2.3.1. Humoral Response
People with antibody deficiencies generate a particularly poor antibody response to vaccination; indeed, “diagnostic vaccination” is used to clinically evaluate individuals for the presence of humoral IEI [
75]. It was therefore an early concern whether these populations would develop any meaningful protection from COVID-19 vaccines, especially those using novel mRNA vaccine platforms. Here, we reviewed 58 papers evaluating the humoral immune response generated by individuals with different IEI to COVID-19 primary vaccines and/or booster doses (
Table S1,
Figure 2). In many cases [
69,
76,
77,
78,
79,
80,
81,
82,
83], studies age-matched IEI patients to healthy controls to compare the robustness and quality of antibody responses to that of the general population. This is important because otherwise, known age-related changes in the immune system would confound the results of each study. The published study designs varied from observational [
71,
77,
83,
84,
85,
86,
87,
88,
89,
90,
91,
92,
93,
94,
95,
96] to interventional studies and clinical trials [
80,
82,
83,
97,
98]. Furthermore, humoral responses were reported by seroconversion, antibody concentration/titers, and, in some cases, virus neutralization/pseudo-neutralization capacity (
Figure 3 and
Table 2). Samples were generally collected between approximately 2 weeks to one-month post-vaccination; with a smaller portion of studies collecting samples at 2-, 3-, or 6-months post-vaccination [
88,
89,
99]. These differences in sampling time post vaccination can have dramatic effects on the status of the immune response. For these reasons, care should be taken when comparing studies as variations in sample collection time points and experimental methodology will affect the quantitative ranges etc. (
Figure 3 and
Table 2).
Overall, seroconversion in populations with IEI was much more variable and at a lower magnitude compared to healthy controls across the primary series and additional doses [
74,
76,
77,
78,
83,
94,
99,
100,
101,
102,
103,
104,
105,
106,
107,
108], regardless of methods of measurement differed. This is due to the heterogenous genetic and immunologic phenotype of disease in these individuals. In one study, individuals with IEI overwhelmingly displayed poorer seroconversion rates (ranging from 16.6%-61.4%) compared to healthy controls (90-100%), after two doses of mRNA vaccine (Comirnaty, SpikeVax) [
73]. Despite the lower seroconversion rate post dose 2 it is clear that booster doses increased seropositivity (76-100%) in IEI patients after 3 vaccine doses [
73,
74,
77].
As described in
Section 1.1, the immune defects in patients with IEI are highly diverse and likely impact the outcomes of COVID-19 vaccination. Most patients examined in the literature had PAD, in line with this group of diseases being the most prevalent IEI group (42% worldwide IEI diagnoses) [
18]. In those receiving an mRNA vaccine, 20.6% to 90.0% of individuals with PAD seroconverted after two doses [
71,
74,
81]. In contrast in a cohort of individuals with primary or secondary antibody deficiencies vaccinated with either Comirnaty or ChAdOx1 nCoV-19, the overall seroprevalence rose from 61.4% to 76.0% after a third dose [
73].
Tandon et al. 2023 assessed the differences between an adult (n = 62) and relatively smaller pediatric (n = 13) PADs cohort, finding that antibody levels were higher in the pediatric cohort after a primary series (Comirnaty), as well as after a monovalent booster, but only for those classified as having mild PADs [
109]. There was otherwise limited research on pediatric PAD cohorts. A study by Shin et al. specifically investigated the clinical and/or immunological factors in a group of mixed PADs that might predict their COVID-19 vaccination response [
110]. They found a lower seroconversion rate and neutralizing antibody levels to vaccination in CVID patients compared to other PAD categories (e.g., IgG deficiency, IgG2 subclass deficiency, and specific antibody deficiency (SpAD)). This was associated with a number of factors including lower baseline IgG, IgA, and proportion of Bmem cells, a higher proportion of CD8
+ Tmem cells, and a co-diagnosis of an autoimmune disorder [
110]. Mizera et al. also found that in their cohort study of various PADs, that all patients in their CVID subgroup had around 5-fold lower titers than other subgroups [
106]. Thus, as stated by Quinti
et al., with the high immunological and clinical diversity within CVID, each patient should be assessed individually for their capacity to mount a protective vaccination response [
111].
