡† ‹ ‰‡‡• ™‹–Š‹ –Š‡

apoptosis pathway.

ϲ͘ϮWKWdK^/^Z>d'Ed/sZ/d/KE/E</;//Ϳ

The exact pathophysiological mechanisms of AKI are unknown. Formerly, the common approach to test for association with genetic markers has been focusing on known variants discovered in other phenotypes, often CKD.^205,206 Frank and colleagues presented a study that performed large-scale genotyping on 887 septic-shock patients to test AKI associating apoptosis-related genes.¹⁴⁸ Š‡‹” ϐ‹†‹‰• ƒ”‡ –Š‡ „ƒ•‹• ‘ˆ ”‡’Ž‹…ƒ–‹‘ ‹ –—†› ǡ ‹

™Š‹…Š –™‘ ‘ˆ –Š‡ ’”‡˜‹‘—•Ž› ‹†‡–‹ϐ‹‡† ”‹• ˜ƒ”‹ƒ–• ™‡”‡ •—……‡••ˆ—ŽŽ›

replicated in a sample of 478 critically ill patients with septic shock.

Š‡ …‘ϐ‹”‡† ƒ••‘…‹ƒ–‹‘• ™‡”‡ ™‹–Š‹ SERPINA4 and SERPINA5 genes. These variants, rs2093266 and rs1955656, respectively, are in tight correlation, which could indicate that one of them, or yet another risk variant that is in linkage with the aforementioned, is driving the association with AKI and the other is a proxy for the phenomenon. The minor alleles of these variants provided protection from AKI, both in the original study and the FINNAKI cohort. An additive genetic test model provided the

’”‹ƒ”›ƒ••‘…‹ƒ–‹‘•‹‰ƒŽǡ™Š‹…Š™ƒ•…‘ϐ‹”‡†‹ƒ†‘‹ƒ–ǡ„—–‘–

in a recessive, test model. Thus, it seems that one copy of the minor allele

•—ˆϐ‹…‡•–‘‹ϐŽ‹…–’”‘–‡…–‹‘ˆ”‘Ǥ

–—†›•—……‡‡†‡†‹…‘ϐ‹”‹‰–Š‡’”‡˜‹‘—•Ž›”‡’‘”–‡†ƒ••‘…‹ƒ–‹‘•

‹–™‘‘—–‘ˆϐ‹˜‡˜ƒ”‹ƒ–•Ǥ‘–Š–Š‡…ƒ•‡•ƒ†…‘–”‘Ž•™‡”‡…”‹–‹…ƒŽŽ›‹ŽŽ

patients from a prospectively collected multicenter study, with robust

†‡ϐ‹‹–‹‘‘ˆˆ‘”–Š‡’Š‡‘–›’‡Ǥ

The association analysis was repeated in a larger patient cohort, with 653 septic patients. No evidence of association was seen, suggesting a difference in AKI phenotype from patients with septic shock. Septic shock is a more severe form of illness with a greater risk of death.¹⁴¹ Thus, these ϐ‹†‹‰•ƒ”‡‘–‰‡‡”ƒŽ‹œƒ„Ž‡–‘ƒŽŽ•‡’–‹…’ƒ–‹‡–•Ǥ‘™‡˜‡”ǡ…ƒ””‹‡”•‘ˆ–Š‡

6 two functional domains, an active site and a heparin-binding site.²⁰⁹ The

functions of kallistatin are relevant in septic AKI. Hence it is plausible that a genetic variant in SERPINA4 could explain the better outcome in relation to AKI in septic shock. In murine models, the association of kallistatin

™‹–Š •—’’”‡••‡† ‹ϐŽƒƒ–‘”› ”‡ƒ…–‹‘ ƒ† ‘”‰ƒ ‹Œ—”›ǡ²¹⁰ diminished mortality in established sepsis,²¹⁰ enhanced survival in lung injury relate to sepsis,²¹¹ and renoprotection in diabetes²¹²•—’’‘”–•–Š‡•‡ϐ‹†‹‰•Ǥ

SERPINA5 encodes a protein that inhibits activated protein C (APC).

