Balance between maternal antiviral response and placental transfer of protection in gestational SARS-CoV-2 infection

Abstract


Introduction 55
The maternal immune response protects the growing fetus from the vertical 56 have been shown to display antiviral response from birth (24). In COVID-19 98 patients the profile of inflammatory cytokine responses has been associated with 99 disease outcomes. While an interferon response is conducive to viral clearance 100 (25), the enduring production of inflammatory cytokines IL-1, IL-6, IL-8, IL-10 101 and IL-18 has been linked to disease severity (26, 27). Pregnant women have 102 been described to mount a mild inflammatory response to SARS-CoV-2 infection 103 that is reflected in the fetus, even in the absence of vertical viral transmission, as

Results 115
Population 116 Our study comprised 72 pregnant women recruited from May 2020 to February 117 2021, before the introduction of COVID-19 vaccination to the general population. 118 Sixty pregnant women tested positive (CoV-2 + ; Supplemental Table 1) and 12 119 tested negative (CoV-2 -; Supplemental Table 2) for SARS-CoV-2 by PCR on 120 nasopharyngeal swabs ( Figure 1A). We divided the 60 SARS-CoV-2 + pregnant 121 women according to gestational age at the time of infection. In the recovered 122 second trimester (2R) group were the ones that tested positive between 85 and 123 154 days before delivery. The recovered third trimester (3R) group was 124 comprised by the ones that tested positive between 20 to 77 days before delivery. 125 The ongoing (3O) group included the ones that tested positive within 11 days of 126 delivery ( Figure 1A). Pregnant women were mandatorily tested for SARS-CoV-2 127 by PCR in the context of monthly health provision, which allowed us to document 128 both the time of viral infection and resolution. Fifty percent of pregnant women 129 were symptomatic, with four pregnant women being admitted to the hospital due 130 to COVID-19 symptomatology, non-invasive oxygen support was provided to 131 other three. Within the SARS-CoV-2 + population, 50 mother-umbilical cord 132 matched samples were collected at the time of birth, and in 82% of the deliveries 133 a nasopharyngeal swab was performed to the newborn, with 3 babies testing 134 positive but remained asymptomatic and one baby that tested negative being 135 admitted to neonatal ICU (Supplemental Table 1). Additional 9 pregnant women 136 were recruited between July of 2021 and February of 2022 having undergone 137 COVID-19 mRNA BNT162b2 vaccine. Maternal and cord blood samples were 138 collected from 6 dyads ( Figure 1A; Supplemental Table 3).

NAbs and with impaired transplacental transfer of IgG-NAbs in 142 symptomatic infections 143
The transfer of antibodies from maternal to fetal circulation relies on the neonatal 144   Figure 1D). To evaluate whether the 160 decrease in anti-spike/RBD antibody transfer in ongoing infection was due to a 161 saturation of the FcRn (31-34), we assessed the transfer ratio of total IgG, which 162 we found to be ~1 in all analyzed groups, including the 3O ( Figure 1E). Moreover, 163 total IgG transfer ratio was also significantly higher than the one of anti-with premature deliveries ( Figure 1F). As expected, total IgA, IgG and IgM 166 remained constant, regardless of gestational age of infection (Supplemental 167 Figure 1C) and we found that only the production of anti-RBD IgM antibodies was 168 positively associated with total IgM levels ( Figure 1 G-I), albeit with a modest R 2 169

value. 170
We then addressed whether the sex of the fetus had an impact on the immune 171 response to infection. We did not observe changes in maternal production of anti-172 spike IgA, IgM or IgG antibodies in function of the sex of the fetus (Supplemental 173 Figure 1D). However, in dyads with ongoing maternal SARS-CoV-2 infection, we 174 observed a significant decrease in anti-spike IgG levels in male neonates which 175 was not observed in female neonates (Supplemental Figures 1 E, F). infection gave rise to higher frequency of NK cells in maternal circulation ( Figure  244 3 G, Supplemental Figure 3C). The implication that early immune response might 245 be important to curtail infection, made us look at IgM antibodies, and we found 246 that anti-spike IgM levels were higher in asymptomatic ongoing infections ( Figure  247 3H). Even though, the presence of maternal symptoms did not significantly 248 impact NK cell frequency in the cord blood ( Figure 3I), we observed that delayed 249 umbilical cord clamping led to higher frequency of NK cells in neonates born to 250 mothers with ongoing infection ( Figure 3J). This enrichment in NK cell frequency 251 in the cord blood appears to be specific and not a simple result of increased blood 252 influx, as delayed clamping did not alter neonatal frequency of B cells ( Figure 3K) 253 nor CD4 + T cells ( Figure 3L). Similarly, delayed clamping had no impact on 254 neonatal NK cell frequency in resolved SARS-CoV-2 infections ( Figure 3M). 255 Altogether, our results indicate that an early immune response conveyed by NK 256 cells, and to a lesser extent by anti-spike IgM, is associated with asymptomatic 257 maternal infection. To pinpoint long-lasting cellular immune changes brought by SARS-CoV-2 262 infection, we compared the B and CD4 + T profiles in pregnant women that had 263 been either infected or vaccinated at similar gestational ages. Delivery occurred second vaccine dose, respectively. The frequency of B and CD4 + T cells was 266 equivalent between SARS-CoV-2 infected (2R and 3R) and COVID-19 267 vaccinated group (Figures 4 A, B). When we looked in more detail to CD4 + T cell 268 activation and migratory status, we observed that SARS-CoV-2 infection did not 269 alter CD4 + T cell activation state ( Figure 4C) but led to an increase in the 270 frequency of CXCR5 + T cells, and to a lesser extent of CCR6 + T cells (Figure 4  2 infected mothers, the increase in cytokine concentration in the cord blood 327 ranged from 2-to 1.10 7 -fold ( Figure 7G). In contrast, neonates born to control 328 non-infected and non-vaccinated mothers displayed a more contained cytokine 329 upregulation both in terms of magnitude, ranging from 2-to ~7500-fold, and in 330 terms of the cytokines upregulated which were restricted to MCP-1, IL-8, IL-1 331 and IL-10 (Supplemental Figure 5). Next, we illustrated through polar plots the 332 possible contribution of maternal symptomatology and neonatal sex to the 333 inflammatory profile of cord blood. In dyads with cord/maternal ratio higher than 334 1.5 for at least one cytokine, there was a trend, albeit not statistically significant, All in all, these data underscore that even asymptomatic maternal infections elicit 341 fetal exposure to at least one inflammatory cytokine, in at least 50% of the cases. Moreover, in this dyad all but two cytokines were present at a higher 355 concentration in the cord blood than in the mother ( Figure 8F). 356