In contrast, several studies report on smaller groups of patients with specific IEI diagnoses. For example, one report of individuals with STAT1-GOF (n=7) found that 57% seroconverted after two doses of vaccine [
112]. Another case report on one individual with XMEN/MAGT1 deficiency reported a reactive spike-specific total antibody response after two doses of mRNA vaccine [
90]. In addition, Timothy et al. also completed a retrospective chart review of eight patients with CTLA-4 deficiency (two children and six adults). Of those who had their SARS-CoV-2-specific IgG levels tested (n=2), antibody titers were lower (<6.25%) than that of the healthy unvaccinated controls [
113]. Multiple studies evaluated antibody responses in individuals with XLA. In all studies patients had nearly nonexistent antibody responses after two doses with either ChAdOx1 nCoV-19 or Comirnaty [
81,
98]. In fact, in one case study of three people with XLA, only one was able to mount a measurable antibody response post-vaccination [
114] (
Figure 2). This is consistent with the inherent B-cell defects in these individuals whereby peripheral B cells are low or absent, resulting in an absence of serum immunoglobulins, and an inadequacy in generating a B-cell response to vaccination [
16,
17]. In this case, the observed post-vaccination antibody response is likely due to the IgRT received by the patient to treat their underlying antibody deficiency (See
Section 3.1).
In several studies with more diverse cohorts of IEI, seroconversion differed by subgroup. Van Leeuwen et al. found significantly lower seroconversion rates in those with CVID, XLA, and CID compared to SpAD, undefined antibody deficiency, and phagocytic defects [
71]. In studies which evaluated possible predictors of antibody responses in IEI population(s) certain factors, such as condition severity [
78,
80], previous history of autoimmunity [
69], underlying immunosuppressive treatment [
78,
80] or low numbers of Bmem cells pre-vaccination [
81], were associated with poor responses (
Figure 2).
This is important as previous studies have established a connection between the humoral and cellular response to infection outcomes in healthy individuals [
115].
Interestingly, in some studies which compared both COVID-19 naïve individuals and those with a history of previous SARS-CoV-2 infection, recovery from natural SARS-CoV-2 infection corresponded with a more efficacious antibody response to the vaccination series (i.e., “hybrid immunity”). This includes a more robust seroconversion rate [
97,
98], higher neutralization capacity [
76], and better protection against the
Delta variant [
116]. Additionally, Abella et al. found that even in the cohort with plasma cell diseases, those with multiple myeloma who had hybrid immunity had a higher plasma neutralization capacity than infection-naïve individuals 3 months post-vaccination [
117]. This provides insights in to whether hybrid immunity provides a more robust immune response than vaccination alone, and the capacity of natural infection to shape immunity.
Recently Abo-Helo et al. found that antibody levels remained poor in people with more severe disorders of humoral immunity 5-6 months after vaccination, suggesting the need for additional doses to preserve sufficient protection [
82]. Van Leeuwen et al. conducted a large multicenter study of patients with PADs, following them up to 6 months past the primary vaccine series, and a through a third dose for smaller subset of 50 CVID patients. Their cohort all received mRNA-1273 for their primary series, and either BNT16b2 or an unspecified vaccine formulation for a third dose if applicable. At 6 months past the primary series, the decline in antibodies ranged from 5.9 to 11.2-fold, measured in GMT of S-specific IgG. However, seropositivity rates at 6 months ranged from 73% to 98%, with those in the XLA subgroup maintaining a relatively low rate of 24% [
118]. Another study of a mixed cohort of more than 50 patients who received at least one dose of any mRNA-based OR inactivated vaccine, or a combination of both, found no significant different in antibody levels across formulations [
119].
There remains limited knowledge about the efficacy and durability of responses generated to different COVID-19 vaccine formulations in people with IEI. Leung et al. studied the humoral response to both Comirnaty and CoronaVac (“Sinovac”) formulations in their cohort of mixed IEIs and found that 73% of individuals were seropositive after a full series of either formulation [
93]. Some studies found that a full Comirnaty vaccination series was more efficacious than ChAdOx1 nCoV-19 [
120], and this superior protection was maintained for about 6 months past the second dose [
73]. Overall, most individuals reported had increased neutralizing activity after a partial or full vaccine series [
69,
70,
71,
73,
83,
93,
98,
116,
120]. Furthermore, a follow-up from the COV-AD study using a live virus neutralization assay with Vero cells, found 99.8% of their study subjects had neutralization activity after a booster dose [
98]. However, whether these are a bona fide response to vaccination or a consequence of receiving SARS-CoV-2-specific antibody-containing IgRT was not evident, although in at least one study, IgRT did not appear to be significantly associated with either titer or neutralization activity [
74]. It is also important to note the sparsity of data surrounding SARS-CoV-2-specific Bmem generation in individuals with IEI both in response to natural infection and SARS-CoV-2 infection, which is necessary to provide information about the generation and durability of the Bmem cells in these populations and their ongoing ability to produce antibody.
Furthermore, to date, some IEI patients may have received five or more vaccine doses; however, information on the efficacy of these later doses is limited. This is likely due to lack of available participants with sufficiently time elapsed since their most recent dose [
98], and also insufficient time elapsed to enable sample analysis and reliable reporting of findings.2.3.2. Memory T-cells
Like humoral immunity, cellular immunity in people with IEI is more variable than that of healthy populations. Recent research focuses attention specifically on the mechanisms of Bmem, Tmem, and innate immune cells following COVID-19 vaccination in IEI populations.
As such, we reviewed 43 papers investigating the cellular response overall and its relationship with the humoral response of IEI populations after receiving COVID-19 vaccination (
Table S2). As with antibody production, IEI patients demonstrated suboptimal CD4
+ and/or CD8
+T-cell responses compared to the general population, ranging from ~30-80% of that of healthy controls [
70,
77,
93,
98,
121,
122]. A number of studies found that individuals with IEI could mount measurable CD4
+ and/or CD8
+T-cell responses [
86,
104,
105,
123], measured as percent of IFN-γ-positive T-cells via IFN-γ release assay (IGRA), ELISpot assay, or flow cytometry (
Figure 2,
Table 2).
Studies that evaluated both cellular and humoral post-vaccination immunity found varying levels of correlation between these two responses, making it difficult to draw definitive conclusions about their relationship [
93,
96,
97]. Sauerwein et al. found an association between intact spike-specific CD4
+ T-cell responses and normal IgG responses, defined as a titer of spike protein 3x the limit or more, in individuals with CVID or milder antibody deficiency [
100]. Oyaert et al. found a positive and significant correlation between anti-spike IgG and IFN-γ levels in IEI and chronic kidney disease (CKD) subgroups [
99]. Notably, Gao et al. were able to confirm a positive correlation between antibodies and spike-specific CD4
+ T-cells in all their study groups, both healthy and immunocompromised [
124].
Other studies found that a negative humoral response was associated with a poor cellular response. Bergman et al. found that low levels of naïve CD4
+ T-cells correlated with a poor antibody response in the CVID participants of their prospective clinical trial [
76]. Similarly, Antoli et al. established that low CD4
+ and CD8
+ T-cell counts were a predictor of poor antibody response specifically in CVID patients [
85]. Barmettler et al. also observed a correlation between low CD4
+ T-helper cells and low specific antibody responses in their mixed cohort of individuals with secondary and severe primary PADs, including complicated CVID/SpAD, activated P13K-syndrome, TACI deficiency, NFKB1 deficiency, HGG, and IgG subclass deficiency. This observation was evident for both mRNA and the Ad26.CoV.S vaccines [
79].
In contrast, several studies reported no association or a discordant relationship between the humoral and cellular responses [
93,
96,
97]. For example, Pulvirenti et al. found nearly no T-cell response in their vaccinated cohort of people with CVID, despite this cohort presenting a positive (albeit suboptimal) IgG response [
97]. Findings from a subsequent paper by Pulvirenti et al. demonstrate that T-cell abnormalities resulting from CVID may play a role in lack of specific antibody responses after vaccination [
91]. Interestingly, in a separate cohort followed by Goschl et al. had a robust specific CD4
+ and CD8
+ T-cell response to vaccination, they found no significant correlation between that response and antibody levels [
125]. Similarly, a case study of five individuals with PAD by Steiner et al. found a robust cellular response, yet no humoral response in their entire cohort [
126]. Mizera
et al., in their study on a cohort of individuals with primary antibody deficiencies receiving 2-4 doses, found that anti-SARS-CoV-2 IgG did not correlate with either CD4+ or CD8+ levels, but rather NK cells [
106]. This underscores the suggestion made by previous papers to measure both the cellular and humoral response to develop a more complete post-vaccination immune profile in each individual [
127,
128].