The functions of this protein C inhibitor are related to tumor growth and

‡–ƒ•–ƒ•‹•ǡ ƒ† ƒ††‹–‹‘ƒŽŽ› ‹– ‹Š‹„‹–• –Š‡ ƒ–‹Ǧ‹ϐŽƒƒ–‘”› ƒ…–‹˜‹–› ‘ˆ

APC.^213,214 In addition to driving the effect of variants in SERPINA4, the variant in SERPINA5 may pose protection against septic AKI through

–Š‹• ‹ϐŽƒƒ–‘”› ˆ—…–‹‘Ǥ Šƒ• ’”‡˜‹‘—•Ž› „‡‡ †‡•…”‹„‡† –‘ Šƒ˜‡

„‡‡ϐ‹…‹ƒŽ ‡ˆˆ‡…–• ‹ •‡˜‡”‡ •‡’•‹•²¹⁵ without conclusive evidence.²¹⁶ The consequences of the variant in SERPINA5 on the expression of the gene, however, remain a question, and the APC is not necessarily inhibited in patients with resilience against AKI.

–—†› ™ƒ• —ƒ„Ž‡ –‘ …‘ϐ‹” –Š‡ ƒ••‘…‹ƒ–‹‘• „‡–™‡‡ ƒ†

the other three variants tested, that locate in the BCL2 and SIK3 genes.

Variants rs8094315 and rs12457893 are located in the intron of BCL2, which encodes the integral outer mitochondrial membrane protein. BCL2 is able to block apoptotic death in certain types of cells.²¹⁷ SIK3, with variant rs625145, encodes a member of a regulatory AMP-activated protein kinase family. In macrophages this protein is found to suppress gene expression

‘ˆ‹ϐŽƒƒ–‘”›‘Ž‡…—Ž‡•Ǥ²¹⁸

Some evidence exists about the role of apoptosis as mechanism in septic AKI. No signs of tubular cell injury by necrosis, but rather with apoptosis, was found in a septic AKI murine model comparing the pathophysiology to that of ischemia-reperfusion AKI.²¹⁹Š‡ϐ‹†‹‰•Šƒ˜‡„‡‡…‘–”‘˜‡”•‹ƒŽ

in human postmortem kidney biopsy studies, in favor of apoptosis in patients who died of septic shock,⁹³ compared to no apoptosis in patients who died of sepsis.⁹⁴ More recently, a varying extent of tubulointerstitial apoptosis was found in patients who died of septic AKI.²²⁰ In a systematic review about histopathological changes in experimental models of septic ǡ–Š‡‘˜‡”ƒŽŽ•–”—…–—”ƒŽ‹Œ—”›™ƒ•ˆ‘—†–‘„‡‹Ž†ǡ™‹–Š‘•’‡…‹ϐ‹…

histopathologic changes and high occurrence of tubular cell apoptosis when acknowledged (21/102 studies).²²¹

It is disputed whether apoptosis has a role in the diverse and temporally evolving syndrome of septic AKI.⁹⁰ Between individuals in septic shock, however, genetic variants in apoptosis-related genes appear to contribute to AKI susceptibility.

ϲ͘ϯ/ZKEDdK>/^DZ>d'Ed/sZ/d/KE/E</;///Ϳ

Several lines of evidence suggest that the HMOX1 promoter repeat polymorphism associates with AKI development. In Study III, the short allele (less than 27 GT repeats) was shown to increase the odds of AKI in septic patients. The reverse, however, was previously reported for cardiac-surgery associated AKI.¹⁴⁹ Leaf and colleagues found an association between the long allele (equal or more than 27 GT repeats) and AKI in 2377 patients that had undergone cardiac surgery with cardiopulmonary bypass.¹⁴⁹ The predisposing event in the form of sepsis or cardiac surgery, and hence, the AKI phenotype, were different in these studies.

In Study III, a prospectively collected sample of cases and controls, from the same cohort of critically ill patients, was studied. The patients were systematically characterized and diagnosed with sepsis, and criteria for AKI were carefully considered. KDIGO staging was employed, however for the purpose of evading bias, the mild cases of AKI (KDIGO stage I) were omitted from the analysis.^5,222 The phenotype of patients with mild AKI was surmised to be somewhat different from the outcome phenotype of severe AKI (KDIGO stages 2 or 3) that was compared to septic controls without AKI.