358
In this study, we investigated the cellular and molecular underpinnings regulating 360 the balance between maternal SARS-CoV-2 clearance and transplacental 361 transfer of protection over the course of pregnancy. We show that while anti-viral 362 cytokine production is short-term, SARS-CoV-2 gestational infection is 363 associated with high concentration of inflammatory cytokines, which could be 364 detected months after the infection. Finally, we describe that even asymptomatic 365 maternal SARS-CoV-2 infection elicits fetal exposure to several inflammatory 366 cytokines, in a significant proportion of neonates. 367 In our cohort, three neonates tested positive for SARS-CoV-2 by PCR. However, 368 we could not exclude that these infections occurred post-nataly, rather than in 369 utero. Indeed, umbilical cord testing for anti-spike IgM, which more accurately with NAbs (10%) by using anti-RBD IgG as an approximate measure for is associated with reduced transplacental transfer of anti-RBD IgG. Moreover, we 410 found that elevated IL-6 was associated with lower anti-RBD IgG transfer, while 411 elevated IL-10 and IL-23 correlated with increased anti-RBD IgG transplacental 412 transfer. It is possible that IL-6, IL-10 and IL-23 could affect NAb transfer 413 efficiency by altering NAb glycosylation and or FcRn-NAb binding avidity (19, 58). 414 Next, we identified how the nature of the neutralizing antibody response would 415 impinge on NAbs placental transfer, as IgA and IgM are not amenable to FcRn-416 mediated transplacental transfer (20). We found that viral neutralization was not 417 restricted to IgG isotype. In fact, purified IgA and IgM were ~3-fold more efficient 418 at neutralizing a SARS-CoV-2 pseudovirus than purified IgG, in ongoing 419 infections. It is possible that in vivo IgM neutralizing activity might even be higher, 420 as IgM has been estimated to trigger a complement response that is many-fold 421 Our study has some limitations. It is difficult to quantify the relative contribution 518 of each antibody isotype to both maternal and fetal protective immunity. We did 519 not assess whether anti-spike/RBD IgA/IgM had undergone affinity maturation 520  Blood samples were immediately processed. 555 In total, we recruited 63 pregnant women suspected of SARS-CoV-2 infection. 558 Three participants were excluded from the study due to loss of biospecimens' 559 integrity. One participant was excluded from the control group due to prior 560 exposure to SARS-CoV-2. (B) Anti-spike and anti-RBD IgG endpoint titers (n=50 dyads).
(D) Correlation between anti-spike and anti-RBD IgG transfer ratio and elapsed time between diagnosis and delivery (n=39).
(J) Cord blood NK cell frequency upon either immediate or delayed umbilical cord clamping, in ongoing maternal infections (3O, n=21).
(M) Cord blood NK cell frequency upon either immediate or delayed umbilical cord clamping, in second and third trimester recovered (2R+3R) infection (n=10).
Data represents mean ± SD for parametric tests, or median ± IQR for non-parametric tests. p values *p < 0.05; ns, not significant were determined by unpaired t test (A, B     Data represents median ± IQR for non-parametric tests. p values *p < 0.05; ns, not significant were determined by Mann-Whitney test (A-F) and Spearman correlation (G-J). Effect sizes were determined by r = correlation coefficient r (A-F).  (G) Ratio of cytokines in cord and maternal blood (n=27).
(J and K) Polar plots of cytokine profile of (J) female and (K) male neonates, in dyads with cord/maternal ratio>1.5 (n=27).
p values ***p < 0.001, ****p < 0.0001; ns, not significant were determined by non-parametric paired Wilcoxon test (A-F) and unpaired t test (H-K). Effect sizes were determined by r = correlation coefficient r (A-F). (F) Cytokine ratio between cord and maternal blood for dyad 26.
Dashed line indicates assay cutoff; nd: not detectable. p values *p < 0.05, ****p < 0.0001; ns, not significant were determined by non-parametric paired Wilcoxon test (A, C, E) and parametric paired t test (E). Effect sizes were determined by r = correlation coefficient r (A, C, E) and d = Cohen's d (E).