Importantly, some studies evaluated the spectrum of T-cell responses by IEI patients to identify specific characteristics or factors that may predict poor or robust responses. Shin et al. found a correlation between baseline immune profiles as well as comorbidities, where poor vaccine responders among their mixed cohort of PADs had autoimmune diseases, low levels of baseline IgG, low naïve CD8
+ T-cells, and higher levels of effector CD8
+ Tmem [
110]. Gao et al. who looked extensively at T-cell profiles between subgroups of immunocompromised individuals, found overall that specific CD8
+ T-cells were generated at a higher rate than CD4
+ T-cells. These frequencies were comparable at 35 days after dose 2 between healthy controls, IEI, HIV, and HSCT, but lower in SOT and CLL [
124]. This study found that individuals enrolled in the trial generated higher levels of higher frequencies of SARS-CoV-2 specific CD4
+ T-cells after two vaccine doses [
124]. Similarly, van Leeuwen et al. conducted a multicenter study of different subgroups of IEI compared to healthy controls. Only the CVID subgroup had significantly lower IFN-γ response rates compared to healthy controls. Within this subgroup, lower responses were associated with age and lymphoproliferative diseases [
71].
Similarly, we identified case reports of vaccination among less researched IEI, including STAT1-GOF and X-linked
SASH3 deficiency, which found that nearly all studied individuals were able to mount both a cellular response and humoral response despite a suboptimal baseline immune profile due to their inborn immune disorders [
112,
129] (
Figure 2). Of note, Bloomfield et al. found that 85% of their STAT1-GOF cohort, including 2 individuals on the JAK inhibitor ruxolitinib, produced a cellular response comparable to healthy controls [
112].
Like with post-vaccination antibody levels, exogenous factors such as natural SARS-CoV-2 infection and vaccine formulation appear to impact T-cell responses in certain settings. Some studies found that SARS-CoV-2 infection increased T-cell activity in antibody-deficient individuals [
130], while another found poor CD4
+ and CD8
+ T-cell responses after both previous infection and subsequent vaccination [
97]. Other studies have reported that prior SARS-CoV-2 infection results in improved immune response after the first [
131] and second [
98] vaccine dose. Additionally, other studies have reported the phenomena of pre-existing cross reactive CD4
+ T-cells that can recognize SARS-CoV-2 [
132]. The COV-AD study follow-up found that receipt of heterologous vaccines by infection-naïve IEI individuals was associated with a significantly higher likelihood of T-cell response generation, compared to homogenous boosting [
73]. Lin et al. found that a third dose increased the amount of SARS-CoV-2-specific CD4
+ and CD8
+ T-cells generated [
87], and Goda et al. were able to demonstrate a 30% increase in T-cell activity after a third dose of Comirnaty following two ChAdOx1 n CoV-19 in individuals with CVID [
120].
The rise of viral variants has sparked questions about the durability of T-cells, as well as whether strain-specific T-cells are broadly protective against variants in the IEI population. While few studies have been published to date in either healthy populations or individuals with IEI, there is some evidence of sustained reactivity.
In terms of longevity, Hurme
et al. conducted a longitudinal cohort study over 22 months in a mixed cohort of primary and secondary antibodies deficiencies, finding a robust cellular response. Specifically, across 1-4 doses of either mRNA-1273, BNT162b2, or ChAdOx1, 95% of their cohort had detectable T-cell responses after the primary series as measured by flow cytometry, dropping down to 78% at 3 months following a fourth dose. In a smaller subset of individuals, they observed that CD4+ cell responses were higher than CD8
+ at the same timepoints when stimulated with wildtype SARS-CoV-2 strain [
102].
Van Leeuwen et al. extended their investigations to observe cellular responses at 6 months post-primary vaccination in their mixed IEI cohort, finding that while IFN levels decreased significantly in most of their subgroups (XLA, CVID, IgG/SPAD), these responses were still detectable in varying levels [
118]. Barouch et al. found that the Ad26.CoV2.S vaccine induces durable cellular immunity with minimal decreases over the course of 8 months in a healthy population [
133]. In another healthy individual cohort, Liu et al. found that both the Ad26.CoV2.S and Comirnaty vaccine induced durable spike-specific CD4
+ and CD8
+ T-cell responses. Here, 82-84% CD8
+ T-cell responses were cross-reactive to the Omicron B.1.1.529 variant [
122]. Sanchez et al. recently demonstrated that people with XLA had an increased T-cell response to variants compared to healthy controls, while individuals with CVID had a reduced response [
134]. More evidence is needed to understand the post-vaccine T-cell response in general as well as its ability to provide protection in the face of future SARS-CoV-2 variants. This coupled with correlation between generation and durability of Tmem and Bmem responses, will provide crucial information about whether, for example, generation of Tmem responses in the absence of Bmem is sufficient for protection from disease.