Similar to previous reports,149,150,223 the frequency distribution of the dinucleotide repeats was bimodal, peaking at 23 and 30 alleles. In Study III,

–Š‡ƒŽŽ‡Ž‡†‹—…Ž‡‘–‹†‡…‘—–•™‡”‡ˆ—”–Š‡”˜‡”‹ϐ‹‡†„›…ƒŽ…—Žƒ–‹‰–Š‡‹

the sequence data of 15 homozygote patients. The validity of the reliability

‘ˆ ‰‡‘–›’‹‰ ™ƒ• –Š—• •–”‡‰–Š‡‡†Ǥ Š‡ ƒŽŽ‡Ž‡ Ž‡‰–Š …Žƒ••‹ϐ‹…ƒ–‹‘

™ƒ• ’‡”ˆ‘”‡† ƒ……‘”†‹‰ –‘ –Š‡ –™‘Ǧ…Žƒ•• …Žƒ••‹ϐ‹…ƒ–‹‘ –Šƒ– ‡ƒ† ƒ†

colleagues employed.¹⁴⁹ As described in the literature, both two^149,224 and three class^150,165…Žƒ••‹ϐ‹…ƒ–‹‘Šƒ•„‡‡ƒ’’Ž‹‡†ǡƒ†ƒƒ††‹–‹‘ƒŽƒƒŽ›•‹•

™ƒ•’‡”ˆ‘”‡†‹–—†›—•‹‰–Š”‡‡Ǧ…Žƒ••…Žƒ••‹ϐ‹…ƒ–‹‘Ǥ

The GT repeat polymorphism in HMOX1 is known to associate to critical illness.^150,165 In addition, the high concentration of the HMOX1 protein product HO-1 has been associated to organ dysfunction in the critically ill.¹⁵⁰ The association between the length of the dinucleotide repeat and the concentration of the protein product HO-1 is known to be inverse.150,151,225

DNA sequences with Z-potential, including pyrimidine-purine alternating sequences like GT repeat, negatively affect transcriptional activity.^225,226 In a transient-transfection assay of HMOX1 inducibility, due to oxidant stress, it was concluded that the short allele in comparison to the long allele was more inducible and the basal expression level higher.²²⁵ The regulation of the expression of HO-1 varies between cell and tissue types.²²⁵

6 Ǧͳ ˆ—…–‹‘ ”‡Žƒ–‡• –‘ Š‡‡ …ƒ–ƒ„‘Ž‹•ǡ •’‡…‹ϐ‹…ƒŽŽ› ‡…Šƒ‹••

to reutilize iron from heme-containing proteins. At the end of the red-blood-cell life-span, which is approximately 120 days, it will be engulfed into the reticuloendothelial system (macrophages) and degraded, after which heme will be free in the cytoplasm of these cells.²²⁷ Energy- and oxygen-dependent enzymatic degradation takes place in the endoplasmic reticulum of these cells:

Heme + NADPH + 3 O₂ heme oxygenase Biliverdin IX + CO + Fe³ + NAPD⁺ The heme oxygenase enzyme cleaves heme-ring containing iron into linear tetrapyrrole biliverdin IX. In this reaction, the catalyzing enzyme heme oxygenase can be either in its type I isomer that is substrate inducible, or in its type II isomer that is constitutive. The enzyme cleaves an alpha-methene bridge producing carbon monoxide (CO); no other reaction in the human body produces CO and thus the amount of heme degradation could be determined by monitoring CO production.²²⁷ Following this, the biliverdin IX is further reduced into bilirubin IX in an energy dependent reaction:

Biliverdin IX + NADPH biliverdin reductace Bilirubin IX + NAPD⁺

Bilirubin, which is poorly water soluble, is transported in plasma bound

–‘ •‡”— ƒŽ„—‹Ǥ Ž„—‹ …‘–ƒ‹• ‘‡ Ž‘™Ǧƒˆϐ‹‹–› •‹–‡ ƒ† ‘‡ Š‹‰ŠǦ

ƒˆϐ‹‹–›•‹–‡ˆ‘”„‹Ž‹”—„‹Ǥ–Š‡Ž‹˜‡”ǡ„‹Ž‹”—„‹‡–‡”•Š‡’ƒ–‘…›–‡•ǡ™Š‡”‡‹–

binds to several detoxifying cytosolic proteins and is rendered more water soluble by conjugation to diglucuronide, after which it is excreted to bile.

Bilirubin diglucuronide is degraded in the intestine by bacterial hydrolases, after which bilirubin is reduced into urobilinogens and further oxidized to urobilins that are excreted in the feces.²²⁷

HO-1 assists cells in surviving stressful situations, in kidneys as well as in other organs.²²⁸†‡ˆ‡•‡ƒ‰ƒ‹•–‘˜‡”„Ž‘™‹ϐŽƒƒ–‘”›”‡ƒ…–‹‘•ǡ–Š‹•

reaction is fundamental.^229–233 However, in association to some malignant diseases, the maximal survival of a cell line is uncalled for.²³³ Moreover, the increase in HO-1 has been associated with a more dismal outcome in the critically ill¹⁵⁰ as well as patients that have been resuscitated from cardiac arrest outside hospital.¹⁶² An optimal therapeutic range for HO-1 expression is suggested for the critically ill.^150,234 The effect of elevated HO-1 in the critically ill appears to be detrimental and exceeding the therapeutic range would explain this.

Increased HO-1 concentrations associate with AKI.²³⁵ Accordingly, Study III indicated that in patients without AKI, the plasma HO-1 concentration was lower. Notwithstanding, kidney protection by HO-1 induction is acknowledged in murine models²³⁶ as well as in association with AKI.^237,238

The experimentally induced models in another species mimicking

”Šƒ„†‘›‘Ž›•‹• ‘” ‹•…Š‡‹ƒǦ”‡’‡”ˆ—•‹‘ ƒ”‡ ƒ ‰”‘•• •‹’Ž‹ϐ‹…ƒ–‹‘ ‘ˆ

the function of HO-1 in the presence of critical illness. Interestingly, the regulation of HMOX1 expression in humans is different from that of mice.^236,239 Nevertheless, in the presence of complicating critical illness the role of HO-1 enzyme induction is seen as an injury repairing response due to the acute stressor.¹⁵⁰

ϲ͘ϰ/E&>DDd/KEͲZ>d'Ed/sZ/d/KE/E</;/sͿ

The attempted replication of several variants in Study IV failed to prove

–Š‡‹‹–‹ƒŽϐ‹†‹‰•–‘„‡–”—‡ǤŠ‡ʹ͹˜ƒ”‹ƒ–•–Šƒ–™‡”‡–‡•–‡†ƒ”‡™‹–Š‹

ͳͺ…ƒ†‹†ƒ–‡‰‡‡•–Šƒ–”‡Žƒ–‡–‘ǡˆ‘”‡šƒ’Ž‡ǡ‹ϐŽƒƒ–‹‘ǡ…‡ŽŽ•—”˜‹˜ƒŽǡ and circulation. The sample size of nearly 3000 critically ill adult patients

‡š…‡‡†•–Š‡•ƒ’Ž‡•‹œ‡•‘ˆ–Š‡•–—†‹‡•’”‡•‡–‹‰–Š‡‹‹–‹ƒŽϐ‹†‹‰•ǤŠ‡

majority of these studies were underpowered and of inadequate quality, as found in Study I. The predisposing setting was most often cardiac surgery, and ICU patients and patients with sepsis were studied less often. Nearly all of the original studies relied on a candidate polymorphism study setting and the biological function of the variant was rarely known. Due to the heterogeneity of the published studies, no meta-analysis was conducted, even in the most recent review article.²⁴⁰ Few replication attempts have been undertaken previously, and the results have been contradicting.

In genetically complex syndromes such as AKI, the realistic assumptions

In genetically complex syndromes such as AKI, the realistic assumptions

In document Genetic predisposition to acute kidney injury in critically ill adults (sivua 50-